University of North Texas

3 related articles

Subscribe to updates

arXiv Oct 9, 2026
CAPABLE lifts a frozen VLA from 24.8% to 59.3% success on an actuator excluded from fault training
CAPABLE is a capability-aware adaptation framework for frozen vision-language-action policies: a temporal encoder shared across joints infers from command-response history and live Jacobian grounding how much commanded motion each joint still realizes, a self-supervised physical-prediction objective ties that representation to behavior, and a bounded residual reinforcement-learning policy corrects the VLA arm action without fault labels or faulty-joint identifiers; across 28 LIBERO tasks it raises success on an actuator excluded from fault training from 24.8% to 59.3%, 17.4 points above a parameter-matched global-history baseline, while preserving healthy performance.
CAPABLE is a capability-aware adaptation framework for frozen vision-language-action policies: a temporal encoder shared across joints infers from command-response history and live Jacobian grounding how much commanded motion each joint still realizes, a self-supervised physical-prediction objective ties that representation to behavior, and a bounded residual reinforcement-learning policy corrects the VLA arm action without fault labels or faulty-joint identifiers; across 28 LIBERO tasks it raises success on an actuator excluded from fault training from 24.8% to 59.3%, 17.4 points above a parameter-matched global-history baseline, while preserving healthy performance.
University of North Texas University of Bremen
Mohammad Khoshnazar（First author）
CAPABLE Residual Reinforcement Learning Vision-Language-Action Policy Fault-Tolerant Robot Control LIBERO
arXiv Sep 30, 2026
DCSD decouples credit direction from magnitude in self-distillation, topping 11 benchmarks and lifting math reasoning by 8.45 points over base models
The work introduces Decoupled Credit Self-Distillation (DCSD), which uses belief-margin probing to set each reasoning step's credit direction and marginal information gain to quantify its contribution magnitude, leaving the privileged teacher to allocate credit only within steps; across 11 mathematical and multimodal reasoning benchmarks DCSD achieves the best overall scores against GRPO, OPSD, RLSD and RLCSD, improving overall score by 8.45 points on mathematical reasoning and 7.01 points on multimodal reasoning over base models, while correcting credit direction for about 6% of tokens and reducing token credit magnitude by roughly 1.5 times.
The work introduces Decoupled Credit Self-Distillation (DCSD), which uses belief-margin probing to set each reasoning step's credit direction and marginal information gain to quantify its contribution magnitude, leaving the privileged teacher to allocate credit only within steps; across 11 mathematical and multimodal reasoning benchmarks DCSD achieves the best overall scores against GRPO, OPSD, RLSD and RLCSD, improving overall score by 8.45 points on mathematical reasoning and 7.01 points on multimodal reasoning over base models, while correcting credit direction for about 6% of tokens and reducing token credit magnitude by roughly 1.5 times.
University of North Texas Adelaide University Curtin University The University of Sydney
Yugu Li（First author）
Decoupled Credit Self-Distillation Belief-Margin Probing Marginal Information Gain On-Policy Self-Distillation Temporal Credit Assignment
arXiv Sep 29, 2026
REALM adds a delay prior and stochastic expression refinement to turn listening from a frozen stare into blinks and smiles on a robot
REALM is a coarse-to-fine framework for audio-driven reactive listening that fuses listener motion history with speaker audio through a Reactive Gated Speaker–Listener Fusion module using a delay-centered attention prior and adaptive gating, then predicts a base motion trajectory with a coarse decoder and augments it with audio-conditioned stochastic residuals in the expression subspace, improving over the evaluated baselines on multiple motion-quality metrics on ViCo and L2L and deploying to an Ameca humanoid robot with a perceptual user study.
REALM is a coarse-to-fine framework for audio-driven reactive listening that fuses listener motion history with speaker audio through a Reactive Gated Speaker–Listener Fusion module using a delay-centered attention prior and adaptive gating, then predicts a base motion trajectory with a coarse decoder and augments it with audio-conditioned stochastic residuals in the expression subspace, improving over the evaluated baselines on multiple motion-quality metrics on ViCo and L2L and deploying to an Ameca humanoid robot with a perceptual user study.
University of North Texas Macquarie University
Peizhen Li（First author）
REALM Reactive Gated Speaker-Listener Fusion Coarse-to-Fine Stochastic Refinement Audio-Driven Listener Motion Generation Human-Robot Interaction

You’re all caught up

CAPABLE lifts a frozen VLA from 24.8% to 59.3% success on an actuator excluded from fault training

DCSD decouples credit direction from magnitude in self-distillation, topping 11 benchmarks and lifting math reasoning by 8.45 points over base models

REALM adds a delay prior and stochastic expression refinement to turn listening from a frozen stare into blinks and smiles on a robot