Events
- Vilém Zouhar (ETH Zurich) on the Last Translation Benchmark, and Philippe Bich (Huawei Research) on low-bit inference for LLMs.96
- TodayTiago Pimentel (ETH Zurich) on causality’s role in ML interpretability and Amir Joudaki (ETH Zurich) on learning dynamics and scaling laws.139
- Haithem Turki from NVIDIA on photorealistic world models and Omar Sanseviero from Google DeepMind on the open Gemma model family.295(+58)
- Jannis Vamvas from the University of Zurich on the challenge of Romansh and Eric Chen from EPFL on reasoning as test-time learning.131
- Fabian Schaipp from Inria on training algorithms for machine learning and Valentina Pyatkin from the Allen Institute and ETH Zurich on training open-source LLMs.289
- Sina Ahmadi from the University of Zurich on low-resource language varieties and Barna Pasztor from ETH Zurich on sample-efficient dataset collection for RLHF.179
- Philippe Bich from Huawei on calibration-free low-precision LLMs and Alejandro Hernandez Cano from EPFL on Apertus, the Swiss open LLM.258(+50)
- Aleksander Ficek from NVIDIA on synthetic generators and verifiers for coding and Matteo Saponati from ETH Zurich on the structures of self-attention.139
- Ivan Vulić from the University of Cambridge and Google DeepMind on vision-language models for spatial reasoning and Tiago Pimentel from ETH Zurich on generalisation in language models.187
- Matej Jusup from the ETH AI Center on mastering board games and Bobby He from the University of Oxford on understanding outlier features.118
10 of 25