Vision-Language Models

Spatial reasoning is a critical capability for intelligent robots, yet current vision-language models (VLMs) still fall short of human-level performance in video-based spatial reasoning. This gap mainly stems from two challenges: a semantic-geometric misalignment that prevents consistent 3D understanding, and the absence of persistent memory to retain 3D representation and understanding across frames. To address these limitations, we present VLM$^2$, a Vision-Language Model with persistent Memory for spatial reasoning with a view-consistent, 3D-aware representation purely from 2D videos. Specifically, we incorporate a dual-memory module consisting of a working memory that operates as a sliding window to focus on immediate context, and an episodic memory that consolidates and stores critical information across frames. This design enables bounded and efficient spatial reasoning under a fixed computational cost. Extensive experiments on multiple benchmarks show that VLM$^2$ achieves state-of-the-art performance among video-based models, significantly advancing the frontier of visual-spatial intelligence.

Visual in-context learning (VICL) solves visual tasks by conditioning on a few input-output demonstrations without any model training. Recent advances in large vision-language models (VLMs) have shown promising VICL capability when the demonstration pair and the query belong to the same vision task, but real use cases often provide mismatched examples, making it unclear whether a VLM should imitate the demonstrated transformation or infer a new one from the query. This raises a fundamental question: Can VLMs perform cross-task VICL where demonstration and query differ? In the paper, we study this cross-task VICL setting and propose T2T-VICL, a collaborative prompt-transfer framework, which converts mismatched visual demonstrations into implicit textual guidance without explicitly naming the tasks. To do so, a large teacher VLM first generates structured descriptions of visual changes and task differences between task pairs, from which we construct a dataset of diverse implicit cross-task relations. We then distill this capability into a lightweight student VLM that produces content-dependent prompts from a task-A demonstration pair and a task-B query. The generated prompt is used to guide a frozen image-editing VLM, and a score-based inference strategy is introduced to rank multiple candidates. Experiments on 12 low-level vision tasks and over 20 evaluated cross-task pairs show that T2T-VICL consistently improves task-aware alignment over fixed prompting and often also improves image fidelity, revealing both the potential and limits of cross-task VICL. Our code is available on GitHub.

Vision-language models have recently emerged as promising planners for autonomous driving, where success hinges on reasoning over spatial structure and dynamic interactions from multimodal input. However, existing models are typically trained without supervision that explicitly encodes relational dependencies, requiring them to infer interactions between traffic entities implicitly from raw sensor data. In this work, we introduce GraphPilot, a model-agnostic approach that conditions language-based driving models on structured relational context in the form of traffic scene graphs. We serialize scene graphs at different abstraction levels and formats and inject them through structured prompts without modifying the underlying architecture or training objective. Extensive closed-loop evaluations on LangAuto and Bench2Drive show persistent improvements across LMDrive, BEVDriver, and SimLingo. Controlled ablations and hidden-state probing further indicate that relational supervision contributes substantially to these improvements while increasing the decodability of scene relations. Open-loop SpaceDrive results on nuScenes further demonstrate transfer to real-world data. These results indicate that diverse architectures can effectively internalize and ground relational priors through scene graph-conditioned training, even without requiring scene graph input at test-time. Code, fine-tuned models, and our scene graph dataset are publicly available at https://github.com/iis-esslingen/GraphPilot.

Multimodal LLMs can answer many questions in vision-centric benchmarks without looking at the image, using linguistic priors and skewed answer distributions. Blind evaluation shows what a model already answers this way, but not what it could learn from the test set's own regularities. Extending partial-input auditing (e.g., hypothesis-only baselines in natural language inference), we argue that benchmark designers should "train on the test set": probe the artifact they release for exploitable patterns. Our Test-set Stress-Test (TsT) cross-validates a text-only Qwen2-7B on the test set's questions and answer options, yielding a benchmark-level score and a per-question bias score s(x); a random forest on hand-crafted features adds a fast, interpretable audit. On the template-based VSI-Bench and CV-Bench, the held-out score is 17.9 and 13.1 points above the model's own zero-shot score. On MMMU it learns little, even though GPT-4o correctly answers 52.3% of its multiple-choice questions without the image, so blind success there reflects pretrained knowledge rather than learnable test-set patterns. Iterative Bias Pruning (IBP) removes the questions with the highest s(x) and re-diagnoses; on VSI-Bench it widens a fine-tuned model's vision-blind gap more than random removal at the same rate. We also release VSI-Bench-Debiased, which lowers a fine-tuned model's blind score from 44.7 to 32.0 while its vision score falls only from 57.1 to 48.7.

Large vision-language models (LVLMs) have achieved impressive performance across a wide range of vision-language tasks, while they remain vulnerable to backdoor attacks. Existing backdoor attacks on LVLMs aim to force the victim model to generate a predefined target pattern, which is either inserted into or replaces the original content. We find that these fixed-pattern attacks are relatively easy to detect, because the attacked LVLM tends to memorize such frequent patterns in the training dataset, thereby exhibiting overconfidence on these targets given poisoned inputs. To address these limitations, we introduce TokenSwap, a more evasive and stealthy backdoor attack that focuses on the compositional understanding capabilities of LVLMs. Instead of enforcing a fixed targeted content, TokenSwap subtly disrupts the understanding of object relationships in text. Specifically, it causes the backdoored model to generate outputs that mention the correct objects in the image but misrepresent their relationships (i.e., bags-of-words behavior). During training, TokenSwap injects a visual trigger into selected samples and simultaneously swaps the grammatical roles of key tokens in the corresponding textual answers. However, the poisoned samples exhibit only subtle differences from the original ones, making it challenging for the model to learn the backdoor behavior. To address this, TokenSwap employs an adaptive token-weighted loss that explicitly emphasizes the learning of swapped tokens, such that the visual triggers and bags-of-words behavior are associated. Extensive experiments demonstrate that TokenSwap achieves high attack success rates while maintaining superior evasiveness and stealthiness across multiple benchmarks and various LVLM architectures.

Class-Incremental Learning (CIL) aims to endow models with the ability to continuously adapt to evolving data streams. Recent advances in pre-trained vision-language models (e.g., CLIP) provide a powerful foundation for this task. However, existing approaches often rely on simplistic templates, such as "a photo of a [CLASS]", which overlook the hierarchical nature of visual concepts. For example, recognizing "cat" versus "car" depends on coarse-grained cues, while distinguishing "cat" from "lion" requires fine-grained details. Similarly, the current feature mapping in CLIP relies solely on the representation from the last layer, neglecting the hierarchical information contained in earlier layers. In this work, we introduce HiErarchical Representation MAtchiNg (HERMAN) for CLIP-based CIL. Our approach leverages LLMs to recursively generate discriminative textual descriptors, thereby augmenting the semantic space with explicit hierarchical cues. These descriptors are matched to different levels of the semantic hierarchy and adaptively routed based on task-specific requirements, enabling precise discrimination while alleviating catastrophic forgetting in incremental tasks. Extensive experiments on multiple benchmarks demonstrate that our method consistently achieves state-of-the-art performance.

Vision-language models (VLMs) have enabled strong zero-shot classification through image-text alignment. Yet, their purely visual inference capabilities remain under-explored. In this work, we conduct a comprehensive evaluation of both language-guided and vision-only image classification with a diverse set of dual-encoder VLMs, including both well-established and recent models such as SigLIP 2 and RADIOv2.5. The performance is compared in a standard setup on the ImageNet-1k validation set and its label-corrected variant. The key factors affecting accuracy are analysed, including prompt design, class diversity, the number of neighbours in k-NN, and reference set size. We show that language and vision offer complementary strengths, with some classes favouring textual prompts and others better handled by visual similarity. To exploit this complementarity, we introduce a simple, learning-free fusion method based on per-class precision that improves classification performance. The code is available at: https://github.com/gonikisgo/bmvc2025-vlm-image-recognition.

Sign Language (SL) enables two-way communication for the deaf and hard-of-hearing community, yet many sign languages remain under-resourced in the AI space. Sign Language Instruction Generation (SLIG) produces step-by-step textual instructions that enable non-SL users to imitate and learn SL gestures, promoting two-way interaction. We introduce BdSLIG, the first Bengali SLIG dataset, used to evaluate Vision Language Models (VLMs) (i) on under-resourced SLIG tasks, and (ii) on long-tail visual concepts, as Bengali SL is unlikely to appear in the VLM pre-training data. To enhance zero-shot performance, we introduce Sign Parameter-Infused (SPI) prompting, which integrates standard SL parameters, like hand shape, motion, and orientation, directly into the textual prompts. Subsuming standard sign parameters into the prompt makes the instructions more structured and reproducible than free-form natural text from vanilla prompting. We envision that our work would promote inclusivity and advancement in SL learning systems for the under-resourced communities.

Video Temporal Grounding (VTG) aims to localize specific video segments corresponding to natural language queries. While recent Large Vision-Language Models (LVLMs) employ Reinforcement Learning to generate Chains-of-Thought (CoT), they typically rely solely on outcome-based supervision. Consequently, this often leads to hallucinations, where the reasoning process becomes disconnected from the visual content and the final prediction. Existing attempts to mitigate this by relying on external supervision from larger models or separate reward models are computationally expensive and prone to rigid patterns. To address these challenges, we propose TAR (Temporal Anchor-Constrained Reasoning), a framework that introduces the temporal anchor (T-anchor) as a transparent and auditable checkpoint mechanism. T-anchor enforces progressive refinement within the CoT, compelling the model to continuously ground its intermediate thoughts in visual evidence and iteratively calibrate temporal predictions, thereby significantly enhancing the faithfulness and autonomy of the reasoning process and final accuracy. Furthermore, we introduce a bootstrapping paradigm that automatically harvests high-quality CoT data using only a standard 7B model, eliminating the dependency on ultra-large models. Extensive experiments demonstrate that TAR achieves state-of-the-art performance and generates faithful, autonomous, and progressively refined reasoning traces.

Large language models (LLMs) have demonstrated remarkable capabilities across diverse language tasks, motivating their extension to vision-language models (VLMs) for multimodal understanding. However, billion-parameter VLMs incur substantial memory and computational costs that hinder deployment in resource-constrained settings. Post-training quantization (PTQ) compresses models and accelerates inference without retraining, yet remains underexplored for VLMs. We identify two intrinsic VLM activation properties in PTQ: (1) visual over-representation, where vision tokens are excessive and often redundant, and (2) the modality gap separating text and vision tokens in the latent feature space. Prior methods largely overlook these properties, leading to quantization performance degradation. To address this mismatch, we propose SalQ-VLM, an importance-aware PTQ framework that prioritizes salient tokens and suppresses redundant vision tokens during calibration. We derive a gradient-driven importance factor that captures token-level importance variance and is theoretically grounded in the relationship among loss perturbation, activation errors, and output gradients. SalQ-VLM obtains this factor through a single lightweight block-wise gradient-caching pass and incorporates it into the layer-wise reconstruction objective. Because SalQ-VLM modifies only calibration, it adds no inference-time operations and remains compatible with existing high-performance kernels. Extensive evaluations across benchmarks and backbones show that SalQ-VLM consistently outperforms strong PTQ baselines, especially under ultra-low-bit quantization. Notably, it improves MME-RealWorld accuracy by 16.45% under INT2g128 quantization.

Currently, a prevalent approach for enhancing Vision-Language Models (VLMs) performance is to encode both the high-resolution version and the thumbnail of an image simultaneously. While effective, this method generates a large number of image tokens. When combined with the widely used Rotary Position Embedding (RoPE), its long-term decay property hinders the interaction between high-resolution tokens and thumbnail tokens, as well as between text and image. To address these issues, we propose ID-Align, which alleviates these problems by reordering position IDs. In this method, high-resolution tokens inherit IDs from their corresponding thumbnail token while constraining the overexpansion of positional indices. Our experiments conducted within the LLaVA-Next framework demonstrate that ID-Align achieves significant improvements, including a 6.09% enhancement on MMBench's relation reasoning tasks and notable gains across multiple benchmarks. Our code is available at the following link: https://github.com/zooblastlbz/ID-Align.

Modern Vision-Language Models (VLMs) can solve a wide range of tasks requiring visual reasoning. In real-world scenarios, desirable properties for VLMs include fast inference and controllable generation (e.g., constraining outputs to adhere to a desired format). However, existing autoregressive (AR) VLMs like LLaVA struggle in these aspects. Discrete diffusion models (DMs) offer a promising alternative, enabling parallel decoding for faster inference and bidirectional context for controllable generation through text-infilling. While effective in language-only settings, DMs' potential for multimodal tasks is underexplored. We introduce LaViDa, a family of VLMs built on DMs. We build LaViDa by equipping DMs with a vision encoder and jointly fine-tune the combined parts for multimodal instruction following. To address challenges encountered, LaViDa incorporates novel techniques such as complementary masking for effective training, prefix KV cache for efficient inference, and timestep shifting for high-quality sampling. Experiments show that LaViDa achieves competitive or superior performance to AR VLMs on multi-modal benchmarks such as MMMU, while offering unique advantages of DMs, including flexible speed-quality tradeoff, controllability, and bidirectional reasoning. On COCO captioning, LaViDa surpasses Open-LLaVa-Next-8B by +4.1 CIDEr with 1.92x speedup. On bidirectional tasks, it achieves +59% improvement on Constrained Poem Completion. These results demonstrate LaViDa as a strong alternative to AR VLMs. Code and models will be released in the camera-ready version.

Despite the impressive capabilities of Large Vision-Language Models (LVLMs), they remain susceptible to hallucinations, where generated content is inconsistent with the input image. Existing training-free hallucination mitigation methods often suffer from unstable performance and high sensitivity to hyperparameter settings, which limits their practicality and broader adoption. In this paper, we propose Decoding with Inter-layer Consistency via Layer Aggregation (DCLA), a training-free decoding mechanism that requires no retraining, fine-tuning, or access to external knowledge bases. Specifically, DCLA constructs a dynamic semantic reference by aggregating representations from previous layers and uses it to correct semantically deviated layers, thereby enforcing inter-layer consistency. Experiments across seven LVLMs and multiple benchmarks demonstrate the generality of DCLA: it surpasses standard decoding by 28.58 MME points on LLaVA1.5-7B and 42.6 MME points on Qwen2.5-VL, while improving POPE accuracy by 2.74 percentage points in the strongest setting.

Vision-Language Models (VLMs) have demonstrated exceptional performance in various multi-modal tasks. Recently, there has been an increasing interest in improving the personalization capabilities of VLMs. To better integrate user-provided concepts into VLMs, many methods use positive and negative samples to fine-tune these models. However, the scarcity of user-provided positive samples and the low quality of retrieved negative samples pose challenges for existing techniques. To reveal the relationship between sample and model performance, we systematically investigate the amount and diversity impact of positive and negative samples (easy and hard) on VLM personalization tasks. Based on the detailed analysis, we introduce Concept-as-Tree (CaT), which represents a concept as a tree structure, thereby enabling the data generation of positive and negative samples with varying difficulty and diversity, and can be easily extended to multi-concept scenarios. With a well-designed data filtering strategy, our CaT framework can ensure the quality of generated data, constituting a powerful pipeline. We perform thorough experiments with various VLM personalization baselines to assess the effectiveness of the pipeline, alleviating the lack of positive samples and the low quality of negative samples. Our results demonstrate that CaT equipped with the proposed data filter significantly enhances the capabilities of VLMs across personalization benchmarks. To the best of our knowledge, this work is the first controllable synthetic data pipeline for VLM personalization.

Humans develop perception through a bottom-up hierarchy: from basic primitives and Gestalt principles to high-level semantics. Current Multimodal Large Language Models (MLLMs) are trained directly on complex downstream tasks. Does it guarantee their performance on these foundational visual capabilities? To systematically investigate this gap, we introduce VisFacor, a benchmark that digitizes 20 vision-centric subtests from FRCT, a well-established cognitive psychology assessment spanning four domains of human visual cognition. Furthermore, we design algorithms to automatically construct and validate unlimited test cases with controllable difficulty. Using VisFacor, we evaluate 39 frontier MLLMs, including both proprietary (e.g., GPT, Gemini) and open-source (e.g., LLaMA, Qwen) models. The best model achieves a score of only 55.9%. Analysis reveals good internal consistency (Cronbach's alpha = 0.94) and construct validity (compared to existing vision benchmarks). Models consistently fail on tasks such as mental rotation, spatial relation inference, and figure-ground discrimination, regardless of model size or prompting strategy. These findings suggest that performance improvements on existing general benchmarks might represent castles in the air instead of a genuine mastery of human-like visual cognition.

The rapid advancement of AI-generated video poses challenges to digital authenticity and security. Current detection methods, often trained on specific datasets, struggle with the ever-evolving landscape of generative techniques and unseen manipulations. We introduce a framework leveraging Vision Language Models (VLMs) for robust AI-generated video detection. Our approach equips the VLM with the ability to reason about video content and use external tools to identify subtle inconsistencies, mirroring human system 2 thinking. Our self-evolving VLM dynamically selects and composes appropriate tools, enhancing its ability to generalize to novel video generation techniques. The modular design promotes interpretability, allowing for a clearer understanding of VLM's decision-making process. To evaluate, we establish the first benchmark VidForensic containing 1.4k+ high-quality AI-generated videos across eight generative models. Experiments show that REVEAL improves F1 scores by 9.1% to 30.2% over top baselines across our datasets for VLMs, notably for GPT-4o, Gemini 1.5 pro, and QWen-VL-Max, and Llava-One-Vision-7B. While open-world AI-video detection remains an open challenge, our results indicate that existing methods fail primarily because they lack tool-enabled, higher-order reasoning.

Large-scale Vision-Language Models (VLMs) have achieved notable progress in aligning visual inputs with text. However, their ability to deeply understand the unique physical properties of non-RGB vision sensor images remains limited. In this paper, we revisit and analyze these limitations and introduce a novel, cost-efficient paradigm that significantly advances sensor image understanding-without requiring extensive training data or any modifications to the existing VLM architectures. Specifically, we propose Sensor-Aware Attributes Fine-Tuning (SAFT) with the Diverse Negative Attributes (DNA) optimization, which leverages minimal sensor-specific data to enable robust learning of non-RGB characteristics and overcome RGB-centric biases inherent in current VLMs. In addition, we present VS-TDX-the first comprehensive, public benchmark designed to rigorously evaluate VLMs' sensor-specific understanding across diverse and realistic scenarios. Through extensive experiments on VLMs and various sensor modalities, we validate that our method consistently delivers superior performance and generalization under resource-constrained and architecture-invariant settings. Our approach provides a practical advance towards scalable deployment of VLMs in increasingly sensor-diverse real-world environments.

Vision-language Models (VLMs) are essential for contextual understanding of both visual and textual information. However, their vulnerability to adversarially manipulated inputs presents significant risks, leading to compromised outputs and raising concerns about the reliability in VLM-integrated applications. Detecting these malicious prompts is thus crucial for maintaining trust in VLM generations. A major challenge in developing a safeguarding prompt classifier is the lack of a large amount of labeled benign and malicious data. To address the issue, we introduce VLMGuard, a novel learning framework that leverages the unlabeled user prompts in the wild for malicious prompt detection. These unlabeled prompts, which naturally arise when VLMs are deployed in the open world, consist of both benign and malicious information. To harness the unlabeled data, we present an automated maliciousness estimation score for distinguishing between benign and malicious samples within this unlabeled mixture, thereby enabling the training of a binary prompt classifier on top. Notably, our framework does not require extra human annotations and is robust to realistic prompt variations, offering strong flexibility and practicality for real-world applications. Extensive experiments show that VLMGuard achieves superior detection results, improving AUROC by 5.39% on average over the state-of-the-art method. Disclaimer: This paper may contain offensive examples; reader discretion is advised. Code is available at: https://github.com/radiolab-ntu/vlmguard.