Salient
-
DeCo: Efficient Decouple-to-Couple Learning for Multi-Task Visual Grounding
2026-09-21 10:58:55Xiaoqiang Lu, Licheng Jiao, Long Sun, Yuting Yang, Xu Liu, Lingling Li, Wenping Ma, Fang LiuAbstract
Multi-task visual grounding requires models to jointly understand linguistic semantics and perform accurate visual localization and segmentation. Despite the success of multimodal large language models, effectively adapting them to multiple grounding objectives remains challenging. Existing methods commonly enforce task cooperation through shared representations, while overlooking the intrinsic conflict between task-oriented feature interests. In this paper, we introduce $\textbf{DeCo}$, an efficient $\textbf{De}$couple-to-$\textbf{Co}$uple learning framework that resolves this dilemma through a two-stage paradigm: task-specific representation decoupling followed by complementary prior coupling. Specifically, we first propose Task-aware Semantic Decoupling (TSD) to route shared visual cues into individual features under salient word-level guidance, alleviating representation interference between localization and segmentation. Furthermore, we observe that segmentation naturally provides informative localization priors due to dense supervision. Based on this insight, we introduce Hybrid Prior Coupling (HPC), which integrates sentence-level semantic prior with mask-derived spatial prior for enhanced grounding. Built upon a frozen multimodal encoder, DeCo requires lightweight trainable parameters while achieving strong generalization across multiple grounding objectives. Extensive experiments on RefCOCO/+, G-Ref, ReferIt, Flickr, DIOR-RSVG, SARVG1.0, RRSIS-D, RIS-LAD, and RefDIOR demonstrate that DeCo achieves state-of-the-art performance on both natural and remote sensing benchmarks. The code and models are available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2609.24409
PDF
-
RegVGGT: Sustainable Visual Geometry Grounding for Streaming via Regulated Memory
2026-09-20 01:42:37Hongbo Mao (Harbin Institute of Technology), Junjun Jiang (Harbin Institute of Technology), Youyu Chen (Harbin Institute of Technology), Jiaxin Zhang (Harbin Institute of Technology), Zhemeng Dong (Harbin Institute of Technology), Xianming Liu (Harbin Institute of Technology)Abstract
3D reconstruction from a lengthy video stream input poses a dilemma for feed-forward reconstruction models (FFRMs), that a whole-stream inference context cannot be retained under limited GPU this http URL studies seek to resolve this problem via a trade-off between the integrity of inference context and GPU memory usage, which either suffer from a rapid memory inflation or degraded context integrity due to artificially capping memory this http URL by our key observation that the initial saliency of a token reliably dictates its long-term importance across the stream, we propose RegVGGT, a training-free token regulation method which aggressively regulates the tokens of incoming this http URL admitting at most 1% of tokens per frame to update the context memory, our method dramatically suppresses memory inflation as the stream this http URL with a FlashAttention-compatible token saliency estimation scheme, RegVGGT is capable of processing thousands of frames on a consumer-grade GPU with negligible compromise to reconstruction this http URL experiments demonstrate that RegVGGT achieves state-of-the-art performance on long-horizon benchmarks across diverse FFRM prediction tasks, surpassing prior FFRM-based stream reconstruction baselines by a large margin.
Abstract (translated)
URL
https://arxiv.org/abs/2609.23286
PDF
-
AdaMerge: Tuning-Free Patch Compression for Multi-Vector Visual Document Retrieval
Abstract
Multi-vector visual document retrieval (VDR) models such as ColPali and ColNomic achieve strong accuracy by representing each document with hundreds to thousands of patch-level embeddings, at substantial storage and latency cost. Existing compression methods either prune unimportant patches or merge similar ones into clusters; the recent state-of-the-art merging method Prune-then-Merge (PtM) consistently outperforms pruning-only baselines at high compression, but requires a per-dataset cluster budget m to be tuned by grid search. We observe that the merge-cosine sequence produced by hierarchical clustering exhibits a sharp cliff separating mergeable redundancy from salient signal, and that the location of this cliff is concentrated in a narrow band across more than 11,000 documents from 14 datasets. This suggests the merge boundary can be detected per document rather than tuned per dataset. Building on this observation, we propose AdaMerge, a plug-and-play compression method that (i) detects each document's own cliff via gap analysis on the merge-cosine trajectory, and (ii) builds attention-weighted cluster centroids to preserve salient signal. On the long-document benchmark ViDoRe-V2 (4 datasets, two backbones), AdaMerge significantly outperforms tuned PtM across the operating range (p < 10^-4); on the short-document benchmark ViDoRe-V1 (10 datasets, two backbones), where all merging methods are already near-lossless, AdaMerge matches tuned PtM without any per-dataset tuning. AdaMerge adds only about 10 ms per document and exposes a single global hyperparameter shared across all datasets and backbones.
Abstract (translated)
URL
https://arxiv.org/abs/2609.22562
PDF
-
Vox-Infinity: Benchmarking the Limits of Long-Context Spoken Language Models
2026-09-18 18:07:47Xize Cheng, Wenxu Jia, Chenyuhao Wen, Dongjie Fu, Zehan Wang, Xinyu Zhang, Tao JinAbstract
Long-context understanding remains a fundamental challenge for large language models, as excessively long inputs often lead models to forget salient information. This issue is even more pronounced in the speech domain, where audio, as a low-compression modality, requires substantially more embeddings than text to preserve both semantic content and acoustic cues. To address this challenge, we introduce \textbf{Vox-Infinity}, the first benchmark specifically designed to evaluate long-context understanding in spoken language models. Vox-Infinity systematically extends audio history along two dimensions: turn count and turn duration. It covers a diverse range of representative scenarios with varying interaction structures and semantic complexity. Crucially, Vox-Infinity provides explicit answer-provenance annotations and organizes samples according to the amount of historical context required to resolve each query, enabling precise and length-aware evaluation. Extensive evaluations of seven representative spoken language models reveal a clear overall recency effect: models generally achieve higher accuracy when answer-supporting evidence is closer to the query, but struggle to retrieve and use evidence located farther back in the dialogue history. Cases and datasets are available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2609.22452
PDF
-
OpenSAL360: Open-Source Crowdsourcing Platform for Omnidirectional Video Saliency Collection
2026-09-18 08:29:09Alexey Bryncev, Andrey Moskalenko, Kira Shilovskaya, Ivan Kosmynin, Dmitriy VatolinAbstract
Omnidirectional video saliency prediction plays an important role in many immersive multimedia applications, including viewport-adaptive streaming and compression, foveated rendering, mesh simplification, perceptual quality assessment. Yet progress in this area remains constrained by the cost and complexity of collecting eye-tracking data with VR headsets, which makes large-scale dataset creation difficult to extend. We present OpenSAL360, the first open-source platform for scalable, low-cost 360° video saliency collection. Unlike conventional VR-based protocols, it requires only a standard screen, mouse, and internet connection, enabling parallel saliency data collection from common crowdsourcing assessors without specialized hardware. We validate our collection protocol against seven well-established VR eye-tracking datasets and conduct ablation studies on key interface, pre-, and post-processing parameters. To demonstrate the effectiveness and scalability of the proposed methodology, we collect and publicly release a saliency dataset covering 500 omnidirectional videos annotated by 2,000+ crowdsourcing assessors, making it, to the best of our knowledge, the largest dataset in this field. We make OpenSAL360 publicly available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2609.21480
PDF
-
NaViRrator: Robot Navigation from Human-Readable Maps through a Learned Visual Route
2026-09-18 04:48:37Ayun Lee, Jiseon Kim, Giseop KimAbstract
Human-readable maps provide an intuitive interface for specifying robot destinations, but connecting their schematic geometry to egocentric observations remains challenging. We present NaViRRator, a framework that translates user-specified start and goal locations on such maps into navigation instructions for a pretrained vision-and-language navigation (VLN) policy. Its core method, RouteScribe, separates route inference from verbalization by first generating an explicit route scaffold in map-image coordinates, which a pretrained vision-language model (VLM) converts into a navigation instruction. We construct the scaffold with start--goal line conditional flow matching (SGL-CFM), which deforms a straight start--goal waypoint sequence into a map-conditioned route. During execution, the VLN policy receives only the instruction and egocentric observations, while the map and scaffold remain upstream, allowing executor replacement without retraining the map-to-language modules. Real-world experiments show higher success rates and success weighted by path length (SPL) than direct map-to-instruction generation, A*-based scaffolding, and Gaussian-source conditional flow matching. Qualitative results further show clearer salient turns and better preservation of the intended maneuver sequence, supporting route-grounded language as a modular interface between human-readable maps and pretrained navigation policies.
Abstract (translated)
URL
https://arxiv.org/abs/2609.21316
PDF
-
Workspace Models: Lightweight Robotic Memory via Saliency-Driven Supervision
2026-09-17 17:59:53Nitish Dashora, Douglas Chen, Idan Shenfeld, John Marangola, Pulkit Agrawal, Max SimchowitzAbstract
Complex robotic manipulation tasks frequently require a long-term memory of past events and actions. As conditioning on full histories renders policies prone to spurious correlations and degrades performance, many approaches to policy memory involve compressing historical information through expensive VLM queries in-the-loop to process only task-salient information. In this paper, we propose an alternative approach in which computationally intensive VLM queries are made during train-time to learn a lightweight latent memory that can be efficiently queried at deployment time. Our representation, which we call the \textbf{workspace token}, is trained by (1) using a VLM to identify current and historical information necessary for completing a task, then (2) distilling these into the workspace token using a set-reconstruction decoder loss. In both simulation and hardware, we show that the workspace token can be used as a drop-in replacement for observations during deployment, enabling policies to solve memory-intensive tasks without the need for VLM reasoning in-the-loop. Interestingly, we found that workspace tokens are not only more lightweight but also lead to better policy performance.
Abstract (translated)
URL
https://arxiv.org/abs/2609.20820
PDF
-
Deep Learning-Based Classification of Cognitive and Resting States Using Electroencephalography Signals
2026-09-17 14:26:12K. A. Januka S. Fernando, Harshit SrivastavaAbstract
The categorization of cognitive and resting states derived from electroencephalography (EEG) signals is crucial for comprehending fluctuations in brain activity linked to various mental states. EEG provides a non-intrusive approach for documenting brain function in both resting and task-oriented cognitive conditions, whilst deep learning techniques enable the automatic extraction of significant patterns from intricate EEG data. This study presents a deep learning framework to distinguish between resting and cognitive states through EEG records. The proposed framework integrates a Convolutional Neural Network (CNN) stacked with a Gated Recurrent Unit (GRU) for the extraction of features from EEG signals. Time-frequency analysis is conducted to explore the salient aspects of signals, and the derived features are then assessed utilizing conventional deep learning and machine learning classifiers, including the suggested 2D-Net architecture. The proposed approach and feature extraction strategy outperform the evaluated comparative methods, achieving accuracies of 83.177% for resting-versus-mathematical task classification, 76.107% for resting-versus-memory task classification, and 83.432% for resting-versus-music task classification. The findings illustrate the efficacy of integrating signal processing with deep learning methodologies to discriminate resting from cognitive states utilizing EEG signals.
Abstract (translated)
URL
https://arxiv.org/abs/2609.20467
PDF
-
GAPrompt++: Multi-Granular Geometry-Aware Point Cloud Prompt for 3D Vision Model
2026-09-17 05:14:34Zixiang Ai, Zhenyu Cui, Yufei Guo, Wenwen Qiang, Lei Chen, Jiwen Lu, Jiahuan ZhouAbstract
Pre-trained 3D vision models have substantially advanced point cloud analysis, yet adapting them to downstream tasks via full fine-tuning is computationally expensive and storage-intensive. Parameter-Efficient Fine-Tuning (PEFT) offers a promising alternative by reducing both adaptation cost and storage burden. However, existing prompting-based approaches ignore the intrinsic geometric structures of point clouds, thereby limiting their adaptation capability. This limitation stems from their inability to encode both fine-grained geometric cues and coarse-grained structural semantics, as well as failing to propagate such information effectively through the model hierarchy. To address these challenges, we propose GAPrompt++, a multi-granular geometry-aware prompting method that provides richer geometric guidance for efficient 3D task adaptation. Specifically, we introduce a Point Shift Prompter that extracts multi-granular geometric features across different scales, enabling instance-specific geometric adjustments during adaptation. Next, a Keypoint Prompter adaptively generates point-level prompts to highlight local geometric saliency and fine-grained structural details. Furthermore, a Prompt Propagation mechanism injects these multi-granular geometric cues throughout the feature extraction hierarchy, strengthening the ability to capture essential geometric characteristics. Extensive experiments show that GAPrompt++ achieves state-of-the-art performance among prompting-based PEFT methods and even surpasses full fine-tuning across diverse benchmarks, while requiring less than 2\% trainable parameters. In addition, to address the saturation of existing evaluation datasets, we construct two more challenging benchmarks derived from 3D Gaussian Splatting and Multi-View Stereo reconstruction, offering diverse and realistic point cloud scenarios to promote future research.
Abstract (translated)
URL
https://arxiv.org/abs/2609.19716
PDF
-
KINO: A Keyframe Interface for VLM Planning and Whole-Body Control in Humanoid Loco-Manipulation
2026-09-16 16:07:33Sitong Chen, Fatemeh Zargarbashi, Jin Cheng, Tianxu An, Stelian CorosAbstract
Humanoid loco-manipulation requires robots to interpret task instructions and scene semantics while executing coordinated whole-body motions. We propose a hierarchical framework that uses motion keyframes as an intermediate representation between Vision-Language Model (VLM) planning and Reinforcement Learning (RL) control. Each keyframe specifies a target whole-body robot pose and, when applicable, an object pose. Given a language instruction, scene observations, and execution feedback, the VLM selects successive task-relevant keyframes from a predefined library. The selected keyframes are retargeted to the current scene to account for object poses and dimensions. A keyframe-conditioned whole-body policy then generates joint-level actions to reach these goals. We introduce a saliency-based keyframe sampling strategy for low-level policy training that improves end-to-end task success rate from 44% to 92% when using sparse VLM keyframes. We evaluate our framework on object pickup, transport, and placement tasks in simulation and on a Unitree G1 humanoid. The system successfully performs both one- and two-handed manipulation and generalises to placement locations beyond the training reference data.
Abstract (translated)
URL
https://arxiv.org/abs/2609.18869
PDF
-
GenStream: Semantic Streaming Framework for Generative Reconstruction of Human-centric Media
2026-09-16 13:20:36Emanuele Artioli, Daniele Lorenzi, Shivi Vats, Farzad Tashtarian, Christian TimmererAbstract
Video streaming dominates global internet traffic, yet conventional pipelines remain inefficient for structured, human-centric content such as sports, performance, or interactive media. Standard codecs re-encode entire frames, foreground and background alike, treating all pixels uniformly and ignoring the semantic structure of the scene. This leads to significant bandwidth waste, particularly in scenarios where backgrounds are static and motion is constrained to a few salient actors. We introduce GenStream, a semantic streaming framework that replaces dense video frames with compact, structured metadata. Instead of transmitting pixels, GenStream encodes each scene as a combination of skeletal keypoints, camera viewpoint parameters, and a static 3D background model. These elements are transmitted to the client, where a generative model reconstructs photorealistic human figures and composites them into the 3D scene from the original viewpoint. This paradigm enables extreme compression, achieving over 99.9% bandwidth reduction compared to HEVC for the continuous data stream. We partially validate GenStream on Olympic figure skating footage and demonstrate potential for high perceptual fidelity under minimal data. While acknowledging the significant computational costs shifted to the client and challenges in generalization, GenStream opens new directions in volumetric avatar synthesis, canonical 3D actor fusion across views, and personalized viewing experiences, laying the groundwork for scalable, intelligent streaming in the post-codec era.
Abstract (translated)
URL
https://arxiv.org/abs/2609.18634
PDF
-
Interpretable Patch-Based Deep Learning for Wildfire Spread Prediction from Ensemble Simulations
2026-09-16 12:17:47Marcin Lawenda, Aleksandra Krasicka, David Caballero, Luis Torres, {\L}ukasz SzustakAbstract
Wildfire spread is traditionally predicted using physics-based simulators, which are physically interpretable but whose cost increases with each additional ensemble member. We ask how well deep learning surrogates can reproduce these simulations at a fraction of this cost, training them on 10,584 fire spread simulations at 2m resolution for the Rectoret region in Catalonia, Spain. Four architectures are compared: a patch-based U-Net, a transfer-learned ResNet-50, a physics-informed network constrained by the wind-driven advection equation and a Swin-Unet transformer. Among the terrain and vegetation variables, only surface fuel load predicts burn probability with any strength (r = 0.27) and including it lowers prediction error by 21%. The remaining variables correlate weakly and are highly duplicative. Next, an experiment with saliency, occlusion and rotation demonstrates the models' learning. Convolutional models rely primarily on distance from the current fire front, while Swin-Unet assigns more weight to fuel and terrain, a finding also noted in an unrelated wildfire dataset. When applied without retraining to the second region, Pedriza, all three convolutional models still predict fire spread, losing accuracy by a small but systematic margin.
Abstract (translated)
URL
https://arxiv.org/abs/2609.18555
PDF
-
Det-LIME: Detector-Aware, Multi-Instance Local Interpretable Model-Agnostic Explanations for Automated Marine Mammal Detection
2026-09-15 17:20:16Jiayi Zhou, David W. Johnston, Brinnae BentAbstract
Despite the rapid uptake of black-box object detectors in marine mammal research and monitoring, explainability techniques are rarely integrated into conservation workflows. Furthermore, most classification-oriented explainability tools are ill-suited to detection tasks involving imagery of social organisms or those with colonial life histories, as they ignore multiple detections within a scene and produce single-instance outputs that blur evidence across individuals. These methods also generate low-resolution, often biologically irrelevant visuals, limiting their utility for debugging, targeted data augmentation, and refined data collection. We proposed Det-LIME, a detector-aware, multi-instance adaptation of Local Interpretable Model-Agnostic Explanations (LIME) that produced instance-specific, box-aligned explanations by combining per-detection weighting, a proximity kernel that emphasizes regions near each box, and Intersection-over-Union-based matching to track the same instance across perturbations. We evaluated Det-LIME on aerial drone imagery for harbor seal detection, with an additional seabird case study to assess generality, and compared it with vanilla LIME, Stabilized LIME, Deterministic LIME, and gradient-based attribution methods. Using the Attribution Ratio and Max Saliency Hit Rate metrics, we showed that Det-LIME consistently improved multi-instance attribution. In practice, these higher-resolution, instance-aware explanations provide insight into model outputs and support post-processing, debugging, and actionable improvements in modeling and data collection or augmentation.
Abstract (translated)
URL
https://arxiv.org/abs/2609.17479
PDF
-
Event-based Selective Attention for Multi-resolution Fast Region of Interest Detection
2026-09-15 13:02:14Luca Peres, Giulia D'Angelo, Chiara Bartolozzi, Oliver RhodesAbstract
Neuromorphic vision systems operate under strict constraints on bandwidth, memory, and energy, particularly at the edge, motivating early mechanisms for data reduction and selective processing. In this work, we investigate a multi-scale training-free, saliency-based, bottom-up visual attention model that operates directly on low-resolution event-based input and selects Regions of Interest (ROI) from the visual scene. The model is evaluated across multiple downscaling factors applied to the incoming event stream, with input resolutions reduced by up to 256x relative to full resolution. Performance is assessed on the Prophesee Automotive dataset, the largest publicly available event-based dataset, demonstrating robust ROI selection across different scales on a real-world use-case. The proposed approach is capable of detecting ROIs belonging to multiple object classes, including various vehicle types, pedestrians, traffic lights, and traffic signs, with accuracy up to 70.8%, while operating at millisecond temporal resolution, 16x finer than the temporal resolution provided by the dataset ground truth. These results highlight the potential of combining early event downscaling with saliency-based attention as an effective front-end for efficient edge neuromorphic vision systems.
Abstract (translated)
URL
https://arxiv.org/abs/2609.17134
PDF
-
High-Fidelity Video Quality Assessment with VQA-Specific Saliency
2026-09-15 10:23:05Hakan Emre Gedik, Shashank Gupta, Alan BovikAbstract
No-reference video quality assessment (NR VQA) has recently seen promising progress with deep learning. However, video data is inherently large, and processing them with deep models incurs high computational cost. This challenge is particularly acute in VQA, where preserving original-resolution cues and dense temporal information is critical for accuracy. Existing efficiency-driven preprocessing strategies, such as fragmenting, reduce computation but alter the input data distribution, limiting effective reuse of pretrained video foundation models (ViFMs). To address these challenges, we propose \textbf{H}igh-\textbf{F}idelity \textbf{V}ideo \textbf{Q}uality \textbf{A}ssessment (\textbf{HFVQA}), a framework built on fixed-size spatio-temporal (ST) patches that is fully compatible with pretrained ViFMs. HFVQA samples ST patches across multiple scales, including the original resolution, with minimal temporal subsampling to preserve low-level quality cues and semantic context. To limit computation, HFVQA introduces a lightweight auxiliary network trained end-to-end with the ViFM encoder to learn \textit{VQA-specific saliency}. Distilled directly from quality supervision, this saliency captures task-specific importance patterns, reflecting that video quality perception is dominated by a small subset of spatio-temporal regions. By combining high-fidelity spatio-temporal cues with learned, task-specific saliency, HFVQA achieves SOTA performance on standard NR VQA benchmarks while processing as little as 12\% of candidate ST patches, making high-fidelity ViFM-based VQA computationally tractable.
Abstract (translated)
URL
https://arxiv.org/abs/2609.16946
PDF
-
FSANet: Frequency-Spatial Aware Network for Image Segmentation
2026-09-15 07:43:08Ruibo Wang, Ziyi Shen, Huaming Wu, Dong Liang, Kun ShangAbstract
Image segmentation remains challenging due to occlusions, poor lighting, and irregular structures. Although transformer-based methods achieve high accuracy, they rely heavily on long-range spatial features, leading to high computational costs and neglecting prior knowledge or noise patterns, resulting in missing details and unclear boundaries. To address these issues, we propose Frequency Spatial Aware Network (FSANet), which integrates prior knowledge with a dual-domain solver to sequentially adapt to diverse segmentation tasks. Specifically, we design three key modules: (1) Structure Prior Module, which recovers overlooked details; (2) Dual-Domain Awareness Module, which captures salient features while disentangling noise; and (3) Edge Estimation Module, which enhances edge awareness for more precise segmentation. In addition, the limited availability of comprehensive segmentation datasets covering various real-world scenarios hinders the performance of existing methods. To address this, we introduce SceneX, a novel open-source dataset featuring 10 challenging non-ideal scenarios, establishing a new benchmark for evaluating and improving the robustness and real-world applicability of the segmentation models. Extensive experiments demonstrate the efficiency and effectiveness of FSANet.
Abstract (translated)
URL
https://arxiv.org/abs/2609.16773
PDF
-
MAETrack: Unleashing the Potential of Pretrained Geometric Priors for 3D Single Object Tracking
2026-09-15 06:21:44Sifan Zhou, Qiwei Wang, Linyue Tan, Ziyu Liu, Ziyu Zhao, Xiaobo LuAbstract
Large-scale pre-training has transformed representation learning in 2D vision, yet its transferability to 3D single object tracking (SOT) remains insufficiently understood. Directly fine-tuning self-supervised 3D encoders, such as masked autoencoders (MAE), often leads to sub-optimal adaptation because the reconstruction objective is not fully aligned with the spatial-temporal matching requirements of tracking. In this paper, we observe that this difficulty can be interpreted as a layer-wise transfer mismatch: shallow layers tend to preserve transferable geometric cues, while deeper layers become increasingly specialized to the reconstruction pretext task and are less suitable for downstream tracking. Based on this observation, we propose MAETrack, a lightweight adaptation framework for transferring pre-training MAE representations to 3D SOT. MAETrack includes Layer-Selective Initialization (LSI), which initializes only the shallow stages of the tracking backbone from pre-trained weights while re-initializing deeper stages, and Geometric Residual Gating (GRG), which reinforces structurally salient regions in the search BEV features before template-search fusion through residual spatial modulation. Extensive experiments on standard 3D SOT benchmarks show that MAETrack consistently improves upon vanilla fine-tuning baselines with limited computational overhead. More broadly, our results suggest that effective transfer from 3D reconstruction pre-training to 3D tracking is not merely a matter of partial fine-tuning, but depends on a tracking-oriented transfer principle that preserves shallow geometry while adapting deeper representations to the downstream objective.
Abstract (translated)
URL
https://arxiv.org/abs/2609.16695
PDF
-
Look Before You Leap: Factual Decoding with Internal Attribution Signals
2026-09-14 15:35:42Hayeong Ryu, JungMin Yun, Byeonggeuk Lim, Sunhee Jo, YoungBin KimAbstract
Hallucination remains a critical challenge in large language models (LLMs), where early factual errors compound through autoregressive generation in a snowballing effect that neither post-hoc correction nor weight-level intervention can effectively preempt. We propose DescaPE (DEcoding Signal Control Against Path Error-snowballing), a decoding framework that leverages internal model signals to suppress hallucination-prone trajectories at inference time. Through sliding-window MLP ablation, we identify a factual-salient layer span within LLMs whose derived signal is selectively elevated for factual tokens and exhibits anomalous spikes at hallucination-prone steps. We train a lightweight probe to approximate this signal from a single forward pass and integrate it into candidate scoring to penalize high-risk continuations while rewarding factually grounded ones. Experiments across five factuality benchmarks on three LLMs demonstrate that DescaPE achieves factuality improvements over decoding-time baselines in multiple settings, while incurring only 1.10x latency overhead in our efficiency evaluation. Our code is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2609.15745
PDF
-
Don't Send What You Don't Need: Question-Guided Token Pruning as a Privacy Defense for Vision-Language Models
2026-09-14 14:48:39Md Khalid Syfullah, Alvi Ataur KhalilAbstract
Visual Question Answering (VQA) with Vision-Language Models (VLMs) is increasingly used in privacy-sensitive and bandwidth-constrained settings. Federated Learning (FL), Split Learning (SL), and U-Shaped Split Learning (USL) keep raw data local, but transmitting all visual tokens across a model partition remains costly and can expose private information. We propose QPriv-VL, a question-guided, privacy-aware token-pruning framework for FL, SL, and USL that prunes visual tokens before transmission based on task utility and privacy sensitivity. Its core component is a lightweight Dynamic Threshold Predictor (DTP) that jointly estimates a sample-specific pruning ratio and a token-level retention mask in one forward pass. DTP combines question relevance, computed from cross-modal similarity between visual patches and the pooled question embedding, with a sensitivity signal derived from frozen DINOv2 features. This allows the model to suppress potentially sensitive regions while preserving patches useful for answering the question, without requiring sensitivity labels. We evaluate QPriv-VL on GQA, OK-VQA, VQAv2, SLAKE, VQA-RAD, and PathVQA against four privacy attack families: FSHA, FORA, iDLG, and attribute-inference membership inference attacks. DTP matches or outperforms fixed-ratio pruning while using substantially fewer transmitted tokens. On VQA-RAD, it reduces membership-inference attack success from 0.99 to 0.76-0.79, lowers FSHA and FORA reconstruction PSNR relative to fixed-ratio pruning, and preserves competitive VQA accuracy using about 40% of the original visual-token budget. A sensitivity exclusion ratio of 1.20 +/- 0.18 indicates preferential removal of privacy-sensitive patches, while explainability analysis shows that retention adapts to question semantics rather than generic visual saliency.
Abstract (translated)
URL
https://arxiv.org/abs/2609.15671
PDF
-
From Model Patterns to Abstract Semantics in Compositional Zero-Shot Learning
2026-09-14 14:32:56Weize Li, Zhicheng Zhao, Fei SuAbstract
Compositional Zero Shot Learning aims to recognize unseen compositions by recombining learned primitives. Recent methods rely on vision language models and attempt to explicitly model contextual variations of primitives through multiple representations. However, such approaches are limited by fixed variant capacity and competition between abstract and concrete semantics. In this work, we present a new perspective that views primitive variations as the context-driven activation of concrete visual cues rather than independent entities. Based on it, we propose CLEAR, a CLoze-style rEAsoning-based Re-ranking framework inspired by human perceptual processes. CLEAR extracts conditional variants from the primitive candidate set in a coarse-to-fine manner, performs cloze-style reasoning to infer high-level semantics, and re-ranks predictions to correct biases toward salient concrete primitives. Extensive experiments demonstrate that CLEAR consistently improves the Base Model and outperforms state-of-the-art methods on the challenging C-GQA and MIT-States datasets. Code is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2609.15649
PDF
