Who's Who of AI

Tim van Erven

125 trust researcher @timvanerven.nl · 1,063 followers
AI research
Why they matter

Researcher with public evidence across AI research.

AI signals
1
past 30d
Sources
1
distinct domains
Discussions
0
past 30d
Latest signal
16d ago
View every signal from Tim van Erven →
Associate professor in machine learning at the University of Amsterdam. Topics: (online) learning theory and the mathematics of explainable AI. www.timvanerven.nl Theory of Interpretable AI seminar: https://tverven.github.io/tiai-seminar
What they're sharing

Articles & links

↻ Tim van Erven reposted
Eugene Vinitsky @eugenevinitsky.bsky.social

These are very odd failures! arxiv.org/abs/2211.00241

Adversarial Policies Beat Superhuman Go AIs arxiv.org
AI Weekly's analysis →
  • An adversarial policy achieves a greater than 97% win rate against KataGo running at superhuman settings.
  • The attack transfers zero-shot to other superhuman Go AIs and works even against KataGo variants adversarially trained to defend.
  • Human experts can execute the winning strategy manually, defeating superhuman Go AIs without any computer assistance.
Read full analysis →
 View on Bluesky →
Their network

In Tim van Erven's orbit

Center = Tim van Erven. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.