Speaker "Alejandra María José Litterio" Details Back



Unveiling ExNLP and Interpretability Models


Title: Unveiling ExNLP and Interpretability Models. Abstract: With the increasing adoption of AI in industries such as finance, legal, medical and defence a critical need for explaining learning based models’ decisions has grown rapidly. Recent explainability approaches in the domain of Natural Language Processing (NLP) focus on models solving tasks that often need human-comprehensible interpretations. Still, can you trust your model? Will it work in deployment? What else can it tell you about the world? We want models to be not only good, but interpretable. Now, how is explainability defined by the research community? Can explainability and interpretability be used interchangeably or are they overlapping terms? Which are the criteria that a model should meet to be performant and equally interpretable? What linguistic knowledge is crucial for certain predictions interpreting the way they represent language? The rise of modern NLP certainly will amplify the challenges of interpretability. In this talk, we seek to refine the discourse on interpretability. Then address why interpretability plays a key role in domain adoption as well as it builds confidence for real-world applications. Finally provide a few guidelines on how evaluation of interpretation methods should and should not be conducted.

Who is this presentation for?

Prerequisite knowledge:
Machine Learning / NLP

What you'll learn?
• Overview Explainable NLP Models and Interpretability: key concepts, definitions. • Why interpretability plays a key role in domain adoption • Criteria to define a EXNLP Model • Model properties and techniques thought to confer interpretability • Practical use cases and real world applications • Guidelines on how evaluation of interpretation methods should and should not be conducted.


Prof. Alejandra M.J. Litterio is a Forensic Linguist, Head of the R&D Department and co-founder at Eye Capital Ltd, a FinTech based in London. She is an expert in NLP leading innovative projects based on the MetaQuant approach and QNLP. She coordinates Eye Capital Quantum Labs, developing new models and frameworks to work in tandem with novel technologies applied to interpreting and coding data in the field of Quantum Computing for the financial markets. Ms. Litterio has been a speaker and panelist internationally. She pursues her Master in Discourse Analysis at Universidad de Buenos Aires. She is actively involved in academia as a researcher and co-director at Centro de Altos Estudios en Tecnología Informática and teaches Computational Linguistics at Universidad Abierta Interamericana.