Xiaomi (China)

Tool-using language-model agents often need dozens of interactions before a single success-or-failure label arrives, leaving no way to tell which step went wrong. STEP-VERIFIER turns a success-labeled corpus of logged interactions into action-level comparisons: the difference in average final return between two continuations that share a common prefix labels which of two actions was better, and a Bradley-Terry ranking head converts those pairs into calibrated per-step scores. On Llama-3.1-8B-Instruct and Mistral-7B-Instruct backbones, it beats outcome-only GRPO on the average metric in all six benchmark-backbone combinations across GAIA, ToolBench and AgentBoard, and its per-action scores agree with human judgments on 81.5% of pairs.

Tool-using language-model agents often need dozens of interactions before a single success-or-failure label arrives, leaving no way to tell which step went wrong. STEP-VERIFIER turns a success-labeled corpus of logged interactions into action-level comparisons: the difference in average final return between two continuations that share a common prefix labels which of two actions was better, and a Bradley-Terry ranking head converts those pairs into calibrated per-step scores. On Llama-3.1-8B-Instruct and Mistral-7B-Instruct backbones, it beats outcome-only GRPO on the average metric in all six benchmark-backbone combinations across GAIA, ToolBench and AgentBoard, and its per-action scores agree with human judgments on 81.5% of pairs.

Knowledge distillation for driving vision-language models previously matched teacher distributions uniformly across answer tokens, diluting planning-relevant semantics and passing on confident teacher errors. RSKD first localizes task-critical phrases in the ground-truth answer, such as traffic participants, states and spatial relations, then uses a frozen sentence encoder to score how semantically consistent the teacher's prediction is with each phrase, so consistent positions take more teacher-distribution supervision and inconsistent ones take more ground-truth supervision. On DrivingVQA, LingoQA and MapLMv2, RSKD reaches 16.60, 0.672 and 81.82, beating the compared SFT, GKD and ToDi student baselines on the two QA benchmarks and nearly matching SFT's 81.85 on MapLMv2.

Knowledge distillation for driving vision-language models previously matched teacher distributions uniformly across answer tokens, diluting planning-relevant semantics and passing on confident teacher errors. RSKD first localizes task-critical phrases in the ground-truth answer, such as traffic participants, states and spatial relations, then uses a frozen sentence encoder to score how semantically consistent the teacher's prediction is with each phrase, so consistent positions take more teacher-distribution supervision and inconsistent ones take more ground-truth supervision. On DrivingVQA, LingoQA and MapLMv2, RSKD reaches 16.60, 0.672 and 81.82, beating the compared SFT, GKD and ToDi student baselines on the two QA benchmarks and nearly matching SFT's 81.85 on MapLMv2.

The authors propose AdaStep, which casts how strongly a group-derived local advantage should modify the trajectory-level signal as a mean-squared-error estimation problem for the latent step advantage and, under an explicit conditional sampling assumption, derives an optimal per-state shrinkage coefficient with a signal-to-total-variance interpretation: it preserves local credit when return variation is attributable to the selected action and suppresses it when downstream randomness dominates; the method needs only lightweight scalar computation, with no critic, additional rollouts, or extra model inference, and improves consistently over baselines across three model backbones on ALFWorld, WebShop, and ScienceWorld at low computational cost.

The authors propose AdaStep, which casts how strongly a group-derived local advantage should modify the trajectory-level signal as a mean-squared-error estimation problem for the latent step advantage and, under an explicit conditional sampling assumption, derives an optimal per-state shrinkage coefficient with a signal-to-total-variance interpretation: it preserves local credit when return variation is attributable to the selected action and suppresses it when downstream randomness dominates; the method needs only lightweight scalar computation, with no critic, additional rollouts, or extra model inference, and improves consistently over baselines across three model backbones on ALFWorld, WebShop, and ScienceWorld at low computational cost.