#01Nov 2, 2026
cs.CV
R$^2$-Omni: Reducing Structural and Precision Redundancy in Omni-LLMs for Long-Form Generation
Bolong Feng, Pengfei Zhou, Zhenglin Wan and 2 more
Omni-modal large language models (Omni-LLMs) typically process large numbers of multimodal tokens from audio-visual inputs, incurring substantial computational costs and a large memory footprint for the key-value (KV) cache. Aggressive token pruning can alleviate this problem while maintaining competitive performance on multiple-choice tasks, which require only a single option as output. However, we observe substantial quality degradation when generating long-form, open-ended responses. We attribute this gap to query-guided selection, which can retain sufficient evidence for answer selection while discarding details needed for comprehensive generation. This motivates reducing overlapping information across tokens, termed structural redundancy, while preserving diverse content. However, the retained representations may still contain precision redundancy due to unnecessarily high numerical precision in their cached values. To address these two levels of redundancy in token representations, we introduce a training-free framework named R$^2$-Omni. First, Structure-Aware Token Consolidation (STC) consolidates redundant representations using spatiotemporal similarity in video and local continuity in audio. Then, Semantic-Sensitivity Guided Precision Allocation (SSPA) allocates value precision to the retained representations according to the estimated reduction in attention-output error under quantized keys, with query attention and audio-visual alignment providing a semantic prior. Extensive experiments on long-form generation benchmarks demonstrate an average 5.0$\times$ reduction in the memory footprint while preserving 97.98% of baseline generation quality and achieving a 1.48$\times$ end-to-end speedup, highlighting the benefits of jointly reducing structural and precision redundancy.
#02Nov 2, 2026
cs.CV
Isekai3D SDG: A Platform for Diverse Synthetic Gameplay Visual Data
Rafal Karp, Gene Liu, Shih-Hsin Li and 10 more
We present Isekai3D Synthetic Data Generation (SDG), a platform for simulating and capturing synthetic 3D gameplay for research in vision and world models, neural graphics, gameplay agents, and other applications. The platform combines reusable environment and character assets with Unreal Engine plugins in a two-stage workflow. First, world simulation generates gameplay while recording the world state, character movements, actions, and camera tracks needed for replay. Second, the recorded simulation is replayed to capture selected outputs, such as rendered video, images, and depth maps, at a specified resolution and frame rate under selected rendering and camera configurations. Separating simulation from capture allows the same recorded gameplay to produce multiple output variants without repeating the original simulation. Visual outputs are accompanied by frame-aligned camera, character, and action metadata. The platform uploads generated data directly in a training-friendly format, reducing intermediate storage, preparation time, and manual processing. This report describes the design of our overall system, including the platform's assets, plugins, and their functionality, explaining how they work together to generate diverse synthetic gameplay data.
#03Nov 2, 2026
cs.CV
3D-BBDA: Backpropagation-Free Black-Box Domain Adaptation for Robust 3D Point Cloud Recognition
Shivangi Rai, Kunal Jangid, Vaibhav Kumar and 1 more
Robust 3D point-cloud recognition is critical for robotics, mapping, and urban sensing, yet point-cloud classifiers often suffer performance drops under domain shifts caused by sensor changes, environmental variation, and acquisition artifacts. Many existing adaptation methods are difficult to deploy in practice because they require access to source data or the internal parameters of the source model, which may be restricted by privacy, licensing, or security constraints. We present 3D-BBDA, a backpropagation-free black-box domain adaptation framework that adapts a source-trained predictor to unlabeled target data using only black-box outputs. Our method performs pseudo-label guided prototype-based alignment and employs a weighted prediction module to fuse source logits with alignment logits, reducing sensitivity to noisy pseudo-labels. To stabilize prototype memory when pseudo-label coverage is incomplete, we construct a Gaussian-perturbed prototype prior from unlabeled target predictions and use it as a fallback for missing or unreliable class prototypes during alignment. Experiments on four corrupted 3D benchmarks demonstrate consistent improvements over strong baselines. On ModelNet-40C, 3D-BBDA achieves mean errors of 32.08% (PointNet) and 29.62% (GDANet), outperforming Source-only (35.88% / 34.03%) and BFTT3D (33.78% / 32.36%). 3D-BBDA requires only a single forward pass with no gradient updates making it suitable for real-time, privacy-preserving deployment in transferable 3D recognition pipelines. Project page is available at https://visdomlab.github.io/3D_BBDA
#04Nov 2, 2026
cs.CV
VidDub: Replacement-Aware Distillation for Two-Step Video Dubbing
Yexin Liu, Yunlong Lin, Zehan Wang and 11 more
Video dubbing aims to synthesize lip movements that match target speech while preserving the speaker identity, head pose, and background of a source video. Existing editors are predominantly trained via self-reconstruction under original paired audio, even though the source mouth becomes a conflicting target when driving speech is replaced. We introduce replacement-aware cross-audio supervision: paired-audio samples use full-frame reconstruction, whereas unpaired replacement-audio samples mask out source-mouth reconstruction losses and supervise the mouth solely through audio-visual synchronization while preserving unmasked facial regions and background. Because offline dubbing provides the complete source video and target speech beforehand, we present VidDub, a full-context diffusion editor distilled into two denoising steps (2 NFE). Our replacement-aware teacher first learns accurate lip synchronization without regressing toward conflicting source mouth movements. Matched trajectory regression then transfers the teacher's CFG-guided trajectories to a two-step student, and distribution matching distillation recovers visual quality after step compression. We will release the curated training dataset, VidDub-Bench, fixed paired- and cross-audio evaluation manifests, preprocessing outputs, and unified evaluation code. Across three training seeds, masking out conflicting source-mouth reconstruction improves cross-audio Sync-C by 0.833; the final two-step model combines strong lip synchronization with superior visual quality, identity preservation, and visual speech recognition across four benchmarks.
#05Nov 2, 2026
cs.CV
Align Once to Explain: Feature Alignment for Scalable B-cosification of Foundational Vision Transformers
Raphael Maser, Siddhartha Gairola, Sukrut Rao and 1 more
Foundational vision models have become the de facto standard for many vision tasks due to their strong performance. However, they are notoriously opaque and remain hard to interpret. We present ALOE (ALign Once to Explain), a one-time, label-free feature alignment approach that efficiently converts foundational vision models into inherently interpretable B-cos variants. Once aligned, the B-cos backbone is used as a drop-in replacement across several downstream tasks-amortizing the cost of interpretability. ALOE is robust across pre-training paradigms (supervised, self-supervised, vision-language) and can preserve teacher-level generalization with orders of magnitude less data than training from scratch. On classification, it strongly outperforms vanilla B-cosification (e.g., +10.5 p.p. top-1 on ImageNet for DINOv3 ViT-B/16), retains strong linear probing, k-NN, and zero-shot transfer competitive with foundational backbones (DINOv3, SigLIP2) across diverse downstream datasets. It also preserves spatially structured features useful for dense prediction, while yielding well-localized and highly human-interpretable explanations by design.