JaxAHT v1.1 released: JAX-based Ad Hoc Teamwork benchmark with Hanabi and LBF tasks
PeterStone_TX · x · 2026-10-12
CarolineWang98's team has released the JaxAHT paper alongside a v1.1 update of the JAX-based Ad Hoc Teamwork benchmark library. New in v1.1: Hanabi (full and mini variants) and LBF 12x12 tasks with heldout evaluation sets and best-response data, plus modules for human data collection/processing, teammate generation, and open-ended training. MIT-licensed on GitHub (26 stars, 900+ commits), aimed at the ad hoc teamwork / MARL research community.
More from Research
- OpenAI raises the plane's chromatic number lower bound to 6, Lean-formalized — fortnow · 2026-10-12
- Snorkel AI Expands Open Benchmark Grants 10x to $30M, Adds Red Team — ajratner · 2026-10-12
- How long is the coast of a neuron? Point-neuron abstraction may hide capacity — neurovium · 2026-10-12
- Community shrinks OpenAI Problem #190 obstruction from 66x66 to 3x4, verified in Lean — BorisMPower · 2026-10-12
- Autoresearch for LLM pretraining needs humans in the loop; stacked changes are the killer — menhguin · 2026-10-12
- COMODO distills video intelligence into IMU sensors for camera-free egocentric AI glasses — flosalim · 2026-10-12