#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
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.
#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
Circumventing Magnitude Collapse via Modular Two-Stage Open-Set Recognition
Laurin van den Bergh, Halil Bisgin, Andres Palechor and 2 more
Open-set recognition (OSR) requires neural networks to classify known classes while rejecting unknown samples, which is critical for real-world deployment. So far, OSR research studied and developed representation learning and postprocessing methods independently and their interactions remain unexplored. In this paper, we reconsider OSR as a modular two-stage pipeline and present the first systematic study of interactions between representation learning and postprocessing methods across dataset scales and usage of known unknown data. First, we discover a failure mode we term \emph{magnitude collapse}, where representation learning methods that utilize known unknown data can suffer performance degradation at large scale and irreversibly destroy discriminative information, despite excelling at small scale. Second, we study the modular interactions and reveal when they can be leveraged for modular performance gains via two-stage processing and where interaction effects amplify performance degradation due to magnitude collapse. Third, we show how these findings can be used to achieve state-of-the-art performance with a simple $K+1$ class classifier baseline and magnitude-aware postprocessing while circumventing magnitude collapse. Finally, our results demonstrate that small-scale evaluations with known unknown data are not predictive of large-scale performance, invalidating current best practices in OSR research. Code available at https://github.com/TwoDigitsOneNumber/Two-Stage-OSR.
#05Nov 2, 2026
cs.CV
ASCoR: Asymmetric Semantic-Spatial Learning for Remote Sensing Domain Generalization
Zitong Wang, Shuchang Lyu, Meng Li and 3 more
Source-only domain generalization (DG) for remote sensing semantic segmentation seeks reliable predictions in unseen domains without target-domain data during training. Although vision foundation models offer transferable knowledge, category recognition remains sensitive to source-specific visual cues, while task-specific adaptation may weaken pretrained transferability. We propose ASCoR, an asymmetric semantic-spatial framework built on SAM3. Its text-conditioned semantic branch uses class-name prompts to provide explicit category guidance. Probe-Once Dynamic Adaptive LoRA distributes a fixed adaptation budget across encoder blocks to limit source-specific specialization. A complementary task-specific dense branch learns multi-scale spatial representations using a separate frozen SAM3 image encoder. A lightweight reliability router integrates the two branches through bounded spatial correction based on local prediction reliability. All components are trained using source-domain supervision only. Experiments on the ISPRS, LoveDA, and OpenEarthMap benchmarks cover nine cross-domain transfers. ASCoR achieves a state-of-the-art mean IoU of 68.61% across six ISPRS transfers, outperforming the previous best method by 2.69 percentage points. It also achieves the best results on all three LoveDA/OpenEarthMap transfers, with 59.07%, 47.95%, and 49.38% mIoU on Rural-to-Urban, Urban-to-Rural, and Urban-to-OpenEarthMap, respectively. Ablation studies, feature visualizations, and spatial reliability analyses support the effectiveness and complementarity of the proposed components. Code: https://github.com/WangZitong42/ASCoR.