1/
Excited to share our recent work: TACO
A Self-Evolving Framework for Efficient Terminal Agents via Observational Context Compression.
mousy-haumea-c5a.notion.site…
Join us for the ATLAS seminar (hybrid) on "Generative Information Retrieval: Progress and Challenges" by Prof. Maarten de Rijke (University of Amsterdam).
All welcome!
📅 18 Feb 2026
⏰ 13:00–14:00
📍 Kilburn_TH 1.3 online
Today is the last day of the main conference, and it began with a keynote from Prof Chenghua Lin.
We have oral and poster presentations, and the day will end with best paper awards and a closing session.
Join us!
Hao’s thesis, “Argument Summarisation via Large Language Models and Diffusion Modelling,” was supervised by Viktor Schlegel, Riza Batista-Navarro, and Goran Nenadic.
Bohao’s thesis, “Enriching Language Model Capabilities: From Dialogue Intelligence and Evaluation to Tabular Data Reasoning,” was supervised by @chenghua_lin.
✨Inspiring seminar by Prof. Roberto Navigli (@RNavigli, Sapienza University of Rome) on “Do Large Language Models Understand Word Meanings?”
Lively audience Q&A and insightful discussion on LLMs, word sense disambiguation, and semantic challenges 🚀 🍻