What Can Summarization Do for NLP? Problem Settings, Task Definitions, and Solutions

ChatGPT and other human-facing generative language models have become a hot topic in the NLP community and in popular society at large. These models represent a success for the field, but also a source of angst because of their potential for disruption. For example, they are known to generate inaccurate or biased text, and may impact the functioning of our research communities and societies. I propose text summarization as a good focusing use case that can help NLP researchers discuss and anticipate such issues, as text summarization is inherently geared towards human consumption and requires semantic processing. I will discuss illustrative studies from my lab on factuality and hallucination, on proposing new task definitions to expand the scope of NLP applications, and on the solutions that we have devised for several of these issues. Just as machine translation has driven many advances in the first decades of NLP (e.g., alignment algorithms, language models, evaluation practices), summarization has the potential to do so for the next decades.

Bio: Jackie Chi Kit Cheung is an associate professor at McGill University's School of Computer Science, where he co-directs the Reasoning and Learning Lab. He is a Canada CIFAR AI Chair and an Associate Scientific Co-Director at the Mila Quebec AI Institute. His research focuses on topics in natural language generation such as automatic summarization, and on integrating diverse knowledge sources into NLP systems for pragmatic and common-sense reasoning. He also works on applications of NLP to domains such as education, health, and language revitalization. He is a consulting researcher at Microsoft Research Montreal.

Hosts: Filip Miletic & Sabine Schulte im Walde

Colloquium for Computational Linguistics and Linguistics in Stuttgart