7384 papers
This paper introduces OmniSpatial, a comprehensive benchmark grounded in cognitive psychology with over 8.4K annotated samples across four major categories, which reveals significant limitations in current vision-language models' spatial reasoning capabilities and explores strategies like PointGraph and SpatialCoT to address them.
The paper proposes UniCUE, a unified framework that directly generates speech from Chinese Cued Speech videos by integrating recognition and generation tasks to overcome the limitations of text-intermediate pipelines, supported by the newly constructed large-scale UniCUE-HI dataset.
This study addresses the challenge of overconfident predictions in closed-world animal classification by demonstrating that feature-based out-of-distribution detection methods, particularly Nearest Class Mean with ImageNet pre-trained features, significantly outperform existing techniques in identifying unknown wildlife species within the context of Africa's Big Five.
The paper proposes Meta-Adaptive Prompt Distillation, a meta-learning framework that distills task-relevant visual features into adaptable soft prompts via an attention-mapper module, significantly outperforming standard in-context learning and parameter-efficient fine-tuning for few-shot Visual Question Answering in Large Multimodal Models.
BitVLA introduces a fully native 1-bit Vision-Language-Action model for robotic manipulation that achieves performance comparable to full-precision baselines while significantly reducing memory footprint and latency through native ternary parameter design and a novel Quantize-then-Distill strategy for the vision backbone.
The paper introduces PDGS, a self-supervised framework that leverages Gaussian Splatting to achieve accurate part-level decoupling and continuous deformation modeling of articulated objects by learning a shared canonical field, while also releasing the RS-Art dataset for real-world evaluation.
VITA introduces a zero-shot value function learning method that enhances the generalization and temporal reasoning of frozen Vision-Language Models through test-time adaptation and dissimilarity-based sampling, enabling robust performance in diverse robotic tasks and improving offline reinforcement learning policies.
The paper introduces VINCIE, a block-causal diffusion transformer trained exclusively on video-derived multimodal sequences through three proxy tasks, which achieves state-of-the-art performance in in-context image editing and demonstrates strong capabilities in multi-concept composition and story generation.
This paper introduces **NIC-RobustBench**, an open-source framework designed to evaluate the adversarial robustness of neural image compression models and their impact on downstream tasks, addressing the current gap in benchmarks that primarily focus only on rate-distortion performance.
This paper proposes Consistency-driven Calibration and Matching (ConCM), a novel framework for Few-Shot Class-Incremental Learning that mitigates knowledge conflicts by integrating memory-aware prototype calibration and dynamic structure matching to achieve state-of-the-art performance on large-scale benchmarks.
This paper introduces VisionDrop, a training-free, visual-only token pruning framework that addresses cross-modal misalignment by selecting informative visual tokens through intra-modal attention and progressive merging, achieving significant inference efficiency gains while maintaining high performance in Large Vision-Language Models.
EchoMimicV3 is an efficient, unified framework that leverages a 1.3B parameter model and novel multi-task and multi-modal strategies to overcome the computational and speed limitations of existing large-scale human animation methods while delivering competitive performance across diverse tasks.
This paper introduces CLiFT, a neural rendering framework that compresses multi-view scene information into adaptive tokens, enabling compute-efficient novel view synthesis with controllable trade-offs between data size, rendering quality, and speed.
The paper introduces Segment Concept (SeC), a novel video object segmentation framework that leverages Large Vision-Language Models to progressively construct high-level object-centric representations, achieving state-of-the-art performance on a new Semantic Complex Scenarios benchmark (SeCVOS) by significantly outperforming existing methods like SAM 2.
This paper presents a unified, closed-loop digital and robotic twinning framework that integrates faster-than-real-time simulations with Stanford's robotic testbeds to validate and verify the performance of spacecraft guidance, navigation, and control systems for rendezvous and formation flying missions.
MonoFusion presents a method for reconstructing dynamic scenes from sparse-view videos by aligning independent monocular reconstructions to achieve high-quality, time- and view-consistent results without the need for expensive dense multi-camera setups.
This paper proposes HGTS-Former, a novel hierarchical hypergraph Transformer that effectively models complex multivariate time series couplings through patch embedding and hypergraph-based aggregation, achieving state-of-the-art performance on various tasks including a new large-scale dataset for nuclear fusion Edge-Localized Mode recognition.
This paper proposes a fast magnetic resonance simulation method that groups isochromats with identical properties to enable shared computations, achieving a 3 to 72-fold reduction in processing time compared to conventional individual isochromat simulations.
This paper proposes a method for improving out-of-distribution robustness by learning invariant and sparse "Causal Delta Embeddings" to represent interventions in the latent space, enabling the unsupervised learning of causal representations from image pairs that significantly outperform baselines in both synthetic and real-world benchmarks.
Uni-CoT introduces a unified Chain-of-Thought framework that leverages a two-level reasoning paradigm and structured training to enable efficient, coherent multimodal reasoning across text and vision, achieving state-of-the-art performance on image generation and editing benchmarks with limited computational resources.