ICML 2026 论文列表

Multimodal large language models (MLLMs) have achieved remarkable progress in visual understanding tasks such as visual grounding, segmentation, and captioning. However, their ability to perceive perceptual-level image features remains limited. In this work, we present UniPercept-Bench, a unified framework for perceptual-level image understanding across three key domains: Aesthetics, Quality, Structure and Texture. We establish a hierarchical definition system and construct large-scale datasets to evaluate perceptual-level image understanding. Based on this foundation, we develop a strong baseline UniPercept trained via Domain-Adaptive Pre-Training and Task-Aligned RL, enabling robust generalization across both Visual Rating (VR) and Visual Question Answering (VQA) tasks. UniPercept outperforms existing MLLMs on perceptual-level image understanding and can serve as a plug-and-play reward model for text-to-image generation. This work defines perceptual-level image understanding in the era of MLLMs and, through the introduction of a comprehensive benchmark together with a strong baseline, provides a solid foundation for advancing perceptual-level multimodal image understanding.

Emerging Omni-modal Large Language Models (OmniLLMs) enable real-time interaction across video, audio, and text but suffer from prohibitive computational costs due to the quadratic complexity of processing continuous streaming inputs. Existing token compression strategies remain suboptimal as they typically rely on biased modality-centric priors or enforce uniform retention policies, neglecting the heterogeneity across layers and the critical role of cross-modality alignment. To address these challenges, we propose OmniFit, a training-free framework that decouples interaction profiling from inference execution. OmniFit incorporates Layer-Adaptive Heterogeneity Profiling (LAHP) to dynamically allocate computational budgets based on layer-wise redundancy and modality preferences, preserving tokens according to the characteristics of each layer. Furthermore, we introduce Alignment-Rectified Token Selection (ARTS), a lightweight mechanism that efficiently identifies tokens semantically aligned with cross-modal cues. Extensive experiments on 3 model series across 10 benchmarks demonstrate that OmniFit establishes a new Pareto frontier, retaining 98\% of model performance with only 20\% token usage and achieves up to 2.31$\times$ end-to-end inference speedup and 2.5$\times$ VRAM saving, significantly outperforming state-of-the-art methods.

The Platonic Representation Hypothesis posits that neural networks trained on different modalities converge toward a shared statistical model of the world. Recent work exploits this convergence by aligning frozen pretrained vision and language models with lightweight alignment layers, but typically relies on contrastive losses and millions of paired samples. In this work, we ask whether meaningful alignment can be achieved with substantially less supervision. We introduce a semi-supervised setting in which pretrained unimodal encoders are aligned using a small number of image-text pairs together with large amounts of unpaired data. To address this challenge, we propose SOTAlign, a two-stage framework that first recovers a coarse shared geometry from limited paired data using a linear teacher, and then refines the alignment on unpaired samples via an optimal-transport-based divergence that transfers relational structure without overconstraining the target space. Unlike existing semi-supervised methods, SOTAlign effectively leverages unpaired images and text, learning robust joint embeddings across datasets and encoder pairs, and significantly outperforming supervised and semi-supervised baselines.

Research on the expressive power of neural networks has identified the minimum depth and width of neural networks that enable universal approximation and memorization. However, existing results are derived under exact arithmetic and cannot be directly applied to real implementations on computers, which can only use a finite set of numbers and inexact machine operations with round-off errors. In this work, we study floating-point ReLU networks that have floating-point parameters and use floating-point operations. Specifically, we investigate their minimum depth and width to represent all functions from the set of floating-point vectors $\mathbb F^d$ to the set of floating-point numbers $\mathbb F$. We first show that the minimum depth for representing all functions from $\mathbb F^d$ to $\mathbb F$ is exactly three, where two layers can be sufficient if we consider a smaller domain and/or codomain. We further show that the minimum width for representing all functions from $\mathbb F^d$ to $\mathbb F$ lies between $2d$ and $2d+4$. In addition, if we restrict the domain to non-negative floats, it lies between $d$ and $d+4$, where it can be smaller for a smaller domain, even beyond $d$. Our results show that the existing results analyzed under exact arithmetic do not extend to the floating-point setup.

Scene Graph Generation (SGG) connects visual perception with structured reasoning, but is limited by scarce annotations and the long-tailed distribution of relational predicates. Training-free methods based on vision-language models (VLMs) reduce supervision requirements, yet often rely on flat graph structures and produce hallucinated relations. We present HSGG, a training-free framework for open-world Hierarchical Scene Graph Generation, whose inference proceeds in two steps. First, bidirectional hierarchical entity perception combines top-down object expansion with bottom-up attribute reasoning to construct multi-level scene hierarchies capturing part–whole semantics. Second, geometry-guided relation reasoning infers valid relations from these structured entities: geometry-aware relation filtering first prunes spatially implausible object pairs using 2D proximity, depth cues, and object scale, and geometry-grounded contrastive relation decoding then suppresses hallucinated predicates by contrasting predictions from a visually grounded expert against a hallucination-prone geometric prior, ensuring relations are both geometrically consistent and semantically coherent. Experiments show that HSGG generalizes effectively to unseen objects and predicates without training, substantially reduces relational hallucinations, and consistently improves downstream reasoning performance.

With the rapid advancement of large multimodal models (LMMs), inference-time overhead has become a key bottleneck for real-world deployment. Existing methods typically prune visual tokens at prefill, assuming the required visual evidence remains static during reasoning. However, we empirically show that visual evidence is strongly step-dependent: only a sparse subset of visual tokens is critical at each decoding step, and the critical set evolves across reasoning. Furthermore, we identify a coupled bottleneck where redundant visual context can steer the model toward query-irrelevant regions, lengthening the reasoning trace. Guided by these insights, we propose **VisionPulse**, a step-wise visual token pruning framework during reasoning. VisionPulse computes a lightweight visual attention mass to estimate the step-wise retention budget by exploiting its strong positive correlation with LMMs' effective visual token usage and retain only the most critical tokens under this budget. By enforcing visual sparsity during reasoning, VisionPulse filters redundant visual context while preserving relevant visual evidence, shortening reasoning traces naturally. Extensive experiments show that VisionPulse only retains 5\% of visual tokens per step with reasoning traces shortened by 11.2\%, while keeping accuracy almost unchanged.

The realization of autonomous scientific experimentation is currently limited by LLMs' struggle to grasp the strict procedural logic and accuracy required by biological protocols. To address this fundamental challenge, we present BioProBench, a comprehensive resource for procedural reasoning in biology. BioProBench is grounded in BioProCorpus, a foundational collection of 27,000 human-written protocols. From this corpus, we systematically constructed a dataset of over 550,000 task instances, offering both a large-scale training resource and a rigorous benchmark with novel metrics. Evaluating 10 mainstream LLMs, we find that while general comprehension is high, performance drops significantly on tasks demanding deep reasoning, quantitative precision, and safety awareness. To demonstrate the value of BioProCorpus in mitigating these issues, we developed \textbf{ProAgent}, grounded in our corpus, ProAgent substantially advances the state-of-the-art. Code and data are available at: https://github.com/YuyangSunshine/bioprotocolbench and https://huggingface.co/datasets/BioProBench/BioProBench .

Foundation models have revolutionized AI, but adapting them efficiently for multimodal tasks, particularly in dual-stream architectures composed of unimodal encoders, such as DINO and BERT, remains a significant challenge. Parameter-Efficient Fine-Tuning (PEFT) methods like Low-Rank Adaptation (LoRA) enable lightweight adaptation, yet they operate in isolation within each modality, limiting their ability in capturing cross-modal interactions. In this paper, we take a step in bridging this gap with Cross-Modal Low-Rank Adaptation (CoLA), a novel PEFT framework that extends LoRA by introducing a dedicated inter-modal adaptation pathway alongside the standard intra-modal one. This dual-path design enables CoLA to adapt unimodal foundation models to multimodal tasks effectively, without interference between modality-specific and cross-modal learning. We evaluate CoLA across a range of vision-language (RefCOCO, RefCOCO+, RefCOCOg) and audio-visual (AVE, AVS) benchmarks, where it consistently outperforms LORA, achieving a relative gain of around 3\% and 2\%, respectively, while maintaining parameter efficiency. Notably, CoLA enables the first multi-task PEFT framework for visual grounding, bridging a key gap in efficient multimodal adaptation. Code is available at https://github.com/peterwisu/CoLA

Understanding how cellular morphology, gene expression, and spatial context jointly shape tissue function is a central challenge in biology. Image-based spatial transcriptomics technologies now provide high-resolution measurements of cell images and gene expression profiles, but existing methods typically analyze these modalities in isolation or at limited resolution. We address the problem by introducing SPATIA, a multi-level generative and predictive model that learns unified, spatially aware representations by fusing morphology, gene expression, and spatial context from the cell to the tissue level. SPATIA also incorporates a spatially conditioned generative framework with confidence-aware OT reweighting and morphology-profile alignment for modeling target-state morphology distributions. Specifically, we propose a confidence-aware flow matching objective that reweights weak optimal-transport pairs based on uncertainty. We further apply morphology-profile alignment to encourage biologically meaningful image generation, enabling the modeling of microenvironment-dependent phenotypic transitions. We assembled a multi-scale dataset consisting of 25.9 million cell-gene pairs across 17 tissues. We benchmark SPATIA against 18 models across 12 tasks, spanning categories such as phenotype generation, annotation, clustering, gene imputation, and cross-modal prediction. SPATIA achieves improved performance over state-of-the-art models, improving generative fidelity by 8\% and predictive accuracy by up to 3\%.

Recent Multimodal Large Language Models (MLLMs) have demonstrated significant progress in perceiving and reasoning over multimodal inquiries, ushering in a new research era for foundation models. However, vision-language misalignment in MLLMs has emerged as a critical challenge, where the textual responses generated by these models are not factually aligned with the given text-image inputs. Existing efforts to address vision-language misalignment have focused on developing specialized vision-language connectors or leveraging visual instruction tuning from diverse domains. In this paper, we tackle this issue from a fundamental yet unexplored perspective by revisiting the core architecture of MLLMs. Most MLLMs are typically built on decoder-only LLMs consisting of a causal attention mechanism, which *limits the ability of the earlier modalities (e.g., images) to incorporate information from the latter modalities (e.g., text)*. To address this problem a MLLM that unlocks causal attention into our proposed modality-mutual attention (MMA) to enable image tokens to attend to text tokens. This simple yet effective design allows MMA to achieve state-of-the-art performance in 12 multimodal understanding benchmarks (**+6.2\% on average across 3 LLMs backbones**) without introducing additional parameters. Our MMA design is intended to be generic, allowing for applications across various modalities, and scalable to accommodate diverse multimodal scenarios.

A plethora of recent works has shown that for wide, overparameterized neural networks, training with Stochastic Gradient Descent (SGD) often leads to interpolation of the training data without sacrificing generalization performance. A key parameter that is not only closely connected to generalization properties, but is also closely tied to other desiderata such as robustness and resistance to adversarial perturbations is the Lipschitz constant of the neural network. While empirically, the Lipschitz constant has been shown not to increase with network width, theoretical findings only provide bounds with logarithmic growth in the width and only for the random initialization of ReLU-networks. In this work, we close this gap for neural networks with smooth activations by showing that, both at random initialization and throughout lazy training, the local Lipschitz constant of deep neural networks does not increase with network width. More precisely, we establish novel non-asymptotic (finite width) upper bounds and corroborate them by numerical experiments.

Instruction tuning aligns large language models, including multimodal ones, with diverse user intents, but scaling to heterogeneous mixtures is hindered by gradient interference and bandwidth-heavy synchronization. We ask whether these two bottlenecks can be addressed jointly by training parts of the mixture independently and reconciling them once in parameter space. We develop a local quadratic theory inside a shared flat basin that yields three results: weight merging produces a curvature-weighted variance reduction; PCA-aligned conflict splitting maximizes this gain along high-curvature directions; and merging additionally acts as spectral filtering with implicit norm regularization. These results directly motivate **MERIT**, a decentralized merge-ready instruction-tuning pipeline that estimates dataset-level gradient conflicts, partitions the mixture along the top PCA conflict axes, fine-tunes each partition independently with no inter-partition communication, and merges once via token-weighted averaging. On Qwen2.5-VL-3B with 136 Vision-FLAN tasks, MERIT improves the 8-benchmark average from 54.3 (joint training) to 57.0. The same recipe scales to a 7B model on a 1.6M-example, 176-source mixture—matching or exceeding centralized joint training with minimal cost overhead—and transfers to text-only FLAN. Our code is available at [https://github.com/naver-ai/merit](https://github.com/naver-ai/merit).

The trustworthiness of evaluation is critical to reliable model comparison and deployment in Video Anomaly Understanding (VAU). However, existing metrics are sensitive to expression styles and normal content, and this field lacks a diagnostic benchmark to validate metric validity and robustness. To bridge this gap, we propose: (1) a Class-Guided Chain-of-Evaluation (CG-CoE) metric, which structures assessment by extracting anomalous events and matching them under a class-specific semantic tolerance boundary, thereby decoupling anomaly semantics from descriptive style; and (2) an anomaly-focused meta-evaluation benchmark with two subsets: Anomalous Event-level Annotations (AEA) for measuring the validity of reflecting VAU models’ anomaly understanding ability and Controlled Variant Pairs (CVP) with fixed anomalies for quantifying robustness to stylistic perturbations. Extensive experiments demonstrate that CG-CoE achieves SOTA validity and robustness.

Charts are a fundamental visualization format for structured data analysis. Enabling end-to-end chart editing according to user intent is of great practical value, yet remains challenging due to the need for both fine-grained control and global structural consistency. Most existing approaches adopt pipeline-based designs, where natural language or code serves as an intermediate representation, limiting their ability to faithfully execute complex edits. We introduce ChartE$^{3}$, an **E**nd-to-**E**nd Chart **E**diting benchmark that directly evaluates models without relying on intermediate natural language programs or code-level supervision. ChartE$^{3}$ focuses on two complementary editing dimensions: local editing, which involves fine-grained appearance changes such as font or color adjustments, and global editing, which requires holistic, data-centric transformations including data filtering and trend line addition. ChartE$^{3}$ contains over 1,200 high-quality samples constructed via an well-designed data pipeline with human curation. Each sample is provided as a triplet of a chart image, its underlying code, and a multimodal editing instruction, enabling evaluation from both visual and structural perspectives. Extensive benchmarking of state-of-the-art multimodal large language models reveals substantial performance gaps, particularly on global editing tasks, highlighting critical limitations in current end-to-end chart editing capabilities.

We investigate unpaired image inverse problems, a challenging setting where only independent, non-paired sets of noisy measurements and clean target signals are available for training. We propose a novel inverse problem solver based on Unbalanced Optimal Transport, called ***Unbalanced Optimal Transport Map for Inverse Problems (UOTIP)***. Our method formulates the reconstruction task—predicting clean target signals from noisy measurements—as learning a UOT Map from noisy measurement distribution to clean signal distribution by incorporating a likelihood-based cost function. By relaxing the exact marginal constraint, the UOT framework provides key advantages to our model: robustness to multi-level observation noise, adaptability to class imbalance between noisy and clean datasets, and generalizability to diverse noise-type scenarios. Furthermore, we theoretically demonstrate that incorporating a quadratic cost term ensures the existence and uniqueness of the transport map by satisfying the twist condition, even for ill-posed inverse problems. Our experiments demonstrate that UOTIP achieves state-of-the-art performance on unpaired image inverse problem benchmarks, across linear and nonlinear inverse problems.

Large language models (LLMs) achieve strong capabilities by scaling model capacity and training data, yet many real-world deployments rely on smaller models trained or adapted from low-resource data. This gap motivates the need for mechanisms to transfer knowledge from large, high-resource models to smaller, low-resource targets. While model merging provides an effective transfer mechanism, most existing approaches assume architecture-compatible models and therefore cannot directly transfer knowledge from large high-resource LLMs to heterogeneous low-resource targets. In this work, we propose a cross-architecture merging framework based on optimal transport (OT) that aligns activations to infer cross-neuron correspondences between heterogeneous models. The resulting transport plans are then used to guide direct weight-space fusion, enabling effective high-resource to low-resource transfer using only a small set of inputs. Extensive experiments across low-resource languages and specialized domains demonstrate consistent improvements over target models.

Vision–Language Models (VLMs) are increasingly deployed with web search tools, yet we still lack benchmarks that isolate a critical capability for real-world use: deciding when to search and how to steer search from ambiguous visual evidence, especially when multiple images provide overlapping or conflicting cues. We introduce NAVIGATE, a novel benchmark centered on images as primary evidence for open-web search planning and multi-step reasoning. It contains 500 questions across 20 domains and spans three difficulty tiers, from single-image, self-contained problems to multi-image joint search and multi-domain composition. Unlike prior benchmarks that specify explicit search targets, NAVIGATE evaluates search decision-making: models must infer whether external search is necessary and iteratively refine search directions based on holistic reasoning over visual cues. Across a broad set of VLMs and search-enabled systems, performance remains low, Gemini-3-Pro-Preview-Search reaches only 36.4% accuracy, highlighting persistent failures in cross-image grounding, search triggering, and search strategy coordination. We will release NAVIGATE publicly.

While Vision-Language-Action (VLA) models hold promise in embodied intelligence, their large parameter counts lead to substantial inference latency that hinders real-time manipulation, motivating parameter sparsification. However, as the environment evolves during VLA execution, the optimal sparsity patterns change accordingly. Static pruning lacks the adaptability required for environment dynamics, whereas fixed-interval dynamic layer pruning suffers from coarse granularity and high retraining overheads. To bridge this gap, we propose **EcoVLA**, a training-free, plug-and-play adaptive pruning framework that supports orthogonal combination with existing VLA acceleration methods. EcoVLA comprises two components: **E**nvironment-aware **A**daptive **P**runing (**EAP**) and **I**nterleaved **I**nference **O**rchestration (**$I^2O$**). EAP is a lightweight adaptive channel pruning method that incorporates the temporal consistency of the physical environment to update sparsity patterns. $I^2O$ leverages the FLOPs bubbles inherent in VLA inference to schedule the pruning method in parallel, ensuring negligible impact on latency. Evaluated on diverse VLA models and benchmarks, EcoVLA delivers state-of-the-art performance, achieving up to 1.60$\times$ speedup with only a 0.4% drop in success rate, and further reaches 2.18$\times$ speedup with only a 0.5% degradation when combined with token pruning. We further validate the effectiveness of EcoVLA on real-world robots. Our code is available [here](https://github.com/Echo-hyt/Ecovla).

Multimodal representation learning harmonizes distinct modalities by aligning them into a unified latent space. Recent research generalizes traditional cross-modal alignment to produce enhanced multimodal synergy but requires all modalities to be present for a common instance, making it challenging to utilize prevalent datasets with missing modalities. We provide theoretical insights into this issue from an \textit{anchor shift} perspective. Observed modalities are aligned with a local anchor that deviates from the optimal one when all modalities are present, resulting in an inevitable shift. To address this, we propose CalMRL to calibrate incomplete alignments caused by missing modalities. CalMRL leverages the priors and the inherent connections among modalities to model the imputation for the missing ones at the representation level. To resolve the optimization dilemma, we employ a bi-step learning method with the closed-form solution of the posterior distribution of shared latents. We validate its mitigation of anchor shift and convergence with theoretical guidance. By equipping the calibrated alignment with the existing advanced method, we offer new flexibility to absorb data with missing modalities, which is originally unattainable. Extensive experiments demonstrate the superiority of CalMRL. The code is released at https://github.com/Xiaohao-Liu/CalMRL.