Date and time: 27.02.25, 5pm CET
Speaker: Antonio Moreno Sandoval, UAM Computer Linguistics Laboratory and Knowledge Engineering Institute
Abstract:
This talk will focus on the application of NLP to the financial domain. Since the end of the previous decade, all major NLP and AI conferences have organized workshops specialized in the processing of financial narratives, through tasks such as summarization, information extraction, entity recognition, sentiment analysis or search for answers. We will go into detail on FinCausal's shared tasks. They address the treatment of causality in financial reports in Spanish and English. We understand a causal relationship as a succession of events where one involves a cause that produces an effect. The study of causes is important to understand decision-making processes in financial analysis. We will see different types of causes and effects, which can be both agents and facts, quantifiable or not, but which should not be confused with other relationships between events such as conditionals, hypotheses or future projections. We will also see how the emergence of models based on transformers and generative AI have changed causality extraction strategies.
Bio:
Antonio Moreno-Sandoval is Professor of Linguistics at the Autonomous University of Madrid, where he directs the LLI and the UAM-IIC Chair in Computational Linguistics. He began his research training in the late 80s in the European Eurotra project. He worked at the UAM-IBM Scientific Center on a QA project, where he completed his doctoral thesis (Grampal, a Spanish morphological processor in prolog). In the early 90s he did a Fulbright postdoctoral stay in the Proteus group at New York University, directed by Ralph Grishman, where he developed a Spanish grammar for information extraction. Upon his return to Spain, he rejoined the LLI-UAM and directed the UAM Spanish Treebank, with funding from NYU. The following decade he focused on compiling spontaneous speech corpora in different languages (Spanish, Japanese, Chinese) some of these resources have been used as speech training and evaluation datasets. Since 2010, his research has returned to text processing, first in the medical field and since 2018 in financial narrative. His relationship with industry and knowledge transfer has been maintained almost uninterruptedly since the late 1980s with companies such as IBM, Daedalus, Fundación Telefónica, Sigma Technologies and since 2010 with the Institute of Knowledge Engineering, where he collaborates as a senior researcher. This research activity has been developed in parallel with teaching and training activities at the UAM, where he has directed 15 doctoral theses and numerous research grants.
Link to the talk: https://youtu.be/4Z0BeN8PvjA?si=hgNVK81zz5xIzLPk
Registration (mandatory) and link
Zoom
Zoom link (mandatory registration): https://zoom.us/webinar/register/WN_XA2LKdSyRoqqQL6DTs4ueA
