#01Oct 30, 2026
cs.CV
A Better Critic Matters for Distribution Matching Distillation
Zongze Wu, Xianbang Wang, Zhengyang Geng and 5 more
Distribution Matching Distillation (DMD) trains a few-step student using the difference between a teacher's score and an online critic's estimate of the student distribution's score. We investigate what the critic needs, in capacity and in information, to provide a reliable training signal. On the capacity side, controlled experiments across seven SiT model sizes on ImageNet show that increasing teacher size while keeping a small critic fixed can degrade student quality and cause training collapse. Scaling the critic alongside the teacher, a strategy we call joint scaling, prevents collapse in our experiments and allows the student to benefit from larger teachers. On the information side, rather than forcing the critic to model a step-agnostic mixture of student outputs, we introduce a step-aware critic (STAC) that conditions on the student's generation step to estimate step-specific scores. Together, joint scaling and STAC consistently improve few-step distillation across class-conditional ImageNet generation, text-to-image generation with SDXL and FLUX.2, and text-to-video generation with Wan2.1.
#02Oct 30, 2026
cs.CV
GazeRIB: Systematic Evaluation of Multi-Reader Gaze-Guided Chest X-ray Learning
Nattakorn Kittisut, Roy M. Gabriel, Marly van Assen and 4 more
Radiologists' visual attention can supervise medical image learning, but how reader panel composition affects its value remains untested. We evaluated GazeRIB (Gaze-guided Radiologist-Informed Blackout), which masks unviewed regions during training and uses unmasked images at inference. Eight experts repeatedly read the same 80 chest X-rays (CXRs). We compared individual readers with three-, five- and eight-reader panels across seven architectures and 10 seeds, evaluating 910 models on a separate 2,897-CXR cohort for four-class severity prediction. Every panel improved macro F1 over the no-gaze baseline by 5.33-7.47 percentage points (pp; Bonferroni-adjusted p < 0.007); no individual reader reached significance on this endpoint. Three readers captured 73-80% of the eight-reader gain in macro F1, the area under the receiver-operating-characteristic curve (ROC-AUC) and the area under the precision-recall curve (AU-PRC). On external validation without retraining, all four panels improved ROC-AUC on SIIM-FISABIO-RSNA by +1.93 to +4.24 pp (p <= 0.004); on MIMIC-Eye no panel differed significantly from baseline, whereas the single-reader average declined (-1.40 pp, p = 0.008). Reader composition therefore influenced performance, although transfer depended on the cohort.
#03Oct 30, 2026
cs.RO
REA: A Routed Embodied Agent Runtime for Asynchronous Brain-Cerebellum Control
Pengfei Zhou, Yixing Ma, Zhengfeiyang Zhang and 5 more
Foundation models are becoming remarkably reliable at high-level reasoning, yet they remain too slow and costly to serve as fine-grained controllers. Conversely, fast policies like VLAs and whole-body controllers offer more efficient reflexes, but struggle to generalize. While pairing a deliberative brain (foundation models) with a reactive cerebellum (fast policies) seems natural for embodied systems, two obstacles remain: methods are built and scored on incompatible stacks, and their asynchronous coordination is difficult. For example, an agent that calls the VLA in the agentic loop can stall the fast execution, while a VLA that calls the agent on errors tends to re-trigger fallbacks repeatedly or completely miss silent failures. To address this, we propose REA (Routed Embodied Agent), a unified runtime where a deliberative System 2, an embodied router System 1 and reactive System 0 policies cooperate under a single golden rule - no running policy waits on the brain. Instead of querying the brain at each failure, REA employs a route memory that stores each policy's measured success ahead of time, a boundary-triggered judge gate that consults the brain on policy execution, and a shared-budget escalation ladder for bounded recovery. Evaluated on 3,000 test episodes across 60 LIBERO-Pro tasks, REA reaches 95.9% success and beats the best single-agent baseline in its own policy pool (94.7%) while reducing control steps by 25% and control time by 17%. In a separate 100-episode RoboTwin 2.0 study, REA achieves 84% success with VLA policies and 86% with VLA+code, compared with 80% for the strongest single VLA baseline MEDex. These results demonstrate learned and programmatic policies operating within the same runtime. REA also transfers to BEHAVIOR-1K and a real Unitree G1 humanoid, achieving the highest observed success among the compared methods.
#04Oct 30, 2026
cs.CV
WorkIng: A Real-World and Synthetic Highway Work-Zone Benchmark Dataset for Autonomous Navigation
Xujun Xie, Dominik Rößle, Adithya Mohan and 2 more
Highway work zones are critical edge cases for autonomous vehicles (AVs), combining atypical geometries and jurisdiction-specific regulations that existing benchmarks fail to capture: real-world datasets lack 360° multi-modal, full-scene 3D work zone annotation, and no single recorded site can span the layouts and regulatory variants needed for generalizable models. We present WorkIng, a dual-domain benchmark and training resource. WorkIng-Highway provides a 360° multi-modal work zone benchmark with synchronized 128-channel LiDAR, six RGB cameras, and RTK-GPS across 32 km, yielding 12,000+ annotated frames and 227,399 tracked 3D bounding boxes. WorkIng-CARLA complements this with procedural simulation spanning 24 regulation-aware scenarios across six canonical topologies and six international traffic standards, populated by 100+ custom 3D assets. Benchmarking representative 3D detection and end-to-end planning architectures shows that real-world data remains the primary performance driver, and that scenario-stratified synthetic augmentation improves architectures across 2D recognition, 3D detection, and trajectory planning. Resources: https://github.com/cvims/WorkIng
#05Oct 30, 2026
cs.CV
Think in 360°: Benchmarking and Bootstrapping Panoramic Video Reasoning
Lidong Lu, Lidong Zou, Xinyi Mao and 5 more
A panoramic video captures all directions around the camera, but relevant evidence can appear in different directions and at different times. Understanding it requires models to connect observations across the viewing sphere and time. We introduce PanoVideoBench, the first manually annotated benchmark for comprehensive panoramic video understanding. Its 1,000 questions span five categories and 15 question types, with diagnostic annotations for target search and panoramic capabilities. Evaluation of 11 proprietary, open-source, and panoramic-specialized models identifies cross-view tracking and sphere-time evidence integration as their weakest capabilities. To strengthen these capabilities, we construct PanoVideoCoT, the first reasoning dataset grounded in real panoramic video, through sphere-time-grounded synthesis that converts verified panoramic observations and successful agentic explorations into 13,465 QA-CoT pairs. Starting from a general video MLLM, PanoVideoReasoner learns from this corpus through Panoramic CoT SFT and Evidence-Privileged On-Policy Self-Distillation, a panoramic instantiation of Vision-OPD's OPSD in which the teacher uses verified perspective views to guide student-generated trajectories. It reaches 55.7% on PanoVideoBench, improving the backbone by 12.5 points and the previous SOTA by 2.6. The gains concentrate on the same spatial, cross-view, and temporal capabilities exposed by the benchmark, showing that grounded sphere-time supervision can equip a general video MLLM with panoramic reasoning.