← 研踪 PaperTrailarXiv:2611.21094 · v1
Conflict-Aware Multi-Teacher On-Policy Distillation in Aligned Action Spaces for Cross-Game Generalization
首次提交 2026-11-16 21:40
版本更新 2026-11-16 21:40
雷达观测 2026-11-17 17:11
摘要
Zero-shot cross-game generalization aims to consolidate accessible source-game experts into a single vision-language model (VLM) that controls unseen games from raw pixels without target-game parameter updates. While multi-teacher on-policy distillation (MOPD) mitigates offline exposure bias and single-game reinforcement learning (RL) overfitting, scaling it to autoregressive VLMs faces two bottlenecks. First, token-to-action misalignment from subword length bias distorts discrete action probabilities and uncertainty margins. Second, multi-teacher policy conflict arises as source teachers under incompatible dynamics issue contradictory closed-loop corrections, causing negative transfer. We propose CAMOPD (Conflict-Aware Multi-Teacher On-Policy Distillation), a post-training framework that resolves multi-teacher policy conflict within an aligned semantic action space. CAMOPD first uses complete-action alignment to normalize command completion likelihoods over each game's legal action set, forming a calibrated categorical policy that provides reliable uncertainty margins for selective on-policy retention and visual specialist probes that can be evaluated across source games. To resolve policy conflict without target-game data, conflict-aware multi-teacher consolidation cross-evaluates these specialists across source games to capture both visual identifiability and inter-teacher compatibility, upweighting compatible teachers while suppressing conflicting corrections. On GameWorld's temporal and directional transfer suites with Qwen3.5-4B, CAMOPD achieves 26/30 (86.7%) and 24/30 (80.0%) Task-1 zero-shot success on unseen GeoDash and Doodle Jump, outperforming prompting, weight-merging, and multi-task baselines.
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-16 21:40 · 最新观测版本
Conflict-Aware Multi-Teacher On-Policy Distillation in Aligned Action Spaces for Cross-Game Generalization
Zero-shot cross-game generalization aims to consolidate accessible source-game experts into a single vision-language model (VLM) that controls unseen games from raw pixels without target-game parameter updates. While multi-teacher on-policy distillation (MOPD) mitigates offline exposure bias and single-game reinforcement learning (RL) overfitting, scaling it to autoregressive VLMs faces two bottlenecks. First, token-to-action misalignment from subword length bias distorts discrete action probabilities and uncertainty margins. Second, multi-teacher policy conflict arises as source teachers under incompatible dynamics issue contradictory closed-loop corrections, causing negative transfer. We propose CAMOPD (Conflict-Aware Multi-Teacher On-Policy Distillation), a post-training framework that resolves multi-teacher policy conflict within an aligned semantic action space. CAMOPD first uses complete-action alignment to normalize command completion likelihoods over each game's legal action set, forming a calibrated categorical policy that provides reliable uncertainty margins for selective on-policy retention and visual specialist probes that can be evaluated across source games. To resolve policy conflict without target-game data, conflict-aware multi-teacher consolidation cross-evaluates these specialists across source games to capture both visual identifiability and inter-teacher compatibility, upweighting compatible teachers while suppressing conflicting corrections. On GameWorld's temporal and directional transfer suites with Qwen3.5-4B, CAMOPD achieves 26/30 (86.7%) and 24/30 (80.0%) Task-1 zero-shot success on unseen GeoDash and Doodle Jump, outperforming prompting, weight-merging, and multi-task baselines.
查看此版本解读 →查看此版本原文 ↗