Beyond Word Embeddings, Complex Question Answering and Common Sense Reasoning
Details
Hi Everyone,
I'm very happy to announce our 12th Israeli NLP meetup, celebrating over 3.5 years of activity and more than 1.6k members! This time we will have a great technical program with two speakers, kindly hosted by Microsoft in their brand new Reactor space in Tel Aviv. In the first talk, Ari Bornstein from Microsoft and Bar Ilan University's NLP lab will tell us about incorporating richer semantic representations into our models, beyond the frequently used pre-trained word embedding methods. In the second talk, Alon Talmor from Tel Aviv University's NLP lab will tell us about his research on complex question answering and common-sense reasoning. See talk abstracts and speaker bio's below.
Looking forward to see you there,
Roee
Schedule:
18:30-19:15 - gathering
19:15-20:00 - Ari’s talk
20:15-21:15 - Alon’s talk
First Talk Abstract - Ari Bornstein
Beyond Word Embeddings
Since the advent of word2vec, word embeddings have become a go-to method for encapsulating distributional semantics in NLP applications. This presentation will review the strengths and weaknesses of using pre-trained word embeddings, and demonstrate how to incorporate more complex semantic representation schemes such as Semantic Role Labeling, Abstract Meaning Representation and Semantic Dependency Parsing into your applications.
Speaker Bio
Aaron (Ari) Bornstein is a Cloud + AI professional with a passion for history, engaging with new technologies and computational medicine. As an Open Source Engineer at Microsoft's Cloud Developer Relations team, he collaborates with Israeli Start-Ups and Communities, to solve real world problems with game changing technologies that are then documented, open sourced, and shared with the developers around the world. To learn more, see: http://aka.ms/israelcda
Second Talk Abstract - Alon Talmor
Complex Question Answering and Common Sense Reasoning
Answering complex questions is a time-consuming activity for humans that requires reasoning and integration of information. Recent work on reading comprehension made headway in answering simple questions based on some relevant document, but tackling complex questions, or question that require commonsense reasoning, is still an ongoing research challenge. In the first part of this talk we will explore a new method for answering Complex Questions. In the second part, "CommonsenseQA: a difficult new dataset for commonsense question answering" will be presented.
Paper links for further reading:
https://arxiv.org/abs/1803.06643
https://arxiv.org/abs/1811.00937
Speaker Bio
Alon Talmor is a software entrepreneur and the former CEO and co-founder of BlueTail, an A.I. startup acquired by Salesforce.com. At Salesforce he was Chief Data Scientist. He has a Masters in Brain Sciences and is pursuing a PhD in NLP at Tel-Aviv University under the supervision of Dr. Jonathan Berant, mainly focused on multi-modal complex question answering and commonsense question answering. In his free time, he lectures on A.I. and brain sciences.