ICML 2026 论文列表

Rapid advances in AI-generated image (AIGI) technology enable highly realistic synthesis, threatening public information integrity and security. Recent studies have demonstrated that incorporating texture-level artifact features alongside semantic features into multimodal large language models (MLLMs) can enhance their AIGI detection capability. However, our preliminary analyses reveal that artifact features exhibit high intra-feature similarity, leading to an almost uniform attention map after the softmax operation. This phenomenon causes attention dilution, thereby hindering effective fusion between semantic and artifact features. To overcome this limitation, we propose a lightweight fusion adapter, TranX-Adapter, which integrates a Task-aware Optimal-Transport Fusion that leverages the Jensen-Shannon divergence between artifact and semantic prediction probabilities as a cost matrix to transfer artifact information into semantic features, and an X-Fusion that employs cross-attention to transfer semantic information into artifact features. Experiments on standard AIGI detection benchmarks upon several advanced MLLMs, show that our TranX-Adapter brings consistent and significant improvements (up to +6% accuracy). Code can be found in supplementary material.

Transformer-based tabular foundation models (TFMs) dominate small to medium tabular predictive benchmark tasks, yet their inference mechanisms remain largely unexplored. We present the first large-scale mechanistic study of layerwise dynamics in 6 state-of-the-art tabular in-context learning models. We explore how predictions emerge across depth, identify distinct stages of inference and reveal latent-space dynamics that differ from those of language models. Our findings indicate substantial depthwise redundancy across multiple models, suggesting iterative refinement with overlapping computations during inference stages. Guided by these insights, we design a proof-of-concept, looped single-layer model that uses only 20% of the original model’s parameters while achieving comparable performance. The code is available at https://github.com/amirbalef/is_one_layer_enough.

Recent work on Neural Network-based methods for nonlinear control use Lyapunov Functions to obtain controllers with guarantees of stability. However, Lyapunov-based methods are fundamentally limited: they cannot be used for smooth blending with formal Region of Attraction (RoA) expansion guarantees, and also fail to certify stability when unstable equilibria or saddle points are present. Density functions provide an alternate stability certificate, and address these limitations by certifying almost everywhere stability, and enable smooth blending of controllers. Learning valid density certificates is challenging due to integrability constraints, and the effect of density-based blending controllers on RoAs is not well understood. In this work, we provide the first guarantee that controllers blended with density functions yield RoAs containing the union of the RoAs achieved by the constituent controllers. Then, we propose a novel exponential characterization of density functions that provably satisfies the integrability condition, and introduce Neural Control Density Functions (NCDFs), that leverage this new parameterization. We also extend NCDFs for synthesizing safe-stable controllers by combining NCDFs with control barrier functions (NCDF-CBFs). Our experiments show that blended controllers obtain superior RoAs to state-of-the-art methods like Neural Lyapunov Control and Sum-of-Squares based techniques.

This work studies the convergence of two-timescale stochastic approximations (SA), a class of iterative algorithms that update two sets of parameters in fast and slow timescales respectively. Notable examples of two-timescale SA in reinforcement learning (RL) include temporal difference learning with gradient correction (TDC) and actor-critic methods. Previously, the stability (i.e., boundedness) and convergence of two-timescale SA were only established under i.i.d. noise. This work instead establishes the stability and convergence of two-timescale SA under Markovian noise, a setup that is more realistic in RL. Notably, we do not need to use any projection operator and the noise does not need to live in a compact space. Our key technical novelty is to control the fast timescale parameter with the running max of the slow timescale parameter, instead of with the current slow timescale parameter, as most prior works do. As a key application, we establish the first almost sure convergence of TDC with eligibility traces under off-policy learning with linear function approximation.

When evaluating Large Language Models (LLMs) in question-answering domains, multiple-choice question answering (MCQA) is widely used because it enables automatic grading. However, MCQA also exposes models to answer options that can be exploited in ways that inflate reasoning ability. We study this phenomenon across $15$ question-answering benchmarks and $27$ LLMs by systematically varying how and when models are exposed to answer options. For non-reasoning LLMs, MCQA can remain a good proxy for free-text performance when any chain-of-thought is produced only before the options are revealed. However, this "decoupled" format is not realizable for most reasoning models: they are designed to emit reasoning tokens whenever they are prompted, so if options are present they inevitably "reason over" the options. In practice, this makes reasoning models particularly effective at extracting signal from options, and can create large, misleading gains over free-text baselines. To characterize how models exploit MCQA, we introduce diagnostic probes that isolate option-only and question-plus-option exploitation pathways, and we quantify how design choices such as distractor strength and "none-of-the-above" answers effect exploitability. Finally, we examined the practice of multiple choice as an error diagnostic: inferring a model's mistake from the wrong option it picks. On benchmarks where reasoning can be expressed as code, we ask models to output code, we then executed it varying the inputs, and compared the resulting input–output behavior, revealing failure modes that MCQA diagnostics obscure. Lastly, we offer practical guidelines when analyzing results from MCQA that better reflect LLMs' genuine reasoning capabilities.

In RGB--D salient object detection, multimodal fusion is commonly implemented via symmetric token interaction, implicitly allowing information to flow in both directions. Under *modality imbalance*---when an auxiliary stream is substantially noisier than a designated primary stream---such symmetry creates a *backflow channel* that injects auxiliary noise into the primary representation and amplifies errors across iterative refinement stages. We formulate fusion in this regime as *directed refinement with one-way safety*: the primary modality defines a guidance field, while only auxiliary representations are iteratively purified, and primary perturbations induced by the auxiliary stream are explicitly bounded. We propose *Hamiltonian Asymmetric Fusion* (HAF), a lightweight unrolled refinement block that updates auxiliary tokens with momentum regularization and gated driving. The refinement force is instantiated by FFT-based spectral global correlation and modulated by a shared learnable spectral response to emphasize reliable frequency components with minimal parameters; a leaky momentum gate and a stable integrator improve multi-step refinement stability. We provide guarantees of auxiliary error contraction and bounded primary perturbation, which symmetric fusion operators do not satisfy under imbalance. Experiments on six RGB--D SOD benchmarks show consistent gains and substantially more graceful degradation under controlled auxiliary corruption.

This paper introduces a rigorous framework for defining generative diffusion models in infinite dimensions via Doob's h-transform. Rather than relying on time reversal of a noising process, a reference diffusion is forced towards the target distribution by an exponential change of measure. Compared to existing methodology, this approach readily generalises to the infinite-dimensional setting, hence offering greater flexibility in the diffusion model. The construction is derived rigorously under verifiable conditions, and bounds with respect to the target measure are established. We show that the forced process under the changed measure can be approximated by minimising a score-matching objective and validate our method on both synthetic and real data.

Automating User Interface (UI) generation substantially improves productivity and accelerates development by reducing engineering time and manual effort. Despite recent progress of Multimodal Large Language Models (MLLMs) in UI2Code, most existing approaches focus on a single HTML/CSS form and fail to systematically incorporate front-end frameworks such as React, Vue, and Angular. Moreover, their outputs are often verbose and hard to reuse at the component level. To address those issues, we propose Deterministic Component Mining (DCM), a multi-stage pipeline that couples MLLMs with a compact intermediate representation to enable multi-framework and component-oriented code generation. Firstly, a lightweight structure model predicts the representation of the DOM tree in JSON format, capturing the coarse layout from a webpage screenshot. Subsequently, we formulate deterministic rules to normalize the predicted DOM tree and mine reusable components with repetitive patterns via structural hashing and clustering, thereby yielding a portable intermediate representation. Finally, we employ a framework-conditioned prompting strategy governed by a binding specification and a file-block protocol to emit HTML/React/Vue/Angular code with explicit component props and repeat constructs. Extensive experiments demonstrate that DCM significantly outperforms baselines on automatic evaluation metrics and component-level reuse, while delivering consistent gains in multi-framework portability and overall code structural quality.

Innovation is a key driving force of human civilization. As the body of knowledge has grown considerably, bridging knowledge across different disciplines, where significant innovation often emerges, has become increasingly challenging. The recent advancements in machine learning models, particularly Large Language Models (LLMs), have provided effective access to extensive knowledge sources and shown impressive abilities in reasoning, rendering significant opportunities for interdisciplinary discovery. Our research aims to understand the capabilities of state-of-the-art LLMs in integrating knowledge from different fields for interdisciplinary research (IDR). To address this fundamental problem, we introduce IDRBench, a pioneering framework that includes both datasets and evaluation tasks: (1) IDR Paper Identification, (2) IDR Idea Integration, and (3) IDR Idea Recommendation. Our study on ten mainstream LLMs provides a comprehensive analysis of their behavior and establishes benchmarks and baselines for future research. To the best of our knowledge, IDRBench is the first to provide a comprehensive investigation of LLMs' IDR capability.

Data cleaning and automated machine learning (AutoML) are both crucial for reliable learning systems, yet are commonly treated as independent or sequential stages. This separation ignores their strong interaction and leads to inefficient use of limited computational budgets.We propose DMCO, a unified framework that jointly optimizes data cleaning and model construction under a fixed resource budget. DMCO reformulates the traditional two-stage pipeline into a time-sliced process, where data cleaning and AutoML are interleaved and adaptively scheduled. We introduce a gradient-based data cleaning sampling strategy with theoretical guarantees for minimizing gradient estimation variance, and integrates it with loss-driven sampling and progressive AutoML fitting to continuously leverage intermediate data quality improvements.Experiments on six real-world datasets show that DMCO consistently outperforms standalone data cleaning and AutoML baselines on both classification and regression tasks, as measured by F1 score and MSE. Under limited budgets, DMCO achieves up to 82.19% of the performance of full data cleaning with exhaustive AutoML, while remaining robust across different AutoML frameworks.

Recent reinforcement learning (RL) based large-thinking models demonstrate impressive expert-level abilities, i.e., software and math, but still rely heavily on verifiable rewards in specific domains, which places a significant bottleneck to extend the performance boundary of general reasoning capabilities. In this work, we propose PretrainZero, a reinforcement active learning framework built on the pretraining corpus to extend RL from domain-specific post-training to general pretraining. PretrainZero features the following characteristics: 1) Active pretraining objectives: inspired by the active learning ability of humans, PretrainZero learns a unified reasoning policy to actively identify reasonable and informative contents from pretraining corpus, and reason to predict these contents by RL. 2) Self-supervised learning: without any verifiable labels, pretrained reward models, or supervised fine-tuning, we directly pretrain reasoners from $3\sim30$B base models on the general Wikipedia corpus using RL, significantly breaking the verification data-wall for general reasoning. 3) Verification scaling: by tackling increasingly challenging masked spans, PretrainZero substantially enhances the general reasoning abilities of pretrained base models. With reinforcement learning on pretraining data, PretrainZero improves Qwen3-4B-Base for 8.43, 5.96 and 10.60 on MMLU-Pro, SuperGPQA and math average benchmarks. In post-training, the pretrained models can also serve as reasoning foundation models for downstream RLVR tasks.

Synthetic Nearest Neighbors (SNN) provides a principled solution to causal matrix completion under missing-not-at-random (MNAR) by exploiting local low-rank structure through fully observed anchor submatrices. However, its effectiveness critically relies on sufficient data availability within each treatment level, a condition that often fails in settings with multiple or complex treatments. In this work, we propose Mixed Synthetic Nearest Neighbors (MSNN), a new entry-wise causal identification estimator that integrates information across treatment levels. We show that MSNN retains the finite-sample error bounds and asymptotic normality guarantees of SNN, while enlarging the effective sample size available for estimation. Empirical results on synthetic and real-world datasets illustrate the efficacy of the proposed approach, especially under data-scarce treatment levels.

In the era of increasingly complex AI models for time series forecasting, progress is often measured by marginal improvements on benchmark leaderboards. However, standard evaluations rely on aggregate metrics (e.g., MSE) that conflate model capability with the intrinsic difficulty of the evaluated instances. To address this, we propose a diagnostic framework anchored in **Spectral Coherence Predictability (SCP)**, which provides an efficient $\mathcal{O}(N\log N)$ per-instance difficulty reference and yields a corresponding linear MSE lower bound. Complementing this, we introduce the **Linear Utilization Ratio (LUR)** to quantify how effectively models exploit linearly predictable structures across frequencies. Experiments on synthetic and real-world benchmarks show that SCP aligns strongly with realized forecasting errors across diverse state-of-the-art forecasters. Using this lens, we uncover ``predictability drift,'' revealing that task difficulty is not static but fluctuates significantly over time and variables. Furthermore, stratified evaluation exposes complementary architectural strengths across distinct frequency bands and difficulty regimes. Overall, we advocate moving beyond leaderboard-style ranking toward a more insightful, predictability-aware evaluation that fosters fairer model comparisons and a deeper understanding of model behavior. Code and data are available at https://github.com/WanjinVon/TS_Predictability.

Large Language Models (LLMs) that continue improving at test-time budgets far beyond their training budgets can solve harder problems by leveraging additional inference compute: we refer to this property as extrapolation. Standard on-policy RL operates on fixed problem distributions and training budgets, giving rise to a distribution shift between train and test that limits the resulting model's extrapolation capabilities. To address this, we introduce RC, an iterative decoding algorithm replacing standard autoregressive decoding that enables models to extrapolate to lengths an order of magnitude longer than those seen during training. RC exploits the asymmetry between summarization and generation capabilities present in LLMs to construct a decoding process that improves consistently over iterations. Its effectiveness can be further increased through training, which amplifies the model’s ability to perform summary-conditioned reasoning while avoiding the challenges of long-horizon RL. Empirically, training a 4B instruction-following model with RC using a 16k-token training budget improves performance on HMMT 2025 from 40% to 70% when evaluated with a 512k-token test budget, substantially surpassing comparably sized LLMs.

Everyday tasks come with a target, and pretraining models around this target is what turns them into experts. In this paper, we study target-oriented language model (LM) pretraining by introducing ***N**euron-**A**ctivated **G**raph Ranking* (NAG-based Ranking), a training-free and interpretable framework for target pretraining data selection. Rather than using black-box representations, our approach directly characterizes each target input by a sparse set of high-impact neurons in any off-the-shelf LLMs. Concretely, we quantify neuron impact and select the most influential neurons across layers into a compact ***N**euron-**A**ctivated **G**raph* (NAG), and rank candidate data by NAG similarity to target examples. We conduct experiments across six benchmarks, where our NAG-based Ranking improves target-oriented pretraining by 4.9\% on average over random sampling, and also outperforms state-of-the-art baselines by 5.3\% accuracy on HellaSwag. It also remains effective under a more applicable multi-target setting, where our best setup surpasses two baselines by 1.1\% and 4.1\%, respectively. Furthermore, we provide a comprehensive analysis on *why* and *how* our NAG works, e.g., deactivating NAG-selected neurons (only 0.12\% of all) causes a 23.5\% performance collapse, and restricting NAG to the final layer incurs a 4.1\% average drop, indicating that NAG captures a sparse ``functional backbone'' for learning target features. We release the code at https://github.com/asillycat/NAG.

Large-scale, high-quality dynamic 3D (4D) assets are essential for learning physically grounded representations, but remain costly to capture and annotate at scale. This limits the viability of supervised 4D learning and motivates zero-shot text-to-4D generation leveraging pretrained diffusion priors. To model complex dynamics, prior methods typically adopt implicit 3D representations (e.g., NeRFs or 3DGS) for their deformation capacity. However, their implicit nature provides limited control over surface topology, which hinders high-fidelity geometry and makes temporally coherent surface reconstruction challenging. To address these limitations, we explore zero-shot text-to-4D mesh generation. However, a structural mismatch arises when combining diffusion-based guidance with topology-constrained meshes: the guidance is noisy and spatially inconsistent, while meshes impose severe topological constraints, making direct vertex-level deformation unstable. In this paper, we introduce TextMesh4D, the first zero-shot framework for text-to-4D that directly generates dynamic meshes by addressing the above challenge at two complementary levels. Geometrically, we shift deformation modeling from vertices to faces via a Jacobian Deformation Field (JDF), enabling topology-aware surface reconstruction through an integrability-enforcing integration formulation. Semantically, we propose a Local-Global Semantic Regularizer (LGSR) that preserves identity over time by jointly constraining local deformation plausibility and global shape consistency. Extensive experiments demonstrate state-of-the-art temporal consistency, structural fidelity, and visual quality, while remaining efficient on a single 24GB GPU.

Time series forecasting (TSF) plays a critical role in decision-making for many real-world applications. Recently, large language model (LLM)- based forecasters have made promising advancements. Despite their effectiveness, existing methods often lack explicit experience accumulation and continual evolution. In this work, we propose MemCast, a learning-to-memory framework that reformulates TSF as an experience-conditioned reasoning task. Specifically, we learn experience from the training set and organize it into a hierarchical memory. This is achieved by summarizing prediction results into historical patterns, distilling inference trajectories into reasoning wisdom, and inducing extracted temporal features into general laws. Furthermore, during inference, we leverage historical patterns to guide the reasoning process and utilize reasoning wisdom to select better trajectories, while general laws serve as criteria for reflective iteration. Additionally, to enable continual evolution, we design a dynamic confidence adaptation strategy that updates the confidence of individual entries without leaking the test set distribution. Extensive experiments on multiple datasets demonstrate that MemCast consistently outperforms previous methods, validating the effectiveness of our approach. Our code is available at https://github.com/Xiaoyu-Tao/MemCast-TS.

Ill-conditioned kernel matrices can make first-order methods for training neural networks converge slowly. We establish non-asymptotic convergence bounds for the Gauss–Newton method in both under- and overparameterized regimes, showing it avoids these conditioning bottlenecks. In the underparameterized setting, Gauss–Newton gradient flow in parameter space induces a Riemannian gradient flow on a low-dimensional submanifold of function space. Using tools from Riemannian optimization, we show that, under an appropriate output scaling, the loss satisfies geodesic Polyak–Lojasiewicz and Lipschitz-smoothness conditions, implying geometric convergence to the optimal in-class predictor at an explicit rate independent of Gram-matrix conditioning. In the overparameterized setting, we identify adaptive, curvature-aware regularization schedules and prove fast geometric convergence to a global optimum for both Gauss–Newton gradient flow and discrete-time Gauss–Newton iterates, with rates independent of the minimum eigenvalue of the neural tangent kernel. Overall, Gauss–Newton can be provably faster in ill-conditioned regimes where first-order methods slow down.

Vision-language foundation models such as CLIP and SigLIP provide widely used representations for multimodal learning systems. While these models are typically compared through downstream performance, such evaluations often do not explain how their representations differ structurally. In this work, we study this problem through the task of *Contrastive Embedding Clustering*: identifying sample subsets that are weakly clustered under one representation but strongly clustered under another. We propose *Kernel Optimization for Discrepancy Analysis (KODA)*, a kernel-based framework for contrastive representation comparison and alignment. KODA constructs unified multimodal kernels through modality-wise kernel composition and formulates discrepancy discovery as a constrained optimization problem that searches for coherent structures in one representation while suppressing coherence in a reference representation. This yields interpretable discrepancy directions associated with specific sample subsets and modality interactions. To scale KODA to large vision-language datasets, we develop randomized low-dimensional approximations of joint kernels using random projections, including Random Fourier Features for shift-invariant kernels. Empirically, KODA identifies consistent and interpretable discrepancy structures across vision-language representations and provides sample subsets for representation alignment. The code is available at https://github.com/yokiwuuu/KODA.