I’ll start with the basic concepts and building blocks of Open Retrieval Question Answering (QA): starting with cross-lingual document retrieval and then proceeding onto reading comprehension including generative long-form QA. The talk will also touch upon domain adaptation capabilities with Question Generation and multi-modality with TableQA and VisualQA. Towards the second half of the talk, I’ll introduce PrimeQA and how one can easily build a QA tool with it while being able to replicate the state-of-the-art in QA.
Avi Sil is a Principal Research Scientist and a Research Manager in the NLP team at IBM Research AI. He manages the Question Answering team (comprising of research scientists and engineers) that works on industry scale NLP and Deep Learning algorithms. His team’s system called `GAAMA’ has obtained the top scores in public benchmark datasets e.g. Natural Questions, TyDI and has published several papers on question answering. He is the Chair of the NLP professional community of IBM. Avi is a Senior Program Committee Member and the Area Chair in Question Answering for ACL and is actively involved in the NLP conferences by giving tutorials (ACL 2018, EMNLP 2021), organizing a workshop (ACL 2018) and also the Demo Chair (NAACL 2021, 2022). He was also the track coordinator for the Entity Discovery and Linking track at the Text Analysis Conference (TAC) organized by the National Institute of Standards and Technology (NIST).