Resumen
With the exorbitant amount of content continuously generated by today’s technological society, to which one has access almost instantaneously, a person can spend excessive time searching and filtering content to perform a task. Therefore, being able to automatically identify, extract, and condense the most relevant information from any document is vital. In Natural Language Processing (NLP) this problem is addressed by the automatic summary generation or summarization field. Advances in the design of neural network architectures, most notably the Transformers architecture, have significantly increased the capabilities of language models. These improvements have enabled language models to generate more coherent texts. Therefore, in the automatic summarization field, the scientific community has begun to create summarization models that adopt an abstractive strategy. With the abstractive strategy, we obtain summaries where the summary is written by identifying the most relevant information of the document and reinterpreting the text of the original document to obtain a coherent, informative, and concise summary as possible. Therefore, the abstractive summarization strategy brings the models closer to how human beings summarize content. This work delves into different challenges within the abstractive strategy of summarization, focusing our efforts on the composition or writing process of summaries. The thesis aims to define in more detail what characterizes the perception of abstractivity in summaries and whether there are grades of variation in that perception, from complete extractive summaries to purely abstract summaries. We also explore ways to determine when two summaries are valid for the exact text despite having notably different wording and structure. Additionally, we study the emotions that arise from certain words in summaries, how the emotions vary from the document to the summary, and how automatic summarization models are influenced by this aspect.
Palabras clave
- Natural Language Processing
- Automatic Summarization
- Artificial Intelligence
- Transformers
- Redes neuronales
- Neural networks
- Modelos de lenguaje
- Language models
- Resúmenes extractivos
- Extractive summaries
- Lingüística computacional
- Computational linguistics
- Generación de texto
- Text generation
- Procesamiento del lenguaje natural
Publicaciones durante la tesis (12)
- 2025 CongresoELiRF-UPV at eRisk 2025: New Approaches to the Detection and Early Detection of Symptoms and Signs of Depression
Andreu Casamayor Segarra , Vicent Ahuir Esteve , Antonio Molina Marco , Lluís Felip Hurtado Oliver
The Conference and Labs of the Evaluation Forum, 1-15
- 2025 CongresoELiRF-UPV at MentalRiskES 2025: Spanish Longformer for Early Detection of Gambling Addiction Risk
Andreu Casamayor Segarra , Vicent Ahuir Esteve , Antonio Molina Marco , Lluís Felip Hurtado Oliver
Iberian Languages Evaluation Forum
- 2024 CongresoELiRF-VRAIN at MentalRiskES 2024: Using LongFormer for Early Detection of Mental Disorders Risk
Andreu Casamayor Segarra , Vicent Ahuir Esteve , Antonio Molina Marco , Lluís Felip Hurtado Oliver
Iberian Languages Evaluation Forum
- 2024 CongresoELiRF-VRAIN at eRisk 2024: Using LongFormers for Early Detection of Signs of Anorexia
Andreu Casamayor Segarra , Vicent Ahuir Esteve , Antonio Molina Marco , Lluís Felip Hurtado Oliver
Conference and Labs of the Evaluation Forum, 803
- 2024 CongresoELiRF-VRAIN at BioLaySumm: Boosting Lay Summarization Systems Performance with Ranking Models
Vicent Ahuir Esteve , Torres, Diego , Encarna Segarra Soriano , Lluís Felip Hurtado Oliver
Workshop on Biomedical Natural Language Processing, 755
- 2024 RevistaAbstractive Summarizers Become Emotional on News Summarization.
Vicent Ahuir Esteve , José Ángel González Barba , Lluís Felip Hurtado Oliver , Encarna Segarra Soriano
Applied Sciences · DOI: 10.3390/app14020713
- 2023 CongresoELiRF-VRAIN at DIPROMATS 2023: Cross-lingual Data Augmentation for Propaganda Detection
Vicent Ahuir Esteve , Lluís Felip Hurtado Oliver , Fernando García Granada , Emilio Sanchis Arnal
Iberian Languages Evaluation Forum
- 2023 CongresoELiRF-VRAIN at PoliticES-IberLEF2023: Dealing with Long Texts in Transformer-based Systems for User Profiling
Vicent Ahuir Esteve , Lluís Felip Hurtado Oliver , Fernando García Granada , Emilio Sanchis Arnal
Iberian Languages Evaluation Forum
- 2023 CongresoELiRF-VRAIN at BioNLP Task 1B: Radiology Report Summarization
Vicent Ahuir Esteve , Encarna Segarra Soriano , Lluís Felip Hurtado Oliver
Workshop on Biomedical Natural Language Processing, 524
- 2022 CongresoGUAITA: Monitorización y análisis de redes sociales para la ayuda a la toma de decisiones
Ferran Pla Santamaría , Lluís Felip Hurtado Oliver , José Ángel González Barba , Vicent Ahuir Esteve , Encarna Segarra Soriano , Emilio Sanchis Arnal , María José Castro Bleda , Fernando García Granada
Congreso de la Sociedad Española para el Procesamiento del Lenguaje Natural (SEPLN), 79
- 2022 CongresoEnhancing Sexism Identification and Categorization in Low-data Situations
Vicent Ahuir Esteve , José Ángel González Barba , Lluís Felip Hurtado Oliver
Iberian Languages Evaluation Forum
- 2022 CongresoDACSA: A large-scale Dataset for Automatic summarization of Catalan and Spanish newspaper Articles
Encarna Segarra Soriano , Vicent Ahuir Esteve , Lluís Felip Hurtado Oliver , José Ángel González Barba
Annual Conference of the North American Chapter of the Association for Computational Linguistics, 5931