2026
Ibrahim, Mossab; Gervás, Pablo; Méndez, Gonzalo
AI-Cinema: A Hybrid Framework for Arabic Movie Scenario Generation with Traditional Storytelling and Cultural Dialogues Journal Article
In: Complexity, vol. 2026, no. 1, pp. 9978799, 2026, ISSN: 1099-0526.
@article{mossab2026aicinema,
title = {AI-Cinema: A Hybrid Framework for Arabic Movie Scenario Generation with Traditional Storytelling and Cultural Dialogues},
author = {Mossab Ibrahim and Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez},
editor = {Penousal Machado},
url = {https://onlinelibrary.wiley.com/doi/epdf/10.1155/cplx/9978799},
doi = {10.1155/cplx/9978799},
issn = {1099-0526},
year = {2026},
date = {2026-03-13},
urldate = {2026-01-01},
journal = {Complexity},
volume = {2026},
number = {1},
pages = {9978799},
abstract = {AI-Cinema is a hybrid neural-symbolic framework addressing the critical challenge of preserving cultural authenticity in Arabic movie scenario generation. The framework integrates transformer-based neural language models (AraT5-base and AraGPT2-medium) with symbolic reasoning encoded in OWL-DL ontologies and SWRL rules to ensure linguistic fluency, narrative coherence, and cultural preservation. AI-Cinema introduces a three-tier architecture comprising a data layer, cultural embedding layer, and scenario generation layer. Central to its design is an attention-based cultural embedding mechanism leveraging ArabicVerbNet (12,500 culturally annotated verbs) and ArabicNameNet (3653 names with regional annotations), complemented by a mathematically grounded harmony function that dynamically balances neural generation with symbolic constraints. In experiments on 2740 Arabic narratives, AI-Cinema achieves a BLEU-4 score of 32.76 (± 0.6), representing a 5.0% relative improvement over AraBERT-Gen (p \< 0.01, paired bootstrap test) and a 27.5 percentage point gain in cultural preservation metric (CPM) compared with MARBERT-Gen (82.3% vs. 54.8%). The framework maintains 92.3% dialectal accuracy across Modern Standard Arabic and six regional dialects, with explicit evaluation of code-switching scenarios. Expert evaluations by 30 Arabic literature scholars demonstrate strong interannotator agreement (Fleiss’ κ = 0.78, p \< 0.001), with 87% of evaluators rating generated narratives as culturally authentic (score ≥ 4 on a 5-point scale). Current limitations include reduced performance on underrepresented dialects (Yemeni: 85.6% and Sudanese: 86.1%) and complex code-switching scenarios (76.3% for 3+ dialects). The framework incorporates transparent labeling mechanisms for AI-generated cultural content to address authenticity concerns. All resources are publicly available at https://github.com/Mossab82/AI-Cinema.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
AI-Cinema is a hybrid neural-symbolic framework addressing the critical challenge of preserving cultural authenticity in Arabic movie scenario generation. The framework integrates transformer-based neural language models (AraT5-base and AraGPT2-medium) with symbolic reasoning encoded in OWL-DL ontologies and SWRL rules to ensure linguistic fluency, narrative coherence, and cultural preservation. AI-Cinema introduces a three-tier architecture comprising a data layer, cultural embedding layer, and scenario generation layer. Central to its design is an attention-based cultural embedding mechanism leveraging ArabicVerbNet (12,500 culturally annotated verbs) and ArabicNameNet (3653 names with regional annotations), complemented by a mathematically grounded harmony function that dynamically balances neural generation with symbolic constraints. In experiments on 2740 Arabic narratives, AI-Cinema achieves a BLEU-4 score of 32.76 (± 0.6), representing a 5.0% relative improvement over AraBERT-Gen (p < 0.01, paired bootstrap test) and a 27.5 percentage point gain in cultural preservation metric (CPM) compared with MARBERT-Gen (82.3% vs. 54.8%). The framework maintains 92.3% dialectal accuracy across Modern Standard Arabic and six regional dialects, with explicit evaluation of code-switching scenarios. Expert evaluations by 30 Arabic literature scholars demonstrate strong interannotator agreement (Fleiss’ κ = 0.78, p < 0.001), with 87% of evaluators rating generated narratives as culturally authentic (score ≥ 4 on a 5-point scale). Current limitations include reduced performance on underrepresented dialects (Yemeni: 85.6% and Sudanese: 86.1%) and complex code-switching scenarios (76.3% for 3+ dialects). The framework incorporates transparent labeling mechanisms for AI-generated cultural content to address authenticity concerns. All resources are publicly available at https://github.com/Mossab82/AI-Cinema.
2025
Ibrahim, Mossab; Gervás, Pablo; Méndez, Gonzalo
CULTURA: A Multi-Agent Neural–Symbolic System for Culturally-Aware Arabic Story Generation Proceedings Article
In: C. Dima, Ferrando (Ed.): PRIMA 2025: Principles and Practice of Multi-Agent Systems, Springer, Cham, 2025, ISBN: 978-3-032-13562-9.
@inproceedings{mossab2025cultura,
title = {CULTURA: A Multi-Agent Neural\textendashSymbolic System for Culturally-Aware Arabic Story Generation},
author = {Mossab Ibrahim and Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez},
editor = {Dima, C., Ferrando, A., Malvone, V.},
url = {https://link.springer.com/chapter/10.1007/978-3-032-13562-9_24},
doi = {10.1007/978-3-032-13562-9_24},
isbn = {978-3-032-13562-9},
year = {2025},
date = {2025-12-15},
urldate = {2025-12-15},
booktitle = {PRIMA 2025: Principles and Practice of Multi-Agent Systems},
volume = {16366},
publisher = {Springer},
address = {Cham},
series = {LNCS},
abstract = {Contemporary Arabic narrative generation systems often fail to capture nuanced cultural authenticity, primarily addressing dialectal variation as a lexical challenge rather than encoding deeper cultural schemas. The distinction between (Egyptian) and (Levantine) for “The story began in Cairo” exemplifies cultural resonance patterns that standardized approaches overlook. We present CULTURA, a neural-symbolic framework orchestrating three specialized agents: (1) CNN-based dialect classification (94.3% accuracy across six Arabic varieties); (2) OWL 2 DL ontology encoding cultural narrative schemas; (3) PPLM-enhanced generation with schema-guided cultural steering. Evaluation on 2,847 prompts across six dialects shows improvements over strong baselines: Dialect F1 = 84.2% vs. 68.9% (AraT5), CMPI = 0.82 vs. 0.49, and human cultural resonance 6.1 ± 0.3 vs. 4.3 ± 0.4 (7-point Likert), all by paired bootstrap, p\<0.001.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Contemporary Arabic narrative generation systems often fail to capture nuanced cultural authenticity, primarily addressing dialectal variation as a lexical challenge rather than encoding deeper cultural schemas. The distinction between (Egyptian) and (Levantine) for “The story began in Cairo” exemplifies cultural resonance patterns that standardized approaches overlook. We present CULTURA, a neural-symbolic framework orchestrating three specialized agents: (1) CNN-based dialect classification (94.3% accuracy across six Arabic varieties); (2) OWL 2 DL ontology encoding cultural narrative schemas; (3) PPLM-enhanced generation with schema-guided cultural steering. Evaluation on 2,847 prompts across six dialects shows improvements over strong baselines: Dialect F1 = 84.2% vs. 68.9% (AraT5), CMPI = 0.82 vs. 0.49, and human cultural resonance 6.1 ± 0.3 vs. 4.3 ± 0.4 (7-point Likert), all by paired bootstrap, p<0.001.
Ibrahim, Mossab; Gervás, Pablo; Méndez, Gonzalo
Neural-Symbolic AI for Culturally Adaptive Arabic Sign-Language Translation with Motion-Capture Avatars Proceedings Article
In: C. Dima, Ferrando (Ed.): PRIMA 2025: Principles and Practice of Multi-Agent Systems, Springer, Cham, 2025, ISBN: 978-3-032-13562-9.
@inproceedings{mossab2025prima,
title = {Neural-Symbolic AI for Culturally Adaptive Arabic Sign-Language Translation with Motion-Capture Avatars},
author = {Mossab Ibrahim and Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez},
editor = {Dima, C., Ferrando, A., Malvone, V.},
url = {https://link.springer.com/chapter/10.1007/978-3-032-13562-9_25},
doi = {10.1007/978-3-032-13562-9_25},
isbn = {978-3-032-13562-9},
year = {2025},
date = {2025-12-15},
urldate = {2025-12-15},
booktitle = {PRIMA 2025: Principles and Practice of Multi-Agent Systems},
volume = {16366},
publisher = {Springer},
address = {Cham},
abstract = {Arabic Sign Language (ArSL) presents formidable communication barriers for 17\textendash23 million deaf individuals across 460 million Arabic speakers. We introduce a neural-symbolic framework addressing dialectal variation through spatiotemporal constraint injection, embedding grammatical rules as optimizable symbolic loss terms. Our multi-agent coordination system employs symbolic validators that inject dialect-specific rules with physics-informed motion synthesis to generate culturally authentic gestures. The framework achieves 40% higher accuracy than Mahmoud et al. [1], with 95% gesture fidelity (F1@0.5 IoU) and 85 ms median latency on consumer hardware. Cultural authenticity scored 4.6/5 by native signers across 15 dialects. Real-world deployments in educational and healthcare settings elevated communication success rates from 45% to 83%. The most persistent errors (3%) stem from Levantine trilateral root conflations\textemdashexemplified by \d{s}a\d{h}\={i}fa ( , ”clear newspaper”) versus \d{s}u\d{h}uf ( , ”multiple newspapers”)\textemdashdemonstrating the model’s capacity for morphological decomposition while maintaining cultural coherence.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Arabic Sign Language (ArSL) presents formidable communication barriers for 17–23 million deaf individuals across 460 million Arabic speakers. We introduce a neural-symbolic framework addressing dialectal variation through spatiotemporal constraint injection, embedding grammatical rules as optimizable symbolic loss terms. Our multi-agent coordination system employs symbolic validators that inject dialect-specific rules with physics-informed motion synthesis to generate culturally authentic gestures. The framework achieves 40% higher accuracy than Mahmoud et al. [1], with 95% gesture fidelity (F1@0.5 IoU) and 85 ms median latency on consumer hardware. Cultural authenticity scored 4.6/5 by native signers across 15 dialects. Real-world deployments in educational and healthcare settings elevated communication success rates from 45% to 83%. The most persistent errors (3%) stem from Levantine trilateral root conflations—exemplified by ṣaḥīfa ( , ”clear newspaper”) versus ṣuḥuf ( , ”multiple newspapers”)—demonstrating the model’s capacity for morphological decomposition while maintaining cultural coherence.
Ibrahim, Mossab; Gervás, Pablo; Méndez, Gonzalo
CST-Onto: An Ontological Framework for Culturally-Grounded Narrative Generation in Arabic Proceedings Article
In: Proceedings of the Thirteenth International Conference on Knowledge Capture K-CAP 2025, pp. 174-181, ACM Association for Computing Machinery (ACM), New York, NY, USA, 2025, ISBN: 9798400718670.
@inproceedings{ibrahim2025kcap,
title = {CST-Onto: An Ontological Framework for Culturally-Grounded Narrative Generation in Arabic},
author = {Mossab Ibrahim and Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez},
url = {https://dl.acm.org/doi/pdf/10.1145/3731443.3771364},
doi = {10.1145/3731443.3771364},
isbn = {9798400718670},
year = {2025},
date = {2025-12-10},
urldate = {2025-12-10},
booktitle = {Proceedings of the Thirteenth International Conference on Knowledge Capture K-CAP 2025},
pages = {174-181},
publisher = {Association for Computing Machinery (ACM)},
address = {New York, NY, USA},
organization = {ACM},
abstract = {Contemporary Arabic narrative generation faces a critical cultural authenticity crisis, where neural models systematically violate classical adab principles despite sophisticated linguistic capabilities. We introduce CST-Onto, a hybrid neuro-symbolic framework addressing this challenge through the first domain-specific OWL/SWRL ontology for Arabic cultural narratives. Our four-tier ontology encodes 87 core cultural archetypes from expert annotations of Kal\={i}la wa Dimna and One Thousand and One Nights [13], expanded to 847 specialized concepts through systematic derivation protocols. The framework formalizes 234 SWRL rules modeling virtue dependencies (e.g., \d{h}ak\={i}m → \d{h}ilm, ʿadl, fa\d{s}\={a}\d{h}a) that directly enhance symbolic-neural integration via constraint-aware neural decoding. Our constraint-injection algorithm maintains sub-28 ms/token latency on NVIDIA V100 GPUs across 1,000 narrative segments, enabling real-time cultural schema enforcement. Evaluation using our novel CRAFT benchmark demonstrates substantial improvements: across 120 native Arabic speakers, we report role coherence (= 0.89), moral-causality consistency (= 0.83), and dialectal authenticity (= 0.92) using validated classifier methodologies. CST-Onto outputs achieved 78% human preference in pairwise comparisons (\< 0.001) against baseline models, representing a 21% improvement in cultural fidelity.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Contemporary Arabic narrative generation faces a critical cultural authenticity crisis, where neural models systematically violate classical adab principles despite sophisticated linguistic capabilities. We introduce CST-Onto, a hybrid neuro-symbolic framework addressing this challenge through the first domain-specific OWL/SWRL ontology for Arabic cultural narratives. Our four-tier ontology encodes 87 core cultural archetypes from expert annotations of Kalīla wa Dimna and One Thousand and One Nights [13], expanded to 847 specialized concepts through systematic derivation protocols. The framework formalizes 234 SWRL rules modeling virtue dependencies (e.g., ḥakīm → ḥilm, ʿadl, faṣāḥa) that directly enhance symbolic-neural integration via constraint-aware neural decoding. Our constraint-injection algorithm maintains sub-28 ms/token latency on NVIDIA V100 GPUs across 1,000 narrative segments, enabling real-time cultural schema enforcement. Evaluation using our novel CRAFT benchmark demonstrates substantial improvements: across 120 native Arabic speakers, we report role coherence (= 0.89), moral-causality consistency (= 0.83), and dialectal authenticity (= 0.92) using validated classifier methodologies. CST-Onto outputs achieved 78% human preference in pairwise comparisons (< 0.001) against baseline models, representing a 21% improvement in cultural fidelity.
Ibrahim, Mossab; Gervás, Pablo; Méndez, Gonzalo
MM-GTA: Morphological Modeling via Graph-based Text Adaptation for Arabic Dialect Transfer Proceedings Article
In: Proceedings of the Thirteenth International Conference on Knowledge Capture (K-CAP 2025), pp. 198-205, ACM Association for Computing Machinery (ACM), New York, NY, USA, 2025, ISBN: 9798400718670.
@inproceedings{mossab2025mmgta,
title = {MM-GTA: Morphological Modeling via Graph-based Text Adaptation for Arabic Dialect Transfer},
author = {Mossab Ibrahim and Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez},
url = {https://dl.acm.org/doi/pdf/10.1145/3731443.3771368},
doi = {10.1145/3731443.3771368},
isbn = {9798400718670},
year = {2025},
date = {2025-12-10},
urldate = {2025-12-10},
booktitle = {Proceedings of the Thirteenth International Conference on Knowledge Capture (K-CAP 2025)},
pages = {198-205},
publisher = {Association for Computing Machinery (ACM)},
address = {New York, NY, USA},
organization = {ACM},
abstract = {Arabic text adaptation confronts severe computational barriers from morphological complexity and dialectal fragmentation. Current graph-based methods employ static representations that fail to capture evolving morphological dependencies, causing substantial semantic degradation (28.0% cross-dialectal error rate) across dialectal boundaries.
We introduce MM-GTA (Morphological Modeling via Graphbased Text Adaptation), the first dynamic-graph framework for Arabic style transfer that jointly optimizes morphological dynamism, dialect-conditioned attention, and cultural-adaptive resource fusion (dialect-aware embedding compression). MM-GTA introduces dynamic graph reconstruction and dialect-conditioned attention, reducing cross-dialect errors by 55.8% versus AraT5 (from 28.0% to 12.3%).},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Arabic text adaptation confronts severe computational barriers from morphological complexity and dialectal fragmentation. Current graph-based methods employ static representations that fail to capture evolving morphological dependencies, causing substantial semantic degradation (28.0% cross-dialectal error rate) across dialectal boundaries.
We introduce MM-GTA (Morphological Modeling via Graphbased Text Adaptation), the first dynamic-graph framework for Arabic style transfer that jointly optimizes morphological dynamism, dialect-conditioned attention, and cultural-adaptive resource fusion (dialect-aware embedding compression). MM-GTA introduces dynamic graph reconstruction and dialect-conditioned attention, reducing cross-dialect errors by 55.8% versus AraT5 (from 28.0% to 12.3%).
We introduce MM-GTA (Morphological Modeling via Graphbased Text Adaptation), the first dynamic-graph framework for Arabic style transfer that jointly optimizes morphological dynamism, dialect-conditioned attention, and cultural-adaptive resource fusion (dialect-aware embedding compression). MM-GTA introduces dynamic graph reconstruction and dialect-conditioned attention, reducing cross-dialect errors by 55.8% versus AraT5 (from 28.0% to 12.3%).
Ibrahim, Mossab; Gervás, Pablo; Méndez, Gonzalo
NarrativeMind: A Dynamic Neural-Symbolic Decoder for Culturally-Authentic Arabic Story Generation Proceedings Article
In: Proceedings of the 31st International Conference on Cooperative Information Systems (CoopIS2025), Scitepress, 2025.
@inproceedings{ibrahim2025narrativemind,
title = {NarrativeMind: A Dynamic Neural-Symbolic Decoder for Culturally-Authentic Arabic Story Generation},
author = {Mossab Ibrahim and Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez},
url = {https://link.springer.com/chapter/10.1007/978-3-032-15538-2_7
https://doi.org/10.1007/978-3-032-15538-2_7
https://nil.fdi.ucm.es/wp-content/uploads/2025/10/NarrativeMind-CoopIS2025.pdf
https://www.insticc.org/Primoris/api/TechnicalProgram/DownloadBookofAbstract?id=4009},
year = {2025},
date = {2025-10-21},
urldate = {2025-10-21},
booktitle = {Proceedings of the 31st International Conference on Cooperative Information Systems (CoopIS2025)},
publisher = {Scitepress},
abstract = {We introduce NarrativeMind, a previously-unexplored cooperative neural-symbolic decoder in Arabic NLP that dynamically injects dialect-specific cultural constraints during generation. Our approach achieving +3.5 BLEU improvement over AraBERT (from 26.3 to 29.8) while reducing MSA bias by 58%, addressing the critical gap where traditional Arabic narrative systems inadequately capture the rhetorical sophistication embedded in morphologically complex literary forms. The hybrid architecture seamlessly integrates classical Arabic structures into BLOOMZ’s decoding through weighted interpolation, preserving essential rhetorical devices. For instance, it maintains (saj´) patterns in (wisdom in words, blessing in currency) and (jinas¯ ) wordplay as demonstrated in (the speaker said to the killer). Unlike rigid frameworks, NarrativeMind adapts fluidly across Modern Standard Arabic and six regional dialects, particularly benefiting under-resourced Maghrebi varieties. Our real-time multi-dialect collaboration employs adaptive constraint weighting, optimizing both BLEU coherence and our novel CulturalScore metric. This metric derives from 2,500 expert-annotated templates spanning classical (maqam¯ at¯ ) to contemporary (h. ikay¯ at sha´biyya ¯ ), ensuring comprehensive cultural representation. MADAR corpus evaluation (n=12,000) demonstrates substantial improvements: BLEU scores reached 29.8 ± 0.4 versus 27.1 ± 0.3 for baselines, with dialectal accuracy achieving κ = 0.76 compared to 0.70. Human evaluation involving 15 linguists and 185 native speakers validates 82.5% cultural authenticity (p \< 0.01), confirming effective cross-regional story co-creation while preserving dialectal integrity.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We introduce NarrativeMind, a previously-unexplored cooperative neural-symbolic decoder in Arabic NLP that dynamically injects dialect-specific cultural constraints during generation. Our approach achieving +3.5 BLEU improvement over AraBERT (from 26.3 to 29.8) while reducing MSA bias by 58%, addressing the critical gap where traditional Arabic narrative systems inadequately capture the rhetorical sophistication embedded in morphologically complex literary forms. The hybrid architecture seamlessly integrates classical Arabic structures into BLOOMZ’s decoding through weighted interpolation, preserving essential rhetorical devices. For instance, it maintains (saj´) patterns in (wisdom in words, blessing in currency) and (jinas¯ ) wordplay as demonstrated in (the speaker said to the killer). Unlike rigid frameworks, NarrativeMind adapts fluidly across Modern Standard Arabic and six regional dialects, particularly benefiting under-resourced Maghrebi varieties. Our real-time multi-dialect collaboration employs adaptive constraint weighting, optimizing both BLEU coherence and our novel CulturalScore metric. This metric derives from 2,500 expert-annotated templates spanning classical (maqam¯ at¯ ) to contemporary (h. ikay¯ at sha´biyya ¯ ), ensuring comprehensive cultural representation. MADAR corpus evaluation (n=12,000) demonstrates substantial improvements: BLEU scores reached 29.8 ± 0.4 versus 27.1 ± 0.3 for baselines, with dialectal accuracy achieving κ = 0.76 compared to 0.70. Human evaluation involving 15 linguists and 185 native speakers validates 82.5% cultural authenticity (p < 0.01), confirming effective cross-regional story co-creation while preserving dialectal integrity.
Méndez, Gonzalo; Gervás, Pablo
Heating up Interactions in an Agent-Based Simulation to Ensure Narrative Interest Proceedings Article
In: 17th International Conference on Agents and Artificial Intelligence (ICAART2025), pp. 693-703, 2025, ISBN: 2184-433X.
@inproceedings{mendez2025heating,
title = {Heating up Interactions in an Agent-Based Simulation to Ensure Narrative Interest},
author = {Gonzalo M\'{e}ndez and Pablo Gerv\'{a}s},
url = {https://www.scitepress.org/publishedPapers/2025/133088/pdf/index.html
https://nil.fdi.ucm.es/wp-content/uploads/2025/07/ICAART2025-ucm.pdf},
doi = {10.5220/0013308800003890},
isbn = {2184-433X},
year = {2025},
date = {2025-01-01},
urldate = {2025-01-01},
booktitle = {17th International Conference on Agents and Artificial Intelligence (ICAART2025)},
volume = {2},
pages = {693-703},
abstract = {Multi-agent systems have become important sources of inspiration for narrative generation systems, withsignificant growth in solutions based on story sifting: identifying the subset of events generated by sucha system that is worthy of being told as a story. Existing systems simulate the romantic behaviour of agentsbased on simple rules that consider models of social norms and relations, and the evolution of affinities betweenagents. The present paper describes an extension to one such simulation that inserts several sources of conflictbetween characters to induce more interesting situations that allows the creation of more engaging stories. Thesystem is empirically shown to give rise with much higher scores on metrics for narrative interest},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Multi-agent systems have become important sources of inspiration for narrative generation systems, withsignificant growth in solutions based on story sifting: identifying the subset of events generated by sucha system that is worthy of being told as a story. Existing systems simulate the romantic behaviour of agentsbased on simple rules that consider models of social norms and relations, and the evolution of affinities betweenagents. The present paper describes an extension to one such simulation that inserts several sources of conflictbetween characters to induce more interesting situations that allows the creation of more engaging stories. Thesystem is empirically shown to give rise with much higher scores on metrics for narrative interest
Gervás, Pablo; León, Carlos; Kumar, Mayuresh; Méndez, Gonzalo; Bautista, Susana
Prompting an LLM Chatbot to Role Play Conversational Situations for Language Practice Proceedings Article
In: 17th International Conference on Computer Supported Education (CSEDU2025), pp. 257–264, INSTICC Scitepress, 2025, ISSN: 2184-5026.
@inproceedings{gervas2025prompting,
title = {Prompting an LLM Chatbot to Role Play Conversational Situations for Language Practice},
author = {Pablo Gerv\'{a}s and Carlos Le\'{o}n and Mayuresh Kumar and Gonzalo M\'{e}ndez and Susana Bautista},
url = {https://nil.fdi.ucm.es/wp-content/uploads/2025/10/CSEDU2025.pdf
https://www.scitepress.org/Papers/2025/132354/132354.pdf},
doi = {10.5220/0013235400003932},
issn = {2184-5026},
year = {2025},
date = {2025-01-01},
urldate = {2025-01-01},
booktitle = {17th International Conference on Computer Supported Education (CSEDU2025)},
volume = {2},
pages = {257\textendash264},
publisher = {Scitepress},
organization = {INSTICC},
abstract = {Chatbots based on Large Language Models (LLMs) have demonstrated a remarkable ability to engage in conversations that are linguistically correct and make sense from a pragmatic point of view. A significant trait of their proven abilities is that, using verbal instructions provided as contributions to an ongoing conversation, they can be configured to provide specific content and/or modify the role that they play in the exchange. The present paper explores the feasibility of developing a framework of prompts designed with such an aim in mind. The prompting should ensure that the chatbot engages a language learner in an interaction where it proposes conversational situations of appropriate complexity, takes part in them playing the role of one of the participants, while monitoring the linguistic correctness of the contributions by the learner and providing feedback on their language performance both proactively and in response to learner requests. The paper reports on an experiment tha t tested this type of functionality students of Spanish as a second language at Aligarh Muslim University in India.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Chatbots based on Large Language Models (LLMs) have demonstrated a remarkable ability to engage in conversations that are linguistically correct and make sense from a pragmatic point of view. A significant trait of their proven abilities is that, using verbal instructions provided as contributions to an ongoing conversation, they can be configured to provide specific content and/or modify the role that they play in the exchange. The present paper explores the feasibility of developing a framework of prompts designed with such an aim in mind. The prompting should ensure that the chatbot engages a language learner in an interaction where it proposes conversational situations of appropriate complexity, takes part in them playing the role of one of the participants, while monitoring the linguistic correctness of the contributions by the learner and providing feedback on their language performance both proactively and in response to learner requests. The paper reports on an experiment tha t tested this type of functionality students of Spanish as a second language at Aligarh Muslim University in India.
Ibrahim, Mossab; Gervás, Pablo; Méndez, Gonzalo
SAHDA: A Semantic-Aware Hybrid Approach for Arabic Spell Checking with Domain Adaptation Proceedings Article
In: 12th International Conference on Information Technology (ICIT2025), pp. 343–348, IEEE IEEE, 2025, ISBN: 9798331508944.
@inproceedings{ibrahim2025sahda,
title = {SAHDA: A Semantic-Aware Hybrid Approach for Arabic Spell Checking with Domain Adaptation},
author = {Mossab Ibrahim and Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez},
url = {https://ieeexplore.ieee.org/abstract/document/11049091/
https://nil.fdi.ucm.es/wp-content/uploads/2025/10/SAHDA-ICIT2025.pdf},
doi = {10.1109/ICIT64950.2025.11049091},
isbn = {9798331508944},
year = {2025},
date = {2025-01-01},
urldate = {2025-01-01},
booktitle = {12th International Conference on Information Technology (ICIT2025)},
pages = {343\textendash348},
publisher = {IEEE},
organization = {IEEE},
abstract = {Arabic spell checking faces inherent linguistic challenges due to its complex word structures and ambiguous spellings, where roots generate diverse forms and identical spellings mask multiple meanings.
Current systems often prioritize technical accuracy over contextual understanding, struggling to preserve regional dialects or resolve semantic ambiguities. This paper introduces SAHDA, a hybrid framework that bridges computational methods with human linguistic expertise. By combining rule-based error detection, neural semantic analysis, and collaborative humanAI refinement, SAHDA addresses both structural errors and meaning-based ambiguities in written Arabic. Key innovations include context-aware dialect preservation to maintain regional linguistic diversity and adaptive learning techniques that enable efficient customization for specialized domains. Evaluations demonstrate SAHDA’s ability to resolve ambiguous terms and dialectal variations more effectively than existing approaches while minimizing overcorrection of valid regional expressions. As an open-source tool, the framework advances inclusive language technology by harmonizing standardization with dialectal richness\textemdash a critical advancement for supporting Arabic’s diverse written traditions in the digital age.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Arabic spell checking faces inherent linguistic challenges due to its complex word structures and ambiguous spellings, where roots generate diverse forms and identical spellings mask multiple meanings.
Current systems often prioritize technical accuracy over contextual understanding, struggling to preserve regional dialects or resolve semantic ambiguities. This paper introduces SAHDA, a hybrid framework that bridges computational methods with human linguistic expertise. By combining rule-based error detection, neural semantic analysis, and collaborative humanAI refinement, SAHDA addresses both structural errors and meaning-based ambiguities in written Arabic. Key innovations include context-aware dialect preservation to maintain regional linguistic diversity and adaptive learning techniques that enable efficient customization for specialized domains. Evaluations demonstrate SAHDA’s ability to resolve ambiguous terms and dialectal variations more effectively than existing approaches while minimizing overcorrection of valid regional expressions. As an open-source tool, the framework advances inclusive language technology by harmonizing standardization with dialectal richness— a critical advancement for supporting Arabic’s diverse written traditions in the digital age.
Current systems often prioritize technical accuracy over contextual understanding, struggling to preserve regional dialects or resolve semantic ambiguities. This paper introduces SAHDA, a hybrid framework that bridges computational methods with human linguistic expertise. By combining rule-based error detection, neural semantic analysis, and collaborative humanAI refinement, SAHDA addresses both structural errors and meaning-based ambiguities in written Arabic. Key innovations include context-aware dialect preservation to maintain regional linguistic diversity and adaptive learning techniques that enable efficient customization for specialized domains. Evaluations demonstrate SAHDA’s ability to resolve ambiguous terms and dialectal variations more effectively than existing approaches while minimizing overcorrection of valid regional expressions. As an open-source tool, the framework advances inclusive language technology by harmonizing standardization with dialectal richness— a critical advancement for supporting Arabic’s diverse written traditions in the digital age.
Gervás, Pablo; López-Calle, José Luis
Accounting for the Importance of Changes in Event Actuality in the Representation of Narrative Proceedings Article
In: CEUR Workshop Proceedings, pp. 117–125, 2025.
@inproceedings{gervas2025accounting,
title = {Accounting for the Importance of Changes in Event Actuality in the Representation of Narrative},
author = {Pablo Gerv\'{a}s and Jos\'{e} Luis L\'{o}pez-Calle},
url = {https://ceur-ws.org/Vol-3964/paper10.pdf},
year = {2025},
date = {2025-01-01},
urldate = {2025-01-01},
booktitle = {CEUR Workshop Proceedings},
volume = {3964},
pages = {117\textendash125},
abstract = {The mainstay of any story is the sequence of events that has happened to the characters involved in it. However, stories often mention events that have not happened \textendash wishes, dreams, duties, beliefs, plans \textendash that are sometimes extremely important for the plot as readers understand it. These events subsequently either become true, or the hope of their ever becoming true is dashed by circumstances. The initial mentions of such non-actual events very often act as drivers of the plot, or establish some of the con!icts that are essential to the story. When it happens, the change in truth status of those statements tends to operate as partial closure for the plot. In those cases, understanding the plot of the story involves being able to represent not only the events themselves but how the initial views on these events evolve throughout the story towards a resolution of the con!ict. The present paper proposes a representation of narrative that includes means for identifying non-actual events of this kind, and for keeping track of when their non-actual status changes and what it evolves to. This representation is shown to capture important aspects of narratives with complex structure in terms of shifting views on the truth value of statements fundamental to the plot, such as whether the hero is alive or whether the heroine is married to the villain or not. },
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The mainstay of any story is the sequence of events that has happened to the characters involved in it. However, stories often mention events that have not happened – wishes, dreams, duties, beliefs, plans – that are sometimes extremely important for the plot as readers understand it. These events subsequently either become true, or the hope of their ever becoming true is dashed by circumstances. The initial mentions of such non-actual events very often act as drivers of the plot, or establish some of the con!icts that are essential to the story. When it happens, the change in truth status of those statements tends to operate as partial closure for the plot. In those cases, understanding the plot of the story involves being able to represent not only the events themselves but how the initial views on these events evolve throughout the story towards a resolution of the con!ict. The present paper proposes a representation of narrative that includes means for identifying non-actual events of this kind, and for keeping track of when their non-actual status changes and what it evolves to. This representation is shown to capture important aspects of narratives with complex structure in terms of shifting views on the truth value of statements fundamental to the plot, such as whether the hero is alive or whether the heroine is married to the villain or not.
Gervás, Pablo
Interpreting Narrations of Events Witnessed: Relying on Location Data to Help Place Embedded Stories Proceedings Article
In: CEUR Workshop Proceedings, pp. 127–135, 2025.
@inproceedings{gervas2025interpreting,
title = {Interpreting Narrations of Events Witnessed: Relying on Location Data to Help Place Embedded Stories},
author = {Pablo Gerv\'{a}s},
url = {https://ceur-ws.org/Vol-3964/paper11.pdf},
year = {2025},
date = {2025-01-01},
urldate = {2025-01-01},
booktitle = {CEUR Workshop Proceedings},
volume = {3964},
pages = {127\textendash135},
abstract = {When interpreting a narrative in which someone tells a story, !rst-person accounts of events that happened to the narrator earlier are relatively easy to process. But cases in which the narrator is telling something that she witnessed \textendash but was not personally involved in \textendash require more elaborate pragmatic inferences based on the assumption that the narrator was present as a witness of the events being told. Appropriate handling of these cases requires means of correct interpretation of where events experienced and events witnessed \textendash by the narrator \textendash are happening in terms of locations in the storyworld. The present paper considers a simple model of how information on changes in location across events may be inferred from narrative discourse, and how this information can be exploited to ensure appropriate temporal placement \textendash with respect to the events of the main or frame story \textendash of the events for the embedded stories that were witnessed by a narrator at some earlier point in the discourse.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
When interpreting a narrative in which someone tells a story, !rst-person accounts of events that happened to the narrator earlier are relatively easy to process. But cases in which the narrator is telling something that she witnessed – but was not personally involved in – require more elaborate pragmatic inferences based on the assumption that the narrator was present as a witness of the events being told. Appropriate handling of these cases requires means of correct interpretation of where events experienced and events witnessed – by the narrator – are happening in terms of locations in the storyworld. The present paper considers a simple model of how information on changes in location across events may be inferred from narrative discourse, and how this information can be exploited to ensure appropriate temporal placement – with respect to the events of the main or frame story – of the events for the embedded stories that were witnessed by a narrator at some earlier point in the discourse.
2024
Gervás, Pablo; Méndez, Gonzalo
Tagging Narrative with Propp’s Character Functions Using Large Language Models Proceedings Article
In: Seventh Workshop on Narrative Extraction From Texts (Text2Story’24), pp. 137–148, 2024.
@inproceedings{gervas2024tagging,
title = {Tagging Narrative with Propp’s Character Functions Using Large Language Models},
author = {Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez},
url = {https://ceur-ws.org/Vol-3671/paper12.pdf},
year = {2024},
date = {2024-01-01},
urldate = {2024-01-01},
booktitle = {Seventh Workshop on Narrative Extraction From Texts (Text2Story’24)},
volume = {3671},
pages = {137\textendash148},
abstract = {The character functions proposed by Vladimir Propp as abstraction of plot structure to understand Russian folk tales have been popular as means of analysing narrative. Several eorts have been carried out to annotate narratives in this way manually or to apply machine learning techniques over texts previously annotated with syntactic and semantic information. The present paper explores the feasibility of annotating directly from the bare text of synopses of the stories by relying on large language models},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The character functions proposed by Vladimir Propp as abstraction of plot structure to understand Russian folk tales have been popular as means of analysing narrative. Several eorts have been carried out to annotate narratives in this way manually or to apply machine learning techniques over texts previously annotated with syntactic and semantic information. The present paper explores the feasibility of annotating directly from the bare text of synopses of the stories by relying on large language models
Gervás, Pablo; Méndez, Gonzalo
Pruning Worlds into Stories: Affective Interactions as Fitness Function Proceedings Article
In: International Conference on Computational Intelligence in Music, Sound, Art and Design (Part of EvoStar), pp. 179–193, Springer 2024.
@inproceedings{gervas2024pruning,
title = {Pruning Worlds into Stories: Affective Interactions as Fitness Function},
author = {Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez},
url = {https://doi.org/10.1007/978-3-031-56992-0_12},
year = {2024},
date = {2024-01-01},
urldate = {2024-01-01},
booktitle = {International Conference on Computational Intelligence in Music, Sound, Art and Design (Part of EvoStar)},
pages = {179\textendash193},
organization = {Springer},
abstract = {An important challenge when trying to find a story to tell about some set of events that has already happened is to identify the elements in that set of events that will make a story that moves the intended audience. One possible criterion is to consider events that involve significant changes in the emotional relations between the characters involved. The present paper explores a computational model of this particular approach to the task of storytelling. An evolutionary solution is used to explore the logs of an agent-based social simulation, using metrics on the evolution of affinity between characters as fitness function, to identify sequences of events that might be good candidates for moving stories.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
An important challenge when trying to find a story to tell about some set of events that has already happened is to identify the elements in that set of events that will make a story that moves the intended audience. One possible criterion is to consider events that involve significant changes in the emotional relations between the characters involved. The present paper explores a computational model of this particular approach to the task of storytelling. An evolutionary solution is used to explore the logs of an agent-based social simulation, using metrics on the evolution of affinity between characters as fitness function, to identify sequences of events that might be good candidates for moving stories.
Ibrahim, Mossab; Gervás, Pablo; Méndez, Gonzalo
Bridging the Past and Future: Extracting Narrative Structures from the 1001 Arabian Nights to Inspire the Generation of New Arabic Tales Journal Article
In: American Journal of Information Science and Computer Engineering, vol. 5, no. 4, pp. 360–370, 2024.
@article{ibrahim2024bridging,
title = {Bridging the Past and Future: Extracting Narrative Structures from the 1001 Arabian Nights to Inspire the Generation of New Arabic Tales},
author = {Mossab Ibrahim and Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez},
year = {2024},
date = {2024-01-01},
journal = {American Journal of Information Science and Computer Engineering},
volume = {5},
number = {4},
pages = {360\textendash370},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Gervás, Pablo; Méndez, Gonzalo
Evolutionary Plot Arcs for a Series of Neurally-Generated Episodes Proceedings Article
In: 15th International Conference on Computational Creativity (ICCC 2024), 2024.
@inproceedings{gervas2024evolutionary,
title = {Evolutionary Plot Arcs for a Series of Neurally-Generated Episodes},
author = {Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez},
url = {https://computationalcreativity.net/iccc24/papers/ICCC24_paper_68.pdf},
year = {2024},
date = {2024-01-01},
urldate = {2024-01-01},
booktitle = {15th International Conference on Computational Creativity (ICCC 2024)},
abstract = {Neural generators produce fluent text for short stories but face difficulties for longer narrative structures. Knowledge-based generators produce coherent narrative structures but have problems generating fluid text. Series built of episodes involve an overarching plot arc for the series and specific plots for each episode. The present paper explores a combination of evolutionary generation of season plot arcs\textendashbased on knowledge resources for plot structures\textendashwith neural generation of episode descriptions\textendashprompted with fragments of season plot arc},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Neural generators produce fluent text for short stories but face difficulties for longer narrative structures. Knowledge-based generators produce coherent narrative structures but have problems generating fluid text. Series built of episodes involve an overarching plot arc for the series and specific plots for each episode. The present paper explores a combination of evolutionary generation of season plot arcs–based on knowledge resources for plot structures–with neural generation of episode descriptions–prompted with fragments of season plot arc
Góngora, Santiago; Chiruzzo, Luis; Méndez, Gonzalo; Gervás, Pablo
PAYADOR: A Minimalist Approach to Grounding Language Models on Structured Data for Interactive Storytelling and Role-playing Games Proceedings Article
In: 15th International Conference on Computational Creativity (ICCC 2024), 2024.
@inproceedings{gongora2024payador,
title = {PAYADOR: A Minimalist Approach to Grounding Language Models on Structured Data for Interactive Storytelling and Role-playing Games},
author = {Santiago G\'{o}ngora and Luis Chiruzzo and Gonzalo M\'{e}ndez and Pablo Gerv\'{a}s},
url = {https://computationalcreativity.net/iccc24/papers/ICCC24_paper_152.pdf},
year = {2024},
date = {2024-01-01},
urldate = {2024-01-01},
booktitle = {15th International Conference on Computational Creativity (ICCC 2024)},
abstract = {Every time an Interactive Storytelling (IS) system gets a player input, it is facing the world-update problem. Classical approaches to this problem consist in mapping that input to known preprogrammed actions, what can severely constrain the free will of the player. When the expected experience has a strong focus on improvisation, like in Role-playing Games (RPGs), this problem is critical. In this paper we present PAYADOR, a different approach that focuses on predicting the outcomes of the actions instead of representing the actions themselves. To implement this approach, we ground a Large Language Model to a minimal representation of the fictional world, obtaining promising results. We make this contribution open-source, so it can be adapted and used for other related research on unleashing the cocreativity power of RPGs.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Every time an Interactive Storytelling (IS) system gets a player input, it is facing the world-update problem. Classical approaches to this problem consist in mapping that input to known preprogrammed actions, what can severely constrain the free will of the player. When the expected experience has a strong focus on improvisation, like in Role-playing Games (RPGs), this problem is critical. In this paper we present PAYADOR, a different approach that focuses on predicting the outcomes of the actions instead of representing the actions themselves. To implement this approach, we ground a Large Language Model to a minimal representation of the fictional world, obtaining promising results. We make this contribution open-source, so it can be adapted and used for other related research on unleashing the cocreativity power of RPGs.
Gervás, Pablo; Méndez, Gonzalo
Distributing Creative Responsibility Between a Knowledge-Based Content Determiner and a Neural Text Realizer Proceedings Article
In: 23rd EPIA Conference on Artificial Intelligence, Thematic Track on AI and Creativity, 2024.
@inproceedings{gervas2024distributing,
title = {Distributing Creative Responsibility Between a Knowledge-Based Content Determiner and a Neural Text Realizer},
author = {Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez},
url = {https://link.springer.com/chapter/10.1007/978-3-031-73497-7_4
https://doi.org/10.1007/978-3-031-73497-7_4},
doi = {10.1007/978-3-031-73497-7_4},
year = {2024},
date = {2024-01-01},
urldate = {2024-01-01},
booktitle = {23rd EPIA Conference on Artificial Intelligence, Thematic Track on AI and Creativity},
abstract = {Knowledge-based solutions for text generation are known to produce outputs that usually sound repetitive and stilted to the human ear. Attention-based neural solutions for generating text have proven to be successful at generating unconstrained prose that is fluent and sounds natural, but they have difficulty in producing texts that comply with a set of restrictions provided as input. The present paper explores combinations of a knowledge-based content generator and a neural text realizer, focusing on how creative responsibility over the final output is distributed over the knowledge-based and the neural modules of the system. A conceptual draft for a story is produced by a knowledge-based solution. The stories are then told using neural generators, with different types of prompt being built as means of requesting specific ways of telling the selected events. The outcomes are evaluated in terms of the percentage of the ideas in the final text that have been contributed by each module.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Knowledge-based solutions for text generation are known to produce outputs that usually sound repetitive and stilted to the human ear. Attention-based neural solutions for generating text have proven to be successful at generating unconstrained prose that is fluent and sounds natural, but they have difficulty in producing texts that comply with a set of restrictions provided as input. The present paper explores combinations of a knowledge-based content generator and a neural text realizer, focusing on how creative responsibility over the final output is distributed over the knowledge-based and the neural modules of the system. A conceptual draft for a story is produced by a knowledge-based solution. The stories are then told using neural generators, with different types of prompt being built as means of requesting specific ways of telling the selected events. The outcomes are evaluated in terms of the percentage of the ideas in the final text that have been contributed by each module.
Gervás, Pablo; León, Carlos; Kumar, Mayuresh; Méndez, Gonzalo; Bautista, Susana
Configuring an LLM Chatbot as Practice Partner for Language Learning Proceedings Article
In: Advances in Artificial Intelligence–IBERAMIA 2024: 18th Ibero-American Conference on AI, Montevideo, Uruguay, November 13–15, 2024, Proceedings, pp. 458, Springer Nature 2024, ISBN: 978-3-031-80365-9.
@inproceedings{gervas2024configuring,
title = {Configuring an LLM Chatbot as Practice Partner for Language Learning},
author = {Pablo Gerv\'{a}s and Carlos Le\'{o}n and Mayuresh Kumar and Gonzalo M\'{e}ndez and Susana Bautista},
url = {https://link.springer.com/book/10.1007/978-3-031-80366-6
https://doi.org/10.1007/978-3-031-80366-6},
doi = {10.1007/978-3-031-80366-6},
isbn = {978-3-031-80365-9},
year = {2024},
date = {2024-01-01},
urldate = {2024-01-01},
booktitle = {Advances in Artificial Intelligence\textendashIBERAMIA 2024: 18th Ibero-American Conference on AI, Montevideo, Uruguay, November 13\textendash15, 2024, Proceedings},
volume = {15277},
pages = {458},
organization = {Springer Nature},
abstract = {Given appropriate prompts, chatbots based on Large Language Models (LLMs) can adapt both the content and the role in which they participate in the conversation. The present paper reports an experiment to configure one such chatbot to engage in conversation with a language learner so that it can propose engaging situations of the appropriate level of complexity, enact specific roles in the conversation and monitor learner responses. The proposed functionality is tested for a classroom of Hindi-speaking students of Spanish language at Aligarh Muslim University in India},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Given appropriate prompts, chatbots based on Large Language Models (LLMs) can adapt both the content and the role in which they participate in the conversation. The present paper reports an experiment to configure one such chatbot to engage in conversation with a language learner so that it can propose engaging situations of the appropriate level of complexity, enact specific roles in the conversation and monitor learner responses. The proposed functionality is tested for a classroom of Hindi-speaking students of Spanish language at Aligarh Muslim University in India
León, Carlos; Villar, Alejandro; Gervás, Pablo; Méndez, Gonzalo
Realidad virtual para entrenamiento en emergencias radiológicas Journal Article
In: Alfa, vol. 57, pp. 52–53, 2024.
@article{leon2024realidad,
title = {Realidad virtual para entrenamiento en emergencias radiol\'{o}gicas},
author = {Carlos Le\'{o}n and Alejandro Villar and Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez},
year = {2024},
date = {2024-01-01},
urldate = {2024-01-01},
journal = {Alfa},
volume = {57},
pages = {52\textendash53},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Gervás, Pablo; López-Calle, Jose Luis
Representing Complex Relative Chronology Across Narrative Levels in Movie Plots. Proceedings Article
In: Text2Story@ ECIR, pp. 65–76, 2024.
@inproceedings{gervas2024representing,
title = {Representing Complex Relative Chronology Across Narrative Levels in Movie Plots.},
author = {Pablo Gerv\'{a}s and Jose Luis L\'{o}pez-Calle},
url = {https://ceur-ws.org/Vol-3671/paper6.pdf},
year = {2024},
date = {2024-01-01},
urldate = {2024-01-01},
booktitle = {Text2Story@ ECIR},
pages = {65\textendash76},
abstract = {Movie narratives over the last few decades have increasingly featured very complex chronologies. These include a number of phenomena ranging from relatively simple use of ashbacks to complete alteration of the chronological order of events (Pulp Fiction) that sometimes force radical revision of the interpretation of the narrative (Memento). Computational treatment of this type of tropes requires a detailed understanding that will require temporal representation of the chronology of the narratives for these movies. The present paper explores the set of challenges that arise when the discourse for a given narrative departs from traditional ordered chronology, and proposes a model for representing the complexities involved. Narratives often convey at dierent points of its span a story -or more - inside the story that is unfolding. That embedded story might account for events already happened in the story being told, or be a dierent story, whether ctional or not, and belong to the world in which the story takes place, or to another world or time. In the same fashion that the embedded story could account for a past or present event, they can narrate a future event, that may or not actually happen. But that is precisely what occurs when we describe a plan, a wish, a promise, something likely to happen, or things that are or were conditioned to past, present or future events. . . All in all, the events mentioned in them are also told, and despite some may be expressed as a chance past and lost, it might still become true; others have as yet not happened, and may still come to be true, or not take place at all. The present paper proposes a representation of events in a story that captures the relevant features of these aspects. },
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Movie narratives over the last few decades have increasingly featured very complex chronologies. These include a number of phenomena ranging from relatively simple use of ashbacks to complete alteration of the chronological order of events (Pulp Fiction) that sometimes force radical revision of the interpretation of the narrative (Memento). Computational treatment of this type of tropes requires a detailed understanding that will require temporal representation of the chronology of the narratives for these movies. The present paper explores the set of challenges that arise when the discourse for a given narrative departs from traditional ordered chronology, and proposes a model for representing the complexities involved. Narratives often convey at dierent points of its span a story -or more - inside the story that is unfolding. That embedded story might account for events already happened in the story being told, or be a dierent story, whether ctional or not, and belong to the world in which the story takes place, or to another world or time. In the same fashion that the embedded story could account for a past or present event, they can narrate a future event, that may or not actually happen. But that is precisely what occurs when we describe a plan, a wish, a promise, something likely to happen, or things that are or were conditioned to past, present or future events. . . All in all, the events mentioned in them are also told, and despite some may be expressed as a chance past and lost, it might still become true; others have as yet not happened, and may still come to be true, or not take place at all. The present paper proposes a representation of events in a story that captures the relevant features of these aspects.
Gervás, Pablo
Challenges and opportunities for computational construction of narratives Journal Article
In: Digital Society Volume 64, pp. 111, 2024.
@article{gervas2024challenges,
title = {Challenges and opportunities for computational construction of narratives},
author = {Pablo Gerv\'{a}s},
url = {https://www.degruyterbrill.com/document/doi/10.1515/9783839469224-007/pdf?licenseType=open-access},
doi = {10.1515/9783839469224-007},
year = {2024},
date = {2024-01-01},
urldate = {2024-01-01},
journal = {Digital Society Volume 64},
pages = {111},
abstract = {As Artificial Intelligence enters a new era where large language models show surprising capabilities for handling language, it becomes important to review the challenges and insights accumulated by the community of esearches that have worked on building computational modelling of literary creativity over the years.The efforts of this community have at times achieved success at small tasks and at other times faced failure when over-ambitious goals were pursued. It would be important for efforts on related tasks undertaken from this point on to keep in mind the insights accumulated on the nature of the task and the challenges it presents. The present paper attempts to summarize some of these insights into a set of challenges that face the automated generation of narratives, but also as a set of opportunities open for the future.
The paper is structured as a review of a number of possible ways to formulate the problem of having a program generate automatically a story on demand, followed by some reflections on how storytelling might be subdivided into a set of related subtasks as suggested by that set of formulations of the problem.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
As Artificial Intelligence enters a new era where large language models show surprising capabilities for handling language, it becomes important to review the challenges and insights accumulated by the community of esearches that have worked on building computational modelling of literary creativity over the years.The efforts of this community have at times achieved success at small tasks and at other times faced failure when over-ambitious goals were pursued. It would be important for efforts on related tasks undertaken from this point on to keep in mind the insights accumulated on the nature of the task and the challenges it presents. The present paper attempts to summarize some of these insights into a set of challenges that face the automated generation of narratives, but also as a set of opportunities open for the future.
The paper is structured as a review of a number of possible ways to formulate the problem of having a program generate automatically a story on demand, followed by some reflections on how storytelling might be subdivided into a set of related subtasks as suggested by that set of formulations of the problem.
The paper is structured as a review of a number of possible ways to formulate the problem of having a program generate automatically a story on demand, followed by some reflections on how storytelling might be subdivided into a set of related subtasks as suggested by that set of formulations of the problem.
Gervás, Pablo
Storytelling systems: Progress towards generation informed by models of the reader Journal Article
In: Interdisciplinary Science Reviews, vol. 49, no. 2, pp. 269–290, 2024.
@article{gervas2024storytelling,
title = {Storytelling systems: Progress towards generation informed by models of the reader},
author = {Pablo Gerv\'{a}s},
url = {https://journals.sagepub.com/doi/10.1177/03080188241262862},
doi = {10.1177/03080188241262862},
year = {2024},
date = {2024-01-01},
urldate = {2024-01-01},
journal = {Interdisciplinary Science Reviews},
volume = {49},
number = {2},
pages = {269\textendash290},
publisher = {SAGE Publications Sage UK: London, England},
abstract = {The present paper attempts to identify specific points of contact between efforts to generate narrative in the field of artificial intelligence and the field of narratology. Such points of contact centre around the tasks of modelling processes related to the construction of narrative in computational terms. A number of existing research efforts to develop systems related to storytelling are reviewed. An analysis of the functionalities that they cover suggests that storytelling might be profitably studied in terms of at least four different sub-tasks: Fabula generation, narrative discourse composition, text generation, and narrative interpretation. An analysis from a computational point of view of the context in which narrative is used as a means of communication between author and reader is presented. This shows that parts of the cognitive processes involved are yet to be modelled, which suggests that the set of sub-tasks identified may continue to grow with further research.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The present paper attempts to identify specific points of contact between efforts to generate narrative in the field of artificial intelligence and the field of narratology. Such points of contact centre around the tasks of modelling processes related to the construction of narrative in computational terms. A number of existing research efforts to develop systems related to storytelling are reviewed. An analysis of the functionalities that they cover suggests that storytelling might be profitably studied in terms of at least four different sub-tasks: Fabula generation, narrative discourse composition, text generation, and narrative interpretation. An analysis from a computational point of view of the context in which narrative is used as a means of communication between author and reader is presented. This shows that parts of the cognitive processes involved are yet to be modelled, which suggests that the set of sub-tasks identified may continue to grow with further research.
2023
Gervás, Pablo; Méndez, Gonzalo; Concepción, Eugenio
Evolutionary combination of connected event schemas into meaningful plots Journal Article
In: Genetic Programming and Evolvable Machines, vol. 24, iss. 7, no. 1, pp. 38, 2023, ISSN: 1573-7632.
@article{gervas2023evolutionary,
title = {Evolutionary combination of connected event schemas into meaningful plots},
author = {Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez and Eugenio Concepci\'{o}n},
url = {https://link.springer.com/content/pdf/10.1007/s10710-023-09454-2.pdf},
doi = {10.1007/s10710-023-09454-2},
issn = {1573-7632},
year = {2023},
date = {2023-01-01},
urldate = {2023-01-01},
journal = {Genetic Programming and Evolvable Machines},
volume = {24},
number = {1},
issue = {7},
pages = {38},
publisher = {Springer US New York},
abstract = {Many of the stories we are exposed to are built from small schemas of connected events involving a set of characters\textendashboy meets girl leads to a relationship or crime leads to revenge. The present paper proposes an evolutionary solution to the task of putting together a story by combining a set of such schemas. This approach presents three challenges: how to mix up the elements in the different schemas, how to instantiate the characters across the schemas and how to tell acceptable combinations from the rest. The present paper applies an evolutionary solution that relies on a genetic representation for these combinations of schemas, and applies as fitness functions a set of metrics on compatibility constraints across schema combinations. Outputs of this procedure are evaluated by human judges in comparison with baseline solutions in which the values for genes are assigned at random. The proposed solution generates a population of story drafts that resemble plot descriptions for simple stories. The results of the comparative evaluation by human judges are positive. The genetic representation of pattern combinations and the metrics on compatibility across pattern pairs provide a valid evolutionary solution for constructing simple plots.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Many of the stories we are exposed to are built from small schemas of connected events involving a set of characters–boy meets girl leads to a relationship or crime leads to revenge. The present paper proposes an evolutionary solution to the task of putting together a story by combining a set of such schemas. This approach presents three challenges: how to mix up the elements in the different schemas, how to instantiate the characters across the schemas and how to tell acceptable combinations from the rest. The present paper applies an evolutionary solution that relies on a genetic representation for these combinations of schemas, and applies as fitness functions a set of metrics on compatibility constraints across schema combinations. Outputs of this procedure are evaluated by human judges in comparison with baseline solutions in which the values for genes are assigned at random. The proposed solution generates a population of story drafts that resemble plot descriptions for simple stories. The results of the comparative evaluation by human judges are positive. The genetic representation of pattern combinations and the metrics on compatibility across pattern pairs provide a valid evolutionary solution for constructing simple plots.
Méndez, Gonzalo; Gervás, Pablo
Using ChatGPT for Story Sifting in Narrative Generation Proceedings Article
In: International Conference on Computational Creativity, 2023.
@inproceedings{mendez2023using,
title = {Using ChatGPT for Story Sifting in Narrative Generation},
author = {Gonzalo M\'{e}ndez and Pablo Gerv\'{a}s},
url = {https://computationalcreativity.net/iccc23/papers/ICCC-2023_paper_124.pdf},
year = {2023},
date = {2023-01-01},
urldate = {2023-01-01},
booktitle = {International Conference on Computational Creativity},
abstract = {The task of selecting a subset of story-worthy events from out of an observed collection of facts\textendashknown as story sifting\textendashis a useful human ability that has yet to be emulated successfully by computational processes. The emergence of Large Language Models (LLMs) has made it necessary to rethink the way of carrying out many tasks that were previously performed using other tools. This short paper explores how the infamous ChatGPT fares when asked to sift stories from the log of an agent-based simulation featuring romantic relations between characters.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The task of selecting a subset of story-worthy events from out of an observed collection of facts–known as story sifting–is a useful human ability that has yet to be emulated successfully by computational processes. The emergence of Large Language Models (LLMs) has made it necessary to rethink the way of carrying out many tasks that were previously performed using other tools. This short paper explores how the infamous ChatGPT fares when asked to sift stories from the log of an agent-based simulation featuring romantic relations between characters.
Góngora, Santiago; Chiruzzo, Luis; Méndez, Gonzalo; Gervás, Pablo
Skill check: Some considerations on the evaluation of gamemastering models for role-playing games Proceedings Article
In: International Conference on Games and Learning Alliance, pp. 277–288, Springer Nature Switzerland Cham Springer, 2023, ISBN: 978-3-031-49064-4.
@inproceedings{gongora2023skill,
title = {Skill check: Some considerations on the evaluation of gamemastering models for role-playing games},
author = {Santiago G\'{o}ngora and Luis Chiruzzo and Gonzalo M\'{e}ndez and Pablo Gerv\'{a}s},
url = {https://link.springer.com/chapter/10.1007/978-3-031-49065-1_27},
doi = {10.1007/978-3-031-49065-1_27},
isbn = {978-3-031-49064-4},
year = {2023},
date = {2023-01-01},
urldate = {2023-01-01},
booktitle = {International Conference on Games and Learning Alliance},
volume = {14475},
pages = {277\textendash288},
publisher = {Springer},
organization = {Springer Nature Switzerland Cham},
abstract = {In role-playing games a Game Master (GM) is the player in charge of the game, who must design the challenges the players face and narrate the outcomes of their actions. In this work we discuss some challenges to model GMs from an Interactive Storytelling and Natural Language Processing perspective. Following those challenges we propose three test categories to evaluate such dialogue systems, and we use them to test ChatGPT, Bard and OpenAssistant as out-of-the-box GMs.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
In role-playing games a Game Master (GM) is the player in charge of the game, who must design the challenges the players face and narrate the outcomes of their actions. In this work we discuss some challenges to model GMs from an Interactive Storytelling and Natural Language Processing perspective. Following those challenges we propose three test categories to evaluate such dialogue systems, and we use them to test ChatGPT, Bard and OpenAssistant as out-of-the-box GMs.
Gervás, Pablo; Méndez, Gonzalo
Evolutionary Story Sifting over the Log of a Social Simulation Proceedings Article
In: Italian Workshop on Artificial Life and Evolutionary Computation, pp. 381–393, Springer Nature Switzerland Cham Springer, 2023, ISBN: 978-3-031-57429-0.
@inproceedings{gervas2023evolutionaryb,
title = {Evolutionary Story Sifting over the Log of a Social Simulation},
author = {Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez},
url = {https://link.springer.com/chapter/10.1007/978-3-031-57430-6_29},
doi = {10.1007/978-3-031-57430-6_29},
isbn = {978-3-031-57429-0},
year = {2023},
date = {2023-01-01},
urldate = {2023-01-01},
booktitle = {Italian Workshop on Artificial Life and Evolutionary Computation},
volume = {1977},
pages = {381\textendash393},
publisher = {Springer},
organization = {Springer Nature Switzerland Cham},
series = {Communications in Computer and Information Science},
abstract = {The process of selecting subsets out of a sequence of events on the grounds that told together they constitute an interesting narrative\textendashknown as story sifting\textendashhas become a topic of interest due to its applicability in video games that automatically develop large scale simulations of story worlds. Existing approaches to story sifting operate by matching subsequences of the available events onto patterns of plot considered to be of interest. The present paper proposes a two stage approach that combines a process of matching small strings of events connected by common sense relations\textendashsuch as asking someone on a date and having them accept, or developing an attachment to someone who has given us a present\textendashand an evolutionary search procedure that explores combinations of this type of paired events into longer sequences that constitute small plot lines about romantic entanglement. This procedure is run over the log of a multi-agent system that simulates a set of characters that develop a set of affective affinities between them as a result of social interactions of a romantic nature.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The process of selecting subsets out of a sequence of events on the grounds that told together they constitute an interesting narrative–known as story sifting–has become a topic of interest due to its applicability in video games that automatically develop large scale simulations of story worlds. Existing approaches to story sifting operate by matching subsequences of the available events onto patterns of plot considered to be of interest. The present paper proposes a two stage approach that combines a process of matching small strings of events connected by common sense relations–such as asking someone on a date and having them accept, or developing an attachment to someone who has given us a present–and an evolutionary search procedure that explores combinations of this type of paired events into longer sequences that constitute small plot lines about romantic entanglement. This procedure is run over the log of a multi-agent system that simulates a set of characters that develop a set of affective affinities between them as a result of social interactions of a romantic nature.
Gervás, Pablo
Segmenting Narrative Synopses on Event Reporting Mode based on Heuristics on Constituency Parses. Proceedings Article
In: Text2Story@ ECIR, pp. 73–81, 2023.
@inproceedings{gervas2023segmenting,
title = {Segmenting Narrative Synopses on Event Reporting Mode based on Heuristics on Constituency Parses.},
author = {Pablo Gerv\'{a}s},
year = {2023},
date = {2023-01-01},
booktitle = {Text2Story@ ECIR},
pages = {73\textendash81},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
López-Calle, José Luis; Gervás, Pablo; Muñoz, Cristóbal Fernández
A Methodology for the Annotation of Narrative Structures in Films Journal Article
In: Comunicación & Métodos, vol. 5, no. 1, pp. 91–107, 2023.
@article{calle2023methodology,
title = {A Methodology for the Annotation of Narrative Structures in Films},
author = {Jos\'{e} Luis L\'{o}pez-Calle and Pablo Gerv\'{a}s and Crist\'{o}bal Fern\'{a}ndez Mu\~{n}oz},
year = {2023},
date = {2023-01-01},
urldate = {2023-01-01},
journal = {Comunicaci\'{o}n \& M\'{e}todos},
volume = {5},
number = {1},
pages = {91\textendash107},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Gervás, Pablo
Improving efficiency and coherence in evolutionary story generation Proceedings Article
In: 14th International Conference on Computational Creativity, Waterloo, Ontario, Canada, 2023.
@inproceedings{gervas2023improving,
title = {Improving efficiency and coherence in evolutionary story generation},
author = {Pablo Gerv\'{a}s},
year = {2023},
date = {2023-01-01},
booktitle = {14th International Conference on Computational Creativity, Waterloo, Ontario, Canada},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Romero-Hernandez, Alejandro; Sagredo-Olivenza, Ismael; Gervás, Pablo; Manero, Borja
Do Gaming Habits or Gender Influence the Interest of School-Age Children in Spanish Dance after Playing a Video Game? Journal Article
In: Available at SSRN 4524900, 2023.
@article{romero2023gaming,
title = {Do Gaming Habits or Gender Influence the Interest of School-Age Children in Spanish Dance after Playing a Video Game?},
author = {Alejandro Romero-Hernandez and Ismael Sagredo-Olivenza and Pablo Gerv\'{a}s and Borja Manero},
year = {2023},
date = {2023-01-01},
journal = {Available at SSRN 4524900},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
López-Calle, Jose Luis; Gervás, Pablo; others,
Una metodología para la anotación de estructuras narrativas en películas Journal Article
In: Comunicación & métodos, vol. 5, no. 1, pp. 91–107, 2023.
@article{lopez2023metodologia,
title = {Una metodolog\'{i}a para la anotaci\'{o}n de estructuras narrativas en pel\'{i}culas},
author = {Jose Luis L\'{o}pez-Calle and Pablo Gerv\'{a}s and others},
year = {2023},
date = {2023-01-01},
urldate = {2023-01-01},
journal = {Comunicaci\'{o}n \& m\'{e}todos},
volume = {5},
number = {1},
pages = {91\textendash107},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
2022
Gervás, Pablo
A Discourse Interpretation Engine Sensitive to Truth Revisions in a Story Proceedings Article
In: Tenth Annual Conference on Advances in Cognitive Systems, Arlington, Virginia, 2022.
@inproceedings{798,
title = {A Discourse Interpretation Engine Sensitive to Truth Revisions in a Story},
author = {Pablo Gerv\'{a}s},
year = {2022},
date = {2022-11-01},
urldate = {2022-11-01},
booktitle = {Tenth Annual Conference on Advances in Cognitive Systems},
address = {Arlington, Virginia},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Gervás, Pablo; Torrente, A.
Emotional Interpretation of Opera Seria: Impact of Specifics of Drama Structure Proceedings Article
In: 14th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management, 2022.
@inproceedings{796,
title = {Emotional Interpretation of Opera Seria: Impact of Specifics of Drama Structure},
author = {Pablo Gerv\'{a}s and A. Torrente},
year = {2022},
date = {2022-10-01},
urldate = {2022-10-01},
booktitle = {14th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Gervás, Pablo
Evolutionary Stitching of Plot Units with Character Threads Proceedings Article
In: WIVACE 2022 XVI International Workshop on Artificial Life and Evolutionary Computation, 2022.
@inproceedings{794,
title = {Evolutionary Stitching of Plot Units with Character Threads},
author = {Pablo Gerv\'{a}s},
year = {2022},
date = {2022-09-01},
urldate = {2022-09-01},
booktitle = {WIVACE 2022 XVI International Workshop on Artificial Life and Evolutionary Computation},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Gervás, Pablo
Generating Plotlines about Attempting to Avert Disasters Proceedings Article
In: The Thirteenth International Conference on Computational Creativity, 2022.
@inproceedings{778,
title = {Generating Plotlines about Attempting to Avert Disasters},
author = {Pablo Gerv\'{a}s},
year = {2022},
date = {2022-06-01},
urldate = {2022-06-01},
booktitle = {The Thirteenth International Conference on Computational Creativity},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Colton, Simon; Llano, Maria Teresa; Hepworth, Rose; Charnley, John; Gale, Catherine V; Baron, Archie; Pachet, François; Roy, Pierre; Gervás, Pablo; Collins, Nick; others,
The beyond the fence musical and computer says show documentary Journal Article
In: arXiv preprint arXiv:2206.03224, 2022.
@article{colton2022beyond,
title = {The beyond the fence musical and computer says show documentary},
author = {Simon Colton and Maria Teresa Llano and Rose Hepworth and John Charnley and Catherine V Gale and Archie Baron and Fran\c{c}ois Pachet and Pierre Roy and Pablo Gerv\'{a}s and Nick Collins and others},
year = {2022},
date = {2022-01-01},
journal = {arXiv preprint arXiv:2206.03224},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Gervás, Pablo; Concepción, Eugenio; Méndez, Gonzalo
Evolutionary construction of stories that combine several plot lines Proceedings Article
In: International Conference on Computational Intelligence in Music, Sound, Art and Design (Part of EvoStar), pp. 68–83, Springer International Publishing Cham Springer, 2022, ISBN: 978-3-031-03788-7.
@inproceedings{gervas2022evolutionary,
title = {Evolutionary construction of stories that combine several plot lines},
author = {Pablo Gerv\'{a}s and Eugenio Concepci\'{o}n and Gonzalo M\'{e}ndez},
url = {https://link.springer.com/chapter/10.1007/978-3-031-03789-4_5},
doi = {10.1007/978-3-031-03789-4_5},
isbn = {978-3-031-03788-7},
year = {2022},
date = {2022-01-01},
urldate = {2022-01-01},
booktitle = {International Conference on Computational Intelligence in Music, Sound, Art and Design (Part of EvoStar)},
volume = {13221},
pages = {68\textendash83},
publisher = {Springer},
organization = {Springer International Publishing Cham},
series = {Lecture Notes in Computer Science},
abstract = {Although the narrative structure of common entertainment products like Hollywood movies or TV series is generally composed of a number of different plot lines combined into a single narrative discourse, efforts on computational modeling of story generation have to this point focused mostly on the construction of stories with a single plot line. The present paper explores an evolutionary solution to the task of building a story that combines more than one plot line into a single linear discourse. This requires: a set of knowledge resources that capture the main features that influence the decisions involved, a representation suitable for evolutionary treatment for discourses with several plot lines, and a set of fitness functions based on metrics related to the quality of the resulting discourses. The proposed solution produces populations of stories with elaborate discourses that combine several subplots.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Although the narrative structure of common entertainment products like Hollywood movies or TV series is generally composed of a number of different plot lines combined into a single narrative discourse, efforts on computational modeling of story generation have to this point focused mostly on the construction of stories with a single plot line. The present paper explores an evolutionary solution to the task of building a story that combines more than one plot line into a single linear discourse. This requires: a set of knowledge resources that capture the main features that influence the decisions involved, a representation suitable for evolutionary treatment for discourses with several plot lines, and a set of fitness functions based on metrics related to the quality of the resulting discourses. The proposed solution produces populations of stories with elaborate discourses that combine several subplots.
Gervás, Pablo
Generating Plotlines about Attempting to Avert Disasters Paper type: Short Paper Proceedings Article
In: Proceedings of the 13th International Conference on Computational Creativity, ICCC 2022, pp. 67–71, 2022.
@inproceedings{gervas2022generating,
title = {Generating Plotlines about Attempting to Avert Disasters Paper type: Short Paper},
author = {Pablo Gerv\'{a}s},
year = {2022},
date = {2022-01-01},
booktitle = {Proceedings of the 13th International Conference on Computational Creativity, ICCC 2022},
pages = {67\textendash71},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Gervás, Pablo; Méndez, Gonzalo; Concepción, Eugenio
Evolutionary Combination of Subplot Patterns into Meaningful Plots Online
2022, visited: 01.01.2022.
@online{gervas2022evolutionaryb,
title = {Evolutionary Combination of Subplot Patterns into Meaningful Plots},
author = {Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez and Eugenio Concepci\'{o}n},
url = {https://www.researchsquare.com/article/rs-2169837/v1
https://www.researchsquare.com/article/rs-2169837/v1.pdf?c=1666919495000},
doi = {https://doi.org/10.21203/rs.3.rs-2169837/v1 },
year = {2022},
date = {2022-01-01},
urldate = {2022-01-01},
abstract = {Purpose: Many of the stories we are exposed to are built from small patterns of connected events involving a set of characters -- boy meets girl leads to a relationship or crime leads to revenge. The present paper studies the computational constraints that apply to the task of putting together a story by combining a set of such patterns. This approach presents three challenges: how to mix up the elements in the different patterns,how to instantiate the characters across the patterns andhow to tell acceptable combinations from the rest.
Methods: The present paper applies an evolutionary solution that relies on a genetic representation for these combinations of patterns, and applies as fitness functions a set of metrics on compatibility constraints across pattern combinations. Outputs of this procedure are evaluated by human judges in comparison with baseline solutions.
Results: The proposed solution generates a population of story drafts that resemble plot descriptions for simple stories. A comparative evaluation by human judges against baselines based on random gene assignment yields positive results.
Conclusion: The genetic representation of pattern combinations and the metrics on compatibility across pattern pairs provide a valid evolutionary solution for constructing simple plots.},
howpublished = {Preprint from Research Square},
keywords = {},
pubstate = {published},
tppubtype = {online}
}
Purpose: Many of the stories we are exposed to are built from small patterns of connected events involving a set of characters -- boy meets girl leads to a relationship or crime leads to revenge. The present paper studies the computational constraints that apply to the task of putting together a story by combining a set of such patterns. This approach presents three challenges: how to mix up the elements in the different patterns,how to instantiate the characters across the patterns andhow to tell acceptable combinations from the rest.
Methods: The present paper applies an evolutionary solution that relies on a genetic representation for these combinations of patterns, and applies as fitness functions a set of metrics on compatibility constraints across pattern combinations. Outputs of this procedure are evaluated by human judges in comparison with baseline solutions.
Results: The proposed solution generates a population of story drafts that resemble plot descriptions for simple stories. A comparative evaluation by human judges against baselines based on random gene assignment yields positive results.
Conclusion: The genetic representation of pattern combinations and the metrics on compatibility across pattern pairs provide a valid evolutionary solution for constructing simple plots.
Methods: The present paper applies an evolutionary solution that relies on a genetic representation for these combinations of patterns, and applies as fitness functions a set of metrics on compatibility constraints across pattern combinations. Outputs of this procedure are evaluated by human judges in comparison with baseline solutions.
Results: The proposed solution generates a population of story drafts that resemble plot descriptions for simple stories. A comparative evaluation by human judges against baselines based on random gene assignment yields positive results.
Conclusion: The genetic representation of pattern combinations and the metrics on compatibility across pattern pairs provide a valid evolutionary solution for constructing simple plots.
Gervás, Pablo; Torrente, Alvaro
Emotional Interpretation of Opera Seria: Impact of Specifics of Drama Structure (Position Paper). Proceedings Article
In: KDIR, pp. 330–336, 2022.
@inproceedings{gervas2022emotional,
title = {Emotional Interpretation of Opera Seria: Impact of Specifics of Drama Structure (Position Paper).},
author = {Pablo Gerv\'{a}s and Alvaro Torrente},
year = {2022},
date = {2022-01-01},
booktitle = {KDIR},
pages = {330\textendash336},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Gervás, Pablo
N-Gram-Driven Word Level Recombination: Exploring a Search Space of Metrically Valid Verse Journal Article
In: Stylistics in Poetry, Prose, and Drama, pp. 203, 2022.
@article{gervas2022n,
title = {N-Gram-Driven Word Level Recombination: Exploring a Search Space of Metrically Valid Verse},
author = {Pablo Gerv\'{a}s},
year = {2022},
date = {2022-01-01},
journal = {Stylistics in Poetry, Prose, and Drama},
pages = {203},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Gervás, Pablo
Evolutionary stitching of plot units with character threads Proceedings Article
In: Italian Workshop on Artificial Life and Evolutionary Computation, pp. 254–265, Springer Nature Switzerland Cham 2022.
@inproceedings{gervas2022evolutionaryc,
title = {Evolutionary stitching of plot units with character threads},
author = {Pablo Gerv\'{a}s},
year = {2022},
date = {2022-01-01},
booktitle = {Italian Workshop on Artificial Life and Evolutionary Computation},
pages = {254\textendash265},
organization = {Springer Nature Switzerland Cham},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Gervás, Pablo
Exploring a Search Space of Metrically Journal Article
In: Computational Stylistics in Poetry, Prose, and Drama, pp. 203, 2022.
@article{gervas2022exploring,
title = {Exploring a Search Space of Metrically},
author = {Pablo Gerv\'{a}s},
year = {2022},
date = {2022-01-01},
journal = {Computational Stylistics in Poetry, Prose, and Drama},
pages = {203},
publisher = {Walter de Gruyter GmbH \& Co KG},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
2021
Gervás, Pablo; Concepción, Eugenio; Méndez, Gonzalo
Assessing MultiPlot Stories: from Formative Analysis to Computational Metrics Proceedings Article
In: International Conference on Computational Creativity, Ciudad de México, México, 2021.
@inproceedings{765,
title = {Assessing MultiPlot Stories: from Formative Analysis to Computational Metrics},
author = {Pablo Gerv\'{a}s and Eugenio Concepci\'{o}n and Gonzalo M\'{e}ndez},
url = {https://computationalcreativity.net/iccc21/wp-content/uploads/2021/09/ICCC_2021_paper_113.pdf},
year = {2021},
date = {2021-09-01},
urldate = {2021-09-01},
booktitle = {International Conference on Computational Creativity},
address = {Ciudad de M\'{e}xico, M\'{e}xico},
abstract = {Recent interest in story generators capable of combining more than on plot line into an elaborate story have been handicapped by the lack of either theoretical material or quantitative metrics to ascertain the quality of outputs of such attempts. The present short paper postulates a set of metrics designed to capture some of the insights elaborated during a formative evaluation of an existing attempt at plot weaving.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Recent interest in story generators capable of combining more than on plot line into an elaborate story have been handicapped by the lack of either theoretical material or quantitative metrics to ascertain the quality of outputs of such attempts. The present short paper postulates a set of metrics designed to capture some of the insights elaborated during a formative evaluation of an existing attempt at plot weaving.
Gervás, Pablo
A Model of Interpretation of Embedded Stories Proceedings Article
In: Text2Story: 4th International Workshop on Narrative Extraction from Texts, CEUR Workshop Proceedings CEUR Workshop Proceedings, Lucca, Tuscany, 2021.
@inproceedings{761,
title = {A Model of Interpretation of Embedded Stories},
author = {Pablo Gerv\'{a}s},
year = {2021},
date = {2021-04-01},
urldate = {2021-04-01},
booktitle = {Text2Story: 4th International Workshop on Narrative Extraction from Texts},
publisher = {CEUR Workshop Proceedings},
address = {Lucca, Tuscany},
organization = {CEUR Workshop Proceedings},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Gervás, Pablo
Computational Models of Narrative Creativity Book Chapter
In: Machado, Penousal; Romero, Juan; Greenfield, Gary (Ed.): Artificial Intelligence and the Arts: Computational Creativity, Artistic Behavior, and Tools for Creatives, Chapter 9, pp. 209–255, Springer International Publishing, Cham, 2021, ISBN: 978-3-030-59475-6.
@inbook{772,
title = {Computational Models of Narrative Creativity},
author = {Pablo Gerv\'{a}s},
editor = {Penousal Machado and Juan Romero and Gary Greenfield},
url = {https://doi.org/10.1007/978-3-030-59475-6_9},
doi = {10.1007/978-3-030-59475-6_9},
isbn = {978-3-030-59475-6},
year = {2021},
date = {2021-01-01},
booktitle = {Artificial Intelligence and the Arts: Computational Creativity, Artistic Behavior, and Tools for Creatives},
pages = {209\textendash255},
publisher = {Springer International Publishing},
address = {Cham},
chapter = {9},
organization = {Springer International Publishing},
abstract = {Ever since the advent of computers there have been research efforts
to emulate computationally the way in which people can create stories. Artificial
intelligence (AI) and the myriad of specific technologies it brought with it gave rise to
more elaborate efforts. Yet the task seems to be a challenging frontier that even today
still defeats efforts to achieve performance comparable to that of humans. In part, the
difficulty lies in the fact that the approaches computers have been following to date
differ significantly from those applied by humans on a number of points. Whereas
computers generally apply a single specific AI technique to these problems, humans
seem to combine several cognitive abilities into the task. Computers focus on the
short end of narrative production (fairy tales and short stories) but humans produce
in a much wider range, with long examples of much greater complexity (novels, plays,
films) being very frequent. Where computers usually apply a one-pass algorithm that
produces a single output or focus on generating small fragments of an interactive
piece in co-operation with a human, the basic mode of operation for human authors
is to work alone and to proceed by iteratively revising a draft based on critical
appraisal of whether it meets very demanding criteria. The present chapter reviews
the existing approaches to computational generation of (emulations of) narrative
under the light of these differences in procedure. To inform this review, some of the
existing models of how humans address the corresponding tasks are described and
the existing computational solutions are examined for similarities and differences
with these models of human performance. The impact of these various aspects on
the quality of computer generated artifacts is considered and an analysis of current
trends for future development is made, with particular emphasis on how addressing
the aspects that have so far been less studied may help change the landscape of the
field.},
keywords = {},
pubstate = {published},
tppubtype = {inbook}
}
Ever since the advent of computers there have been research efforts
to emulate computationally the way in which people can create stories. Artificial
intelligence (AI) and the myriad of specific technologies it brought with it gave rise to
more elaborate efforts. Yet the task seems to be a challenging frontier that even today
still defeats efforts to achieve performance comparable to that of humans. In part, the
difficulty lies in the fact that the approaches computers have been following to date
differ significantly from those applied by humans on a number of points. Whereas
computers generally apply a single specific AI technique to these problems, humans
seem to combine several cognitive abilities into the task. Computers focus on the
short end of narrative production (fairy tales and short stories) but humans produce
in a much wider range, with long examples of much greater complexity (novels, plays,
films) being very frequent. Where computers usually apply a one-pass algorithm that
produces a single output or focus on generating small fragments of an interactive
piece in co-operation with a human, the basic mode of operation for human authors
is to work alone and to proceed by iteratively revising a draft based on critical
appraisal of whether it meets very demanding criteria. The present chapter reviews
the existing approaches to computational generation of (emulations of) narrative
under the light of these differences in procedure. To inform this review, some of the
existing models of how humans address the corresponding tasks are described and
the existing computational solutions are examined for similarities and differences
with these models of human performance. The impact of these various aspects on
the quality of computer generated artifacts is considered and an analysis of current
trends for future development is made, with particular emphasis on how addressing
the aspects that have so far been less studied may help change the landscape of the
field.
to emulate computationally the way in which people can create stories. Artificial
intelligence (AI) and the myriad of specific technologies it brought with it gave rise to
more elaborate efforts. Yet the task seems to be a challenging frontier that even today
still defeats efforts to achieve performance comparable to that of humans. In part, the
difficulty lies in the fact that the approaches computers have been following to date
differ significantly from those applied by humans on a number of points. Whereas
computers generally apply a single specific AI technique to these problems, humans
seem to combine several cognitive abilities into the task. Computers focus on the
short end of narrative production (fairy tales and short stories) but humans produce
in a much wider range, with long examples of much greater complexity (novels, plays,
films) being very frequent. Where computers usually apply a one-pass algorithm that
produces a single output or focus on generating small fragments of an interactive
piece in co-operation with a human, the basic mode of operation for human authors
is to work alone and to proceed by iteratively revising a draft based on critical
appraisal of whether it meets very demanding criteria. The present chapter reviews
the existing approaches to computational generation of (emulations of) narrative
under the light of these differences in procedure. To inform this review, some of the
existing models of how humans address the corresponding tasks are described and
the existing computational solutions are examined for similarities and differences
with these models of human performance. The impact of these various aspects on
the quality of computer generated artifacts is considered and an analysis of current
trends for future development is made, with particular emphasis on how addressing
the aspects that have so far been less studied may help change the landscape of the
field.
2020
Concepción, Eugenio; Gervás, Pablo; Méndez, Gonzalo
Exploring Baselines for Combining Full Plots into Multiple-plot Stories Journal Article
In: New Generation Computing, pp. 1-41, 2020, ISSN: 1882-7055.
@article{758,
title = {Exploring Baselines for Combining Full Plots into Multiple-plot Stories},
author = {Eugenio Concepci\'{o}n and Pablo Gerv\'{a}s and Gonzalo M\'{e}ndez},
url = {https://link.springer.com/article/10.1007%2Fs00354-020-00115-x},
doi = {10.1007/s00354-020-00115-x},
issn = {1882-7055},
year = {2020},
date = {2020-10-01},
journal = {New Generation Computing},
pages = {1-41},
chapter = {1},
abstract = {Many of the stories at the core of narrative entertainment involve a number of plot lines that combine to give them interest.
The present paper sets out to solve the problem of how several different plot lines, each one of them complete in its own sense, can be combined into a single linear sequence that works reasonably well as a plot. Starting from a brief review of how existing storytelling systems address the task, a representation for plots and plot templates is proposed that allows the combination of several subplots into a single plot line. Four strategies for weaving plots are proposed, two taken from literary studies and two computational baselines, and a formative evaluation of a set of stories produced by these solutions is presented. Finally, open issues, promising avenues of future work and the relation to previous work are discussed.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Many of the stories at the core of narrative entertainment involve a number of plot lines that combine to give them interest.
The present paper sets out to solve the problem of how several different plot lines, each one of them complete in its own sense, can be combined into a single linear sequence that works reasonably well as a plot. Starting from a brief review of how existing storytelling systems address the task, a representation for plots and plot templates is proposed that allows the combination of several subplots into a single plot line. Four strategies for weaving plots are proposed, two taken from literary studies and two computational baselines, and a formative evaluation of a set of stories produced by these solutions is presented. Finally, open issues, promising avenues of future work and the relation to previous work are discussed.
The present paper sets out to solve the problem of how several different plot lines, each one of them complete in its own sense, can be combined into a single linear sequence that works reasonably well as a plot. Starting from a brief review of how existing storytelling systems address the task, a representation for plots and plot templates is proposed that allows the combination of several subplots into a single plot line. Four strategies for weaving plots are proposed, two taken from literary studies and two computational baselines, and a formative evaluation of a set of stories produced by these solutions is presented. Finally, open issues, promising avenues of future work and the relation to previous work are discussed.
León, Carlos; Gervás, Pablo; Delatorre, Pablo; Tapscott, Alan
Quantitative Characteristics of Human-Written Short Stories as a Metric for Automated Storytelling Journal Article
In: New Generation Computing, vol. 38, 2020.
@article{769,
title = {Quantitative Characteristics of Human-Written Short Stories as a Metric for Automated Storytelling},
author = {Carlos Le\'{o}n and Pablo Gerv\'{a}s and Pablo Delatorre and Alan Tapscott},
year = {2020},
date = {2020-10-01},
journal = {New Generation Computing},
volume = {38},
chapter = {635},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Hervás, Raquel; Bautista, Susana; Méndez, Gonzalo; Galván, Paloma; Gervás, Pablo
Predictive Composition of Pictogram Messages for Users with Autism Journal Article
In: Journal of Ambient Intelligence and Humanized Computing, 2020.
@article{756,
title = {Predictive Composition of Pictogram Messages for Users with Autism},
author = {Raquel Herv\'{a}s and Susana Bautista and Gonzalo M\'{e}ndez and Paloma Galv\'{a}n and Pablo Gerv\'{a}s},
doi = {10.1007/s12652-020-01925-z},
year = {2020},
date = {2020-01-01},
journal = {Journal of Ambient Intelligence and Humanized Computing},
abstract = {Communication is a basic need for every person. However, there are many people who present disabilities that prevent communication through natural language. Augmentative and Alternative Communication (AAC) systems, including those based on pictograms, attempt to facilitate the communication for people with this kind of difficulties. In this paper we present PictoEditor, an Augmentative and Alternative Communication application for the composition of pictogram messages for users with autism that incorporates prediction functionalities. Although such functionalities have been widely studied in text-based Augmentative and Alternative Communication tools, they have not been applied to pictogram based ones. The results show that prediction based on frequency of use of specific pictograms improves the immediate availability of the desired pictograms, but the improvement with prediction based on sequencing of pseudo-syntactic types of pictogram is not as clear.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Communication is a basic need for every person. However, there are many people who present disabilities that prevent communication through natural language. Augmentative and Alternative Communication (AAC) systems, including those based on pictograms, attempt to facilitate the communication for people with this kind of difficulties. In this paper we present PictoEditor, an Augmentative and Alternative Communication application for the composition of pictogram messages for users with autism that incorporates prediction functionalities. Although such functionalities have been widely studied in text-based Augmentative and Alternative Communication tools, they have not been applied to pictogram based ones. The results show that prediction based on frequency of use of specific pictograms improves the immediate availability of the desired pictograms, but the improvement with prediction based on sequencing of pseudo-syntactic types of pictogram is not as clear.
2019
Gervás, Pablo
Generating a Search Space of Acceptable Narrative Plots Proceedings Article
In: 10th International Conference on Computational Creativity (ICCC 2019), UNC Charlotte, North Carolina, USA, 2019.
@inproceedings{753,
title = {Generating a Search Space of Acceptable Narrative Plots},
author = {Pablo Gerv\'{a}s},
year = {2019},
date = {2019-07-01},
urldate = {2019-07-01},
booktitle = {10th International Conference on Computational Creativity (ICCC 2019)},
address = {UNC Charlotte, North Carolina, USA},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}