Image_Caption
-
GARD: Gradient-Guided Adversarial Representation Defense against Backdoors in Multimodal Contrastive Learning Models
2026-11-10 14:30:36Yuhong Wang, Yujun Cai, Lei Feng, Songze LiAbstract
Multimodal contrastive learning models such as CLIP are vulnerable to backdoor attacks with diverse trigger structures, ranging from localized patches to globally distributed perturbations, with recent attacks increasingly optimizing these triggers in the representation space. Existing post-hoc defenses based on clean fine-tuning may leave backdoors active, while trigger-inversion methods depend on accurately estimating the attack target and trigger. We observe that backdoored encoders exhibit structure-dependent representation sensitivity: global perturbations preferentially expose distributed backdoors, whereas local patch perturbations are more effective for localized backdoors. Based on this observation, we propose Gradient-Guided Adversarial Representation Defense (GARD), a target- and trigger-agnostic post-hoc defense. GARD optimizes complementary global and local adversarial probes to maximally displace clean-image representations, and then realigns the perturbed representations with detached clean anchors while preserving image-text alignment. The resulting model requires no additional computation at inference time. Across six visual backdoor attacks and two CLIP visual backbones, GARD reduces the attack success rate below 1% in all 12 evaluated attack-backbone combinations while maintaining competitive clean accuracy. These results demonstrate the effectiveness of adversarial representation probing for repairing backdoored multimodal contrastive encoders.
Abstract (translated)
URL
https://arxiv.org/abs/2611.14143
PDF
-
Marrying Capacity and Precision for Efficient 2D Gaussian Image Representation
2026-11-10 09:21:35Lin Dong, Qi Ming, Liuqian Wang, Wentao Chao, Yufei Gao, Chengming Liu, Lei ShiAbstract
2D Gaussian representations provide an explicit and efficient way to represent images, but their rate-distortion performance under limited budgets depends critically on how Gaussian capacity and parameter precision are allocated. Existing methods typically optimize these two resources separately, overlooking content-dependent capacity demands and heterogeneous quantization sensitivity. We present CP-GS, a capacity-precision coupled Gaussian splatting framework for budget-aware 2D Gaussian image representation that jointly optimizes capacity allocation and attribute precision under a shared rate constraint. CP-GS first constructs a content-adaptive Gaussian representation by dynamically reallocating a fixed Gaussian budget from redundant regions to areas with higher representation demand. It then introduces hierarchical distortion-aware representation quantization, using efficient rendering-based calibration to allocate precision across Gaussian attributes and further refine it within heterogeneous attribute groups. Finally, it jointly optimizes Gaussian capacity and attribute precision by constructing candidate representations with different Gaussian counts and precision configurations and selecting operating points according to decoded distortion and actual serialized size. Experiments on DIV2K and Kodak show that CP-GS achieves improved rate-distortion performance over representative 2D Gaussian and implicit neural representations across a range of budgets, while maintaining a fixed overall representation budget and low optimization cost. These results demonstrate the benefit of jointly optimizing spatial capacity and parameter precision for compact Gaussian image representation.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13784
PDF
-
Repository-Context-Guided Image Understanding for Agentic Software Development
2026-11-09 08:33:52Fumiki Yohena, Toshiaki Sota, Shota Suzuki, Tomoko Okamoto, Michiaki TatsuboriAbstract
Coding agents increasingly receive UI screenshots as part of bug reports, yet degraded image quality - caused by low resolution, compression artifacts, or small rendered text - can prevent them from identifying the relevant UI elements and source files. Vision-language models become less reliable as image quality deteriorates, while OCR-assisted approaches may propagate recognition errors from degraded inputs. Both approaches are limited when the visual evidence required for localization is no longer clearly observable in the query screenshot. We propose augmenting coding agents with repository-rendered reference images: crisp screenshots generated from the same codebase that the agent is tasked with modifying. These references provide readable visual evidence of the application's UI while preserving the current, potentially bug-containing implementation state. Because they are generated from the available repository rather than requiring a known bug-free version, they can be produced automatically for reproducible application states without manual image annotation. We evaluate the approach on a controlled benchmark derived from SWE-bench Multimodal by injecting visual UI bugs into a real web application. We vary screenshot degradation and reference-image provision and measure whether the coding agent identifies the correct source file. The results show that reference images improve file localization especially for degraded image quality, increasing the gold file match rate under fourfold downsampling with JPEG compression. We confirmed with further analysis that the proposed approach recognizes text approximately 12 percentage points better than language biased OCR under the same degradation.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13026
PDF
-
Attribute-Evidence Learning with Multimodal Prompt for Generalized Category Discovery
2026-11-08 04:01:26Wei Zhang, Zhu Teng, Zhixiang Chi, Junnan Zou, Wenxin Luo, Linfeng Ye, Jianping Fan, Konstantions N. PlataniotisAbstract
Generalized category discovery (GCD) aims to discover novel categories from unlabeled data by leveraging limited labeled samples from old classes. Existing methods that rely solely on visual cues often neglect the multimodal nature of human perception, prompting recent approaches to incorporate textual information such as category tags and attributes. However, these approaches typically entangle global tags and local attributes within a shared global image representation, rendering attributes unreliable and potentially providing misleading cues that can drive the model toward incorrect predictions. To this end, we propose AptM-GCD, a novel multimodal framework that exploits apt textual guidance through attribute evidence learning and multimodal prompt reasoning. Specifically, we design an Attribute-Evidence Learning (AEL) mechanism that treats attributes as localized visual evidence rather than global category descriptors. By interacting attribute embeddings with patch-level visual tokens, AEL identifies reliable attribute-related regions and extracts discriminative local evidence, thereby reducing the impact of noisy or irrelevant attribute cues. Building upon the extracted evidence, we further develop a Multimodal Prompt Learning (MPL) mechanism to bridge local attribute semantics and global visual understanding. MPL converts fine-grained textual evidence into visual prompts and injects them into a multimodal prompt transformer, enabling attribute information to enhance category discrimination through contextual reasoning without enforcing direct global alignment between modalities. Extensive experiments on six datasets demonstrate that AptM-GCD consistently achieves SoTA performance across both fine-grained and generic benchmarks. Moreover, the proposed framework exhibits strong generalization ability, delivering consistent gains under unimodal and class-imbalanced GCD settings.
Abstract (translated)
URL
https://arxiv.org/abs/2611.12975
PDF
-
Multimodal Training with Knowledge Traction for Video Anomaly Detection
2026-11-06 14:50:13Xiao TanAbstract
The task of weakly supervised video anomaly detection is challenging. Existing approaches mainly model video frame features and determine whether anomalies are present based on their magnitudes. In this work, we propose a novel method that incorporates external knowledge into training to improve model performance, robustness, and generalization. We leverage textual prompts and a large language model (LLM) to establish prior information for the specific task, and employ dictionary learning to normalize and compress this information. Moreover, during experiments, we observed inconsistencies in the textual descriptions extracted from similar video frames by image caption models. To address this, we devise a method that leverages an LLM and auxiliary information to ensure consistency in the textual descriptions of similar frames within a video. Finally, we introduce a feature processing approach that incorporates prior knowledge and cross-modal attention mechanisms. This approach learns the relationships and characteristics among different modalities to construct more accurate and reliable feature representations, thereby improving the accuracy and stability of downstream tasks. Experimental results demonstrate that our method surpasses other state-of-the-art approaches.
Abstract (translated)
URL
https://arxiv.org/abs/2611.09536
PDF
-
BEiTScore++: Reference-free Image Captioning Evaluation with BEiT-3 and Conformalized Quantile Regression
2026-11-06 14:23:34Gon\c{c}alo Gomes, Bruno Martins, Chrysoula ZervaAbstract
Reliable uncertainty estimation is important for deploying evaluation metrics in real-world settings, where both input ambiguity and domain shifts are common. In this work, we propose an uncertainty-aware extension of the BEiTScore image captioning metric, producing conformally calibrated confidence intervals alongside alignment scores. Our approach models predictive distributions over human judgments, enabling both accurate scoring and principled uncertainty quantification in a single forward pass. We evaluate our BEiTScore++ metric across multiple benchmarks with human-annotated ratings, showing that it preserves the ranking correlation of BEiTScore, while producing intervals whose width tracks the prediction error more closely than Monte Carlo dropout on most benchmarks. The estimated uncertainty also grows under image degradation and on out-of-domain artwork images, making BEiTScore++ a practical tool for uncertainty-aware evaluation of image-caption alignment.
Abstract (translated)
URL
https://arxiv.org/abs/2611.09505
PDF
-
OpenView: Empowering MLLMs with Out-of-view VQA
2026-11-06 06:36:32Qixiang Chen, Cheng Zhang, Chi-Wing Fu, Jingwen Ye, Jianfei CaiAbstract
Recent multimodal large language models (MLLMs) show great potential in natural image understanding. Yet, they perform well, mainly on reasoning in-view contents within the image frame. This paper presents the first study on out-of-view (OOV) understanding, i.e., the ability to reason about objects, activities, and scenes beyond the visible frame of a perspective view. Our technical contributions are threefold. First, we design OpenView, a four-stage pipeline that leverages panoramic imagery for large-scale and diverse multi-choice VQA synthesis, where full-scene coverage enables flexible view framing while preserving global scene awareness. Second, we curate OpenView-Dataset, a high-quality synthetic dataset from diverse real-world panoramas to empower MLLMs upon supervised fine-tuning. Third, we build OpenView-Bench, a benchmark that jointly measures choice and rationale accuracy for interpretable and diagnosable evaluation. Experimental results show that despite having a large gap from human performance in OOV VQA answer selection, multiple MLLMs, upon empowered by OpenView, can consistently boost their performance, uplifted from 48.3% to 62.3% on average. Code, benchmark, and data are available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2512.18563
PDF
-
UniMore: Fine-Grained Supervision for Unified Understanding and Generation
2026-11-06 05:11:14Haoyue Bai, Yinyi Luo, Yicheng He, Qingsong Wen, Sharon Li, Jindong WangAbstract
Unified multimodal models combine image understanding and generation within one system, creating an opportunity to improve both capabilities through paired image--text supervision. We examine the outputs of a pretrained unified model in both directions and observe a loss of specificity: generated images can contain the requested objects with incorrect attributes, while captions identify a scene without expressing its distinguishing details. These failures call for supervision that specifies both the visual content to realize and the descriptive differences to distinguish. We propose UniMore, a post-training method that derives these complementary learning tasks from the fine-grained content of image--text pairs. The generation route predicts grounded crops from their corresponding phrases, matching the scope of the condition and visual target while using the native visual objective. The understanding route combines positive-caption fitting with ranking against captions that omit or alter details, conditioned on the full image. The two routes are optimized independently, and their learned updates are combined into a single unified model without changing its inference procedure. Experiments show gains across multiple understanding and generation benchmarks.
Abstract (translated)
URL
https://arxiv.org/abs/2611.08996
PDF
-
VoluCompose: Constructing Compact Volumetric Visual Representations for Medical Image Understanding
2026-11-05 19:53:54Zanting Ye, Yunrui Lin, Shengyuan Liu, Jing Hao, Chen Hu, Yefeng Zheng, Lijun LuAbstract
Medical vision--language models can interpret volumetric images, but processing full volumes is costly because hundreds of slices produce long, redundant visual-token sequences. Existing compression methods often couple content selection with budget determination. We observe that neighboring slices share visual features and that attention concentration varies across visual encoders. We propose VoluCompose, a training-free, model-agnostic framework that separates what visual content to retain from how many tokens to allocate. Its hyperbolic token composition module groups contiguous slices into ordered support units and uses the Lorentz distance to account for feature magnitude and direction during token allocation, anchor selection, and refinement. Its capacity control module uses visual-encoder attention entropy to allocate more tokens when attention is more dispersed. We evaluate VoluCompose on two tasks across three datasets and four backbones. Across nine calibrated settings, VoluCompose attains the highest mean quality--retention hypervolume among the compared methods. It achieves a mean relative quality of 121.68\% against the uncompressed references while retaining 22.78% of visual tokens on average. Compression remains robust even at a 10% token budget. Across settings, the lowest ROUGE-L ratio to the uncompressed reference is 96.4\% for VoluCompose. The corresponding minima for the compression baselines range from 15.0% to 70.5%. Code is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2611.08649
PDF
-
Aligning Policy Encoders with a Shape Teacher
2026-11-05 16:44:24Madelon Bernardy, Mohammad Mahdi Derakhshani, David M Knigge, Riccardo Valperga, Artem MoskalevAbstract
3D-aware representations can improve robotic manipulation by providing geometric information that is robust to changes in viewpoint, object position, and occlusion. However, using 3D inputs at deployment requires additional sensing and processing, making image-based policies more practical. We ask whether an image-based policy can acquire 3D shape information during training while retaining a purely image-based interface at deployment. We propose shape-teacher alignment, a training-time mechanism that uses a pretrained 3D shape representation as an alignment target for an image-based policy encoder. Specifically, we use Point-MAE features from individual objects as targets for corresponding image patches, transferring geometric structure to the policy without changing its deployment inputs or architecture. On LIBERO, shape-teacher alignment produces image representations that are more structured around object shape and corresponding parts, while substantially accelerating policy convergence and improving success rates. For pretrained policies, convergence is accelerated by 4.7x and 5.6x on the Object and Spatial suites, respectively, without requiring the shape teacher or depth input at deployment.
Abstract (translated)
URL
https://arxiv.org/abs/2611.08275
PDF
-
Discovering Semantic Granularity in Images through Clustering of Pretrained Visual Representations
Abstract
Semantic granularity, which refers to the levels at which visual representations organize image content into meaningful entities, is fundamental to image understanding. However, whether pretrained visual representations inherently encode semantic structures at multiple granularities remains poorly understood. This paper presents a representation-driven approach for discovering semantically meaningful image regions at different granularities through clustering of pretrained visual representations without semantic supervision. Specifically, we concatenate selected feature maps extracted from a pretrained visual model and apply a singular value decomposition (SVD)-based projection to derive pixel-wise principal representation map tensor (PRMs). The resulting PRMs captures latent semantic structures embedded in pretrained visual representations, making them accessible to subsequent semantic clustering. We then introduce the semantic continuity rate (SCR), based on the silhouette principle, to quantify the clustering behavior of PRMs at different candidate granularities. Furthermore, we analyze the signal-to-noise ratio (SNR) of SCR profiles to assess the PRMs suitability for discovering semantic-like meaningful regions at multiple granularities. Our results indicate that PRMs exhibiting a higher proportion of relatively high SNR values over an initial range of cluster numbers tend to yield more semantically meaningful regions across multiple granularities. We evaluate the proposed approach on the COCO-2017 and VOC-2012 validation datasets using Hungarian-matched intersection over union (HIoU). The results demonstrate competitive segmentation performance among the compared methods, supporting the feasibility of discovering useful semantic granularities through clustering PRMs without semantic supervision.
Abstract (translated)
URL
https://arxiv.org/abs/2611.02686
PDF
-
VINO: Revisiting 3D Representation Learning with Volume Transformers
2026-11-04 22:00:30Adrian Kruse, Kadir Yilmaz, Daan de Geus, Francis Engelmann, Bastian LeibeAbstract
Self-supervised learning has given rise to vision foundation models that learn general-purpose image representations transferable across downstream tasks, yet representation learning for 3D scenes has not achieved comparable success. Motivated by this, we identify key limitations of current 3D representation learning approaches and revisit their core design choices. We first replace the commonly used hierarchical PTv3 backbone with the Volume Transformer, a ViT-like architecture for 3D scenes. We then redesign two core objectives of modern approaches built around student-teacher networks: matching representations across views and predicting masked content from visible context. For the first objective, we separate the roles of teacher and student: the teacher provides clean, consistent targets while the student learns invariance to stronger perturbations. For the second objective, we prevent the student from accessing the masked geometry and instead require it to infer missing representations from visible context only. Finally, we redesign 2D-to-3D knowledge distillation, replacing direct regression of view-dependent image features with view-consistent supervision. Together, these changes form VINO, a framework that learns semantic representations of 3D scenes without labels. VINO substantially improves dense linear probing across indoor and outdoor benchmarks and transfers strongly to out-of-domain settings.
Abstract (translated)
URL
https://arxiv.org/abs/2611.07176
PDF
-
Frequency-Adaptive Gaussian Grids for Efficient Image Representation
2026-10-30 14:22:36Anatolii Evdokimov, Stavros Diolatzis, Bingxuan Li, Akshay Jindal, Patsorn Sangkloy, Qi SunAbstract
Neural representation has emerged as a powerful tool to mitigate the demand for increasingly higher-resolution images. However, a fundamental trade-off remains between structured (e.g., feature grids, which offer spatial consistency) and unstructured (e.g., 2D Gaussians, which adaptively capture local variances) representations. This choice significantly affects how well an algorithm represents an image. Here, we present Gaussian-Grid, a hybrid neural representation that adaptively integrates structured and unstructured primitives. By dynamically partitioning computation between 2D Gaussian distributions and their attached multi-resolution feature grids based on the local frequency content, Gaussian-Grid automatically optimizes its architecture to match the underlying complexity of the target signal. We evaluate our approach using high-resolution natural images and demonstrate its applications as GPU-native dataset storage, showing $1.24$-$2.14\times$ decoding speed improvement and PCIe data traffic reduction of $23\times$-$176\times$ over JPEG + ZIP data archiving. The results show that Gaussian-Grid maintains small memory footprints while achieving high reconstruction fidelity across a variety of datasets.
Abstract (translated)
URL
https://arxiv.org/abs/2610.38007
PDF
-
Latent-Lens: Visual Perception in Small Language Models Through Latent Communication
2026-10-30 13:53:44Mo Shahdloo, Faegheh Sardari, Mathew SalvarisAbstract
Bridging independent vision-language models (VLMs) and small language models (SLMs) in a modular multi-agent system typically forces them to exchange information as discrete text, which creates a lossy bottleneck for spatial and fine-grained details. This constraint is especially acute for SLMs, whose limited capacity makes them more reliant on the quality of the input signal. Recent latent-communication methods exchange states between language models. How a VLM can pass visual information to a separately pretrained text-only SLM though remains unexplored. We propose \textbf{Latent-Lens}, a lightweight codec that bridges the representation gap between a frozen VLM and a frozen SLM directly in latent space, enabling them to communicate purely through hidden states. The codec consists of a domain encoder that projects hidden states from the VLM into the SLM's embedding space, and an attention decoder formalised as a LoRA that teaches the SLM to attend to these visual prefixes, while training only ${\sim}1.6$M parameters (0.2\% of the combined model size). On visual question answering (CLEVR, GQA), Latent-Lens achieves $81.0\pm7.9$\% and $65.4\pm5.4$\%, compared with $58.3\pm5.7$\% and $61.8\pm1.0$\% for text communication with the same LoRA budget (+30.2--69.8pp relative to frozen text communication), while on image captioning (Flickr8k) it improves over the same baseline by +16.5 (+18.8 frozen). We also evaluate Latent-Lens with TinyLlama-1.1B and InternVL2-1B, which use different architectures, vocabularies, and hidden dimensions.
Abstract (translated)
URL
https://arxiv.org/abs/2610.41751
PDF
-
CompoSET: A Controlled Single-Edit Benchmark for Vision-Language Compositionality
2026-10-30 01:18:34Laure Delisle, Pietro PeronaAbstract
Vision-language compositionality is the ability to identify which properties, relations, quantities, and states bind to which entities in both an image and a caption, and to match those bindings across modalities. Existing benchmarks evaluate this ability through image-text matching: a model ranks candidate image-caption match similarities and succeeds when the correct match comes on top. In such datasets, paired images differ by more than a single binding: scene, layout, and content all differ, with 97% of pixels changing on average. Such pairs do not isolate compositionality: results are confounded by global image differences. We introduce Compositional Single-EdiT (CompoSET), a benchmark dataset in which each base image-caption pair has an edited counterpart, or foil, that preserves the scene but differs by a single controlled edit, such as color, count, spatial relation, or object state. Base and foil images preserve naturalistic scene complexity yet are visually near-identical: only 8% of pixels change on average. Our dataset has many foils per base scene (~22) in order to separate scene difficulty from edit difficulty, a distinction one-foil-per-base benchmarks cannot make. CompoSET contains 1,776 instances across 80 naturalistic scenes, spanning 16 edit types in 4 semantic categories, with captions at 3 verbosity levels. We test 22 dual-encoder VLMs and find that all models perform at or below chance level on CompoSET, despite scoring well on prior benchmarks. This suggests that apparent progress under prior image-text matching protocols may be an artifact arising from confounds in benchmark design, and that compositional understanding in dual-encoder VLMs is far from solved. CompoSET addresses the confounds and offers a better controlled benchmark to measure future progress.
Abstract (translated)
URL
https://arxiv.org/abs/2610.41118
PDF
-
Proving and Improving the Platonic Representation Hypothesis with JEPAs
2026-10-29 03:49:53Randall Balestriero, Yann LeCunAbstract
Whether models trained on different modalities converge to the same representation of the world is central to AI. The Platonic Representation Hypothesis (PRH) conjectures that they do as they scale, but without a proof it offers neither a guarantee nor a recipe. We turn the hypothesis into a theorem. First, we prove a \emph{Platonic Representation Theorem}: Joint-Embedding Predictive Architectures (JEPAs) trained separately on any number of modalities learn the same representation up to one rotation whenever their positive pairs are two observations of one scene--\emph{good views}--and their misalignment is bounded otherwise. Second, we show that next-token prediction is a JEPA with \emph{bad views}, parts of a single document, which explains why language models align with vision only through scale. Third, we validate the theory from 145k captions to the texts of 940M web images, and uncover a scaling law of alignment that predicts unseen runs within seed noise. Good views need 42 times fewer training samples than next-token prediction on the same corpus to reach the same alignment, and a 109M-parameter text JEPA aligns with DINOv2 and CLIP's image encoders better than every language model we evaluated, up to 72B parameters, and than CLIP's text encoder: next-token prediction is a remarkably inefficient route to platonic representations. As a showcase, one rotation turns separately trained encoders into an image captioner (CIDEr .838 on COCO). Even that rotation needs no pair: on the latest pair-free benchmark, our text JEPA beats every zero-shot accuracy reported with language models up to 8B parameters, e.g., 64\% against at most 56\% top-5 on ImageNet-100. Positive views, and not scale alone, are thus the lever to build platonic representations.
Abstract (translated)
URL
https://arxiv.org/abs/2610.39302
PDF
-
What Constitutes Food Similarity? Visual Representations and Retrieval in Indian Cuisine
2026-10-27 19:11:43Saransh Kumar Gupta, Akshadha Sharma, Lipika Dey, Partha Pratim Das, Ramesh JainAbstract
Food similarity underlies retrieval, recommendation, substitution, and other food computing tasks, yet food representations are typically evaluated by identity preservation. We ask what happens to food similarity beyond identity: how different visual representations organize foods into relational neighborhoods, what culinary relationships these neighborhoods reveal or obscure, and how consistently these structures emerge across representations. Indian cuisine provides a challenging setting in which visually similar foods may have different culinary identities, while foods within a culinary family can vary substantially in appearance. Using an expanded Khana dataset and seven pretrained visual representations spanning convolutional, transformer-based, self-supervised, and image-text aligned models, we evaluate image-to-image retrieval and image-to-dish neighborhoods, examining culinary-family structure, contrast cases, and cross-representation agreement. We also aggregate image representations to compare dish-level visual similarity across models and against recipe-derived similarity for the same dishes. Image-level neighborhoods vary substantially across representations, including similarities across culinary-family boundaries. At the dish level, aggregation reveals common visual structure despite local variation, but this structure corresponds less strongly with recipe-derived similarity. These findings show that recognition performance alone cannot characterize relational structure: food similarity depends on representation, modality, and abstraction level rather than being a fixed property of the food collection. Understanding these structures is therefore a prerequisite to combining multiple food information modalities into a coherent notion of similarity.
Abstract (translated)
URL
https://arxiv.org/abs/2610.37867
PDF
-
Honeycomb-Surfels: Compact Hexagonal Primitives for Efficient Scene Rendering and Reconstruction
2026-10-27 17:59:59Prakhar Kaushik, Soumava Paul, Alan YuilleAbstract
Novel view synthesis and 3D reconstruction have advanced rapidly with radiance fields and splatting, yet their memory and computational demands limit their use in resource-constrained settings. We introduce Hexels: novel planar hexagonal surface elements with a first-order spatial color field for efficient 3D scene and 2D image representation. Hexels exploit locally planar regions with smoothly varying appearance using fewer, larger primitives. Their hexagonal support provides an area- and boundary-efficient domain, while the local color field captures appearance variation within each primitive. Across 34 scenes, Hexels improve PSNR over memory-matched 2DGS by 4.53-7.38 dB across datasets while requiring lower training VRAM. We further ablate Hexels by showing that per primitive, Hexels outperform 2DGS by upto 4.74 dB while using lower training VRAM and having 2.31 times faster FPS. Even in unconstrained settings, Hexels offers complementary quality-memory trade-offs. For efficient 3D surface reconstruction, Hexels reduce error by 28.9% compared with memory-matched 2DGS across 15 DTU scans. On 6 Tanks and Temples surface scenes, Hexels achieves the highest mean F1 among compared methods (34.38%, versus 31.72% for 2DGS). We further also show that Hexels, as a general primitive, improves 2D image representation over Image-GS. Hexels push the Pareto frontier on the quality-efficiency trade-off of surface-aligned primitives.
Abstract (translated)
URL
https://arxiv.org/abs/2610.37668
PDF
-
PerceptRL: Making Visual Perception Rewardable in MLLMs from Answers to Interactions
2026-10-27 16:04:10Tongkun Guan, Zhibo Yang, Mingkun Yang, Haocheng Wang, Wei Shen, Xiaokang YangAbstract
Fine-grained visual perception remains a bottleneck during post-training of multimodal large language models (MLLMs), as perceptual errors can constrain gains from reasoning-focused optimization. Although image-caption pre-training improves perception, its influence can weaken as subsequent stages change data mixtures and objectives. Continuing caption supervision during post-training is challenging because subjective, information-dense descriptions complicate verifiable reward design. We present PerceptRL, which strengthens perception at two complementary levels. Perception RL uses a multi-agent pipeline to extract verified errors from dense image captions and convert the corresponding visual facts into QA pairs with fixed answer labels and rule-based rewards. Chain QA links correlated errors, while multiple-choice QA combines unrelated errors. A specialized route targets high-error domains with tailored verification, and a domain-agnostic general route covers errors beyond these domains. Perception Agentic RL addresses fine-grained and dense cases unresolved by single-turn Perception RL through multi-turn inspection with visual tools. We observe coordinate drift even in models trained to use tools, where incorrect crops or annotations distort evidence and propagate errors. Coordinate Injection corrects spatial operations in cold-start trajectories, while a grounding-aware process reward supervises intermediate spatial fidelity alongside final-answer correctness. Training proceeds through shared cold-start SFT, single-turn Perception RL, and multi-turn Perception Agentic RL. Across seven benchmarks under a shared tool-use protocol, PerceptRL achieves 71.6% average accuracy, surpassing the base model with the same tools and the strongest evaluated frontier VLM by 15.7 and 5.2 percentage points, respectively.
Abstract (translated)
URL
https://arxiv.org/abs/2610.37471
PDF
-
Cross-modal Steering Vectors Exist Across Both Image and Text Generation
2026-10-26 20:42:35Donghao Zeng (Mila, McGill University), Nishant Subramani (Carnegie Mellon University), Tamar Rott Shaham (MIT CSAIL), Benno Krojer (Mila, McGill University, Northeastern University)Abstract
Are concepts and mechanisms unified across modalities in current AI models, or instead implemented separately for each one? We operationalize this broad question of modality unification by studying cross-modal steering: 1) finding vectors that successfully steer generation on tokens of one modality, which then 2) transfer zero-shot to steer generation when added to tokens of another modality. Our study goes beyond simple correlational alignment (e.g. representational similarity) because steering is fundamentally causal. We study cross-modal steering in several unified multimodal models (UMMs), since they possess shared parameters for language and image generation, focusing on these two modalities. Across diverse steered concepts such as emotion and object size, we show that steering vectors transfer zero-shot in both directions (text-to-image & image-to-text). Even for concepts that are highly unique to one modality (e.g. formal vs. colloquial language), steering vectors often induce a consistent, interpretable effect in the other modality, such as formal clothing and design. This transfer has a geometric signature too: despite a clear modality gap separating text and image representations, steering vectors with strong cross-modal transferability tend to point in similar directions for the same concept. Our findings shed light on how vision and language interact inside unified multimodal models and point towards more general, modality-agnostic approaches for controlling their behavior.
Abstract (translated)
URL
https://arxiv.org/abs/2610.36269
PDF
