Tetsuya Sakai (Waseda University, Tokyo, Japan)
Virtual Keynote (31st March 2023 at 9am IST)
I would like to discuss with the audience a few keywords which I believe should be considered as foundation pillars of modern research practices, namely, fairness, awareness, and sustainability. Other important pillars such as ethics are beyond the scope of this keynote.
Mounia Lalmas (Spotify, UK)
Keynote (3rd April 2023 at 9:30am IST)
One of Spotify’s missions is “to match fans and creators in a personal and relevant way”. This talk will share some of the research work aimed at achieving this, from using machine learning to metric validation, and illustrated through examples within the context of Spotify’s home and search. An important aspect will focus on illustrating that when aiming to personalize for both recommendation and search, it is important to consider the heterogeneity of both listener and content. One way to do this is to consider the following three angles when developing machine learning solutions for personalization: (1) Understanding user journey; (2) Optimizing for the right metric; and (3) Thinking about diversity.
Karen Spärck Jones Award Keynote 2022
William Yang Wang (University of California, Santa Barbara, USA)
Keynote (4th April 2023 at 9:00am IST)
A key goal for Artificial Intelligence is to design intelligent agents that can reason with heterogeneous representations and answer open-domain questions. The advances in large language models (LLMs) bring exciting opportunities to create disruptive technologies for question answering (QA). In this talk, I will demonstrate that major challenges for open-domain QA with LLMs include the capability to reason seamlessly between textual and tabular data, to understand and reason with numerical data, and to adapt to specialized domains. To do this, I will describe our recent work on teaching machines to reason in semi-structured tables and unstructured text data. More specifically, I will introduce (1) Open Question Answering over Tables and Text (OTT-QA), a new large-scale open-domain benchmark that combines information retrieval and language understanding for multihop reasoning over tabular and textual data; (2) FinQA and ConFinQA, two challenging benchmarks for exploring the chain of numerical reasoning in conversational finance question answering. I will also describe other exciting research directions in open-domain question answering.