TU Darmstadt & RIKEN-AIP Joint Workshop
Online joint workshop with talks from TU Darmstadt and RIKEN-AIP on LLM interpretability, AI oversight, and machine learning.
- When
- Tue, October 6, 2026 · 16:00–19:00 JST
- Where
- Online
- Organizer
- RIKEN Center for Advanced Intelligence Project
- Language
- EN
- Source
- Doorkeeper
Summary
TU Darmstadt and RIKEN-AIP hold a joint online workshop bringing together machine learning researchers from both institutions. The session opens with short introductions to the UKP Lab at TU Darmstadt by Iryna Gurevych and to RIKEN-AIP by Masashi Sugiyama, then moves into a research talk program.
Talks cover complementary intelligence and going beyond imitation (Subhabrata Dutta), evaluation, oversight and monitoring of increasingly capable AI systems (Takashi Ishida), automation and cybersecurity when humans face the unknown (Fares Fourati), linear representations in LLM interpretability (Benjamin Heinzerling), post-training dependable agents for conflicting evidence and long-term memory (Anmol Goel), exploration and test-time scaling via variational learning (Nico Daheim), and geometric-structure-preserving machine learning for Hamiltonian dynamics (Takaharu Yaguchi).
The workshop runs online from 16:00 to 19:00 JST (9:00 to 12:00 in Germany, UTC+2) on October 6, 2026, and closes at 18:50 JST. The content is research-level and aimed at people following current work in machine learning, LLM interpretability, and AI safety.
About the community
The host is a Japanese national research center for artificial intelligence that runs a steady program of public seminars, lectures, and joint workshops with universities and labs in Japan and abroad. Sessions are research-level and usually held online in English, so the regular audience is a mix of academic researchers, graduate students, and industry engineers who follow machine learning closely. Events recur most weeks and registration is open to anyone.
#ai#machine-learning#llm#research#interpretability#ai-safety#online