DiVeR, 결정 중요도 학습으로 VLA 추론 성능 향상
DiVeR: Decision-Critical Verifier Learning for VLA Test-Time Scaling
- 매체
- arXiv
- 게시
- 중요도
- 4/5
- 화제 지수
- —
한국어 요약
Vision-Language-Action(VLA) 정책의 추론 단계 스케일링 효율을 높이기 위해 샘플링된 행동 표현의 분산으로부터 결정 중요도를 추정하는 검증기 학습 방식 DiVeR을 제안했다. LIBERO, RoboCasa, 실제 로봇 실험에서 검증기 기반 행동 선택의 효율성을 개선했다.
원문 발췌 · English
Scaling robot data and model capacity has improved Vision-Language-Action (VLA) policies, but further progress is constrained by the high cost of robotic data. Verifier-guided test-time scaling offers an efficient alternative by sampling multiple action candidates and selecting the one most likely to lead to task success at inference time. Existing classification-based verifiers learn from trajectory-level outcomes but treat all visited states equally, even though their value for candidate discrimination can vary across a trajectory. At many states, plausible actions are similar and provide limited discrimination signal, while only a sparse set of decision-critical states admits meaningfully different actions that can substantially affect downstream outcomes. To address this, we propose DiVeR, which estimates decision criticality from the dispersion of sampled action representations. DiVeR then uses this signal to reweight verifier learning toward states where action selection is most consequential, without requiring step-level annotations or additional environment interaction. Across LIBERO, RoboCasa, and real-world experiments on a Franka Research 3 robot, DiVeR consistently improves task success through more effective verifier-guided action selection, while adding negligible verifier inference overhead.