Lecture
LIACS Research Seminar - nr 1
- Date
- Friday 1 December 2023
- Time
- Address
-
Snellius
Niels Bohrweg 1
2333 CA Leiden - Room
- 413
The LIACS Research Seminar, a new monthly seminar series "for-and-by" LIACS researchers, where both speakers and audience are from LIACS. The goal is to provide a platform for the research clusters within LIACS to present our research to each other.
The first speaker will be our very own Suzan Verberne, on Friday 1 December at noon, with free lunch included.
Note: also if the topic is not directly related to your research, please do come! Because such a seminar is an opportunity to learn from each other, about the research highlights within LIACS.
Speaker: Suzan Verberne, professor Natural Language Processing
Title: Finding trustworthy information in the age of LLMs
Abstract: In this talk I will discuss the challenges and opportunities of Large Language Models (LLMs). Generative LLMs such as ChatGPT are currently used by the general public for a write range of tasks: as writing assistant, programming assistant, for information finding, and even as a therapist. Unfortunately, although LLMs can produce probable and fluent text, their output is not always correct: hallucination and bias are common problems. I will first explain the workings of transformer-based language models, and then discuss ways we leverage the power of (open-source) LLMs while at the same time reducing the risk of producing unfaithful information.