What we're about
Upcoming events (2)
This is a partner event with our friends at the Hungarian Natural Language Processing Meetup. For the Zoom link, please register at: https://www.meetup.com/Hungarian-nlp/events/273884624/ Abstract: Knowledge graphs are increasingly becoming important in the AI world as an enabling technology for data integration and analytics, semantic search and question answering, and other cognitive applications. However, developing and maintaining large knowledge graphs in a manual way is too expensive and time consuming. To accelerate things, methods and techniques from the areas of information extraction and natural language processing (NLP) can be very helpful. In this talk we'll see the main NLP tasks that knowledge graph mining involves, the factors that affect how easy or difficult the execution of these tasks can be, and some common pitfalls that we need to avoid in order to mine high quality knowledge graphs. About the Speaker Panos Alexopoulos, author of the recently published O'Reilly book : Semantic Modeling for Data, has been working for more than 12 years at the intersection of data, semantics, language and software, contributing in building semantics-powered systems that deliver value to business and society. Born and raised in Athens, Greece, Panos currently works as Head of Ontology at Textkernel, in Amsterdam, Netherlands, where he leads a team of data professionals (Linguists, Data Scientists and Data Engineers) in developing and delivering a large cross-lingual Knowledge Graph in the HR and Recruitment domain. Prior to Textkernel, he worked at Expert System Iberia (former iSOCO) in Madrid, Spain, as a Semantic Applications Research Manager, and at IMC Technologies in Athens, Greece, as a Semantic Solutions Architect and Ontologist. Academically, Panos holds a PhD in Knowledge Engineering and Management from National Technical University of Athens, and has published ~60 papers at international conferences, journals and books. He strives though to present his work and experiences in all kinds of venues, trying to bridge the gap between academia and industry so that they can benefit from one another. For more info, visit: http://www.panosalexopoulos.com/ https://learning.oreilly.com/library/view/semantic-modeling-for/9781492054269/ This is a partner event with our friends at the Hungarian Natural Language Processing Meetup. For the Zoom link, please register at: https://www.meetup.com/Hungarian-nlp/events/273884624/
This is a joint meetup with Bay Area NLP. For the Zoom link, RSVP at: https://www.meetup.com/Bay-Area-NLP/events/272678814/ Meghana Ravikumar from SigOpt will join us to talk about finding the architecture for your language models. Summary: Most pretrained language models are too large for many use cases, especially if they need to run on devices like phones, and they are not adapted to your domain. So, how can you distill a language model to a size that makes them easier to deploy, while at the same time adapting them to your task? In this talk, we will explore how to reduce the size of BERT while retaining its capacity, using a Question Answering task as a worked example. We will cover fine-tuning and distillation using a teacher-student model. We will show how multimetric hyperparameter optimization can find the optimal architecture for the student model. Finally, we explore the trade-offs of our hyperparameter decisions to draw insights for our student model’s architecture. Bio: Meghana is a Machine Learning Engineer at SigOpt. She has worked with machine learning in academia and in industry and previously worked in biotech, employing NLP to mine and classify biomedical literature.