ICML 2026 Papers

Detecting hallucinations in large language models is a critical open problem with significant implications for safety and reliability. While existing hallucination detection methods achieve strong performance in question‑answering tasks, they remain less effective on tasks requiring reasoning. In this work, we revisit hallucination detection through the lens of out‑of‑distribution (OOD) detection, a well‑studied problem in areas like computer vision. Treating next‑token prediction in language models as a classification task allows us to apply OOD techniques, if we bring to bear appropriate modifications to account for the structural differences in large language models. We show that approaches based on OOD detection yield training-free, single-sample based detectors, achieving strong accuracy in hallucination detection in reasoning tasks. Overall, our work suggests that reframing hallucination detection as OOD detection provides a promising and scalable pathway toward language model safety.

Dynamic graph learning, which focuses on modeling the merging, vanishing, and reconnection of nodes and edges, is crucial for real-world applications. In dynamic graphs, node neighborhoods often exhibit diverse and time-evolving topologies, including hierarchical, grid-like, and cyclic patterns. Existing methods typically embed graphs into a single curvature space, which limits the quality of node representations when the embedding geometry is not aligned well with the local graph topology. In this paper, we propose **DyGMoCE**, a **Dy**namic **G**raph Transformer with a **M**ixture **o**f **C**urvature-aware **E**xperts, which efficiently embeds each node at every timestamp into an adaptive curvature space. Specifically, DyGMoCE incorporates a mixture-of-experts framework to both the attention and feed-forward modules, where each expert operates on a Riemannian manifold with a distinct curvature. Then, motivated by the geometric continuity across the experts, we introduce a routing mechanism with a ranking constraint. To improve efficiency, we design a fast Riemannian attention module for DyGMoCE, achieving an average speedup of 27.5% and memory reduction of 52.6%. Notably, the fast Riemannian attention module is broadly applicable to Transformer models with sequence inputs. Extensive experimental results show that DyGMoCE significantly outperforms other state-of-the-art methods.

We study sequential decision-making in partially observable environments against strategic, adaptive opponents, modeled as partially observable Markov games (POMGs). The central challenge is to learn latent dynamics from partial observations while facing an adversary whose behavior depends on the learner's strategy, making standard regret notions inadequate. We prove that an epoch-based optimistic maximum-likelihood algorithm achieves \(\tilde O(\sqrt{T})\) policy regret for fixed problem parameters, with explicit dependence on the horizon, adversary memory, confidence radius, and the aggregate Eluder dimension of the observable-operator class. The algorithm selects one policy per geometrically growing epoch using confidence sets built cumulatively from past data, which keeps the cost of comparing adversary responses across policies logarithmic in \(T\). We also prove a lower bound matching the \(\sqrt{T}\) and aggregate-Eluder-dimension dependence, up to problem-dependent and logarithmic factors. Finally, we extend the framework to horizon-adaptive guarantees and adversaries with geometric fading memory.

Real-world graph datasets often arise from mixtures of populations, where graphs are generated by multiple distinct underlying distributions. In this work, we propose a unified framework that explicitly models graph data as a mixture of probabilistic graph generative models represented by graphons. To characterize and estimate these graphons, we leverage graph moments (motif densities) to cluster graphs generated from the same underlying model. We establish a novel theoretical guarantee, deriving a tighter bound showing that graphs sampled from structurally similar graphons exhibit similar motif densities with high probability. This result enables principled estimation of graphon mixture components. We show how incorporating estimated graphon mixture components enhances two widely used downstream paradigms: graph data augmentation via mixup and graph contrastive learning. By conditioning these methods on the underlying generative models, we develop graphon-mixture-aware mixup (GMAM) and model-aware graph contrastive learning (MGCL). Extensive experiments on both simulated and real-world datasets demonstrate strong empirical performance. In supervised learning, GMAM outperforms existing augmentation strategies, achieving new state-of-the-art accuracy on 6 out of 7 datasets. In unsupervised learning, MGCL performs competitively across seven benchmark datasets and achieves the lowest average rank overall.

Knowledge distillation (KD) is a powerful paradigm for compressing large language models (LLMs), whose effectiveness depends on intertwined choices of divergence direction, optimization strategy, and data regime. We break down the design of existing KD methods and present a unified view that establishes connections between them, reformulating KD as a reweighted log-likelihood objective at the token level. We further propose Hybrid Policy Distillation (HPD), which integrates the complementary advantages of forward and reverse KL to balance mode coverage and mode-seeking behaviors, and combines off-policy data with lightweight approximate on-policy sampling. We validate HPD on long-generation math reasoning as well as short-generation dialogue and code tasks, demonstrating improved optimization stability, computational efficiency, and final performance across diverse model families and scales.

Diffusion-based camouflaged object detection (COD) has recently shown great potential. In contrast to existing approaches that rely on multiple sample steps to refine the predicted masks, we propose CODiff, which reformulates the diffusion process to enable one-step mask prediction while maintaining competitive accuracy. Specifically, we first establish the theoretical feasibility of one-step sampling for COD. Based on this, we design a dedicated network for one-step inference with a global semantic guidance mechanism to guide the denoising process globally and hierarchical condition integration blocks to provide fine-grained structural semantics. In addition, we design a straight-forward regularization to learn better intermediate features by bridging the representation gap between the condition backbone and the diffusion model. Extensive experiments demonstrate that CODiff achieves state-of-the-art performance across multiple benchmarks, improving MAE by over 22\% on the challenging COD10K dataset. Code is available at https://github.com/KiiSooo/CODiff.

Model-based reinforcement learning (MBRL) improves sample efficiency by leveraging learned dynamics models, but often suffers from unstable training due to dynamics model learning mismatch: models are trained on data from historical policies while being queried under the continually updated current policy. This mismatch can cause policy-relevant local model error to remain large even as global prediction error decreases, leading to oscillatory updates. We present a finite-horizon performance analysis that decomposes the policy performance gap into global model error, policy-induced distribution shift, and historical policy mixture effects, showing that minimizing global error alone is insufficient for stable optimization. Motivated by this analysis, we propose Prioritized Model Experience Replay (PMER), a lightweight replay mechanism that prioritizes high-error transitions during dynamics model training. PMER implicitly emphasizes policy-relevant regions without explicit policy distance estimation and integrates seamlessly into Dyna-style MBRL frameworks. Experiments on MuJoCo benchmarks demonstrate improved stability, faster convergence, and higher sample efficiency.

Modern deep learning for asset allocation typically separates forecasting from optimization. We argue this creates a fundamental mismatch where minimizing prediction errors fails to yield robust portfolios. We propose the Signature Informed Transformer to address this by unifying feature extraction and decision making into a single policy. Our model employs path signatures to encode complex path dependencies and introduces a specialized attention mechanism that targets geometric asset relationships. By directly minimizing the Conditional Value at Risk we ensure the training objective aligns with financial goals. We prove that our attention module rigorously amplifies signature derived signals. Experiments across diverse equity universes show our approach significantly outperforms both traditional strategies and advanced forecasting baselines. The code is available at: https://github.com/Yoontae6719/Signature-Informed-Transformer-For-Asset-Allocation

Offline reinforcement learning (RL) offers a powerful paradigm for data-driven control. Compared to model-free approaches, offline model-based RL (MBRL) explicitly learns a world model from a static dataset and uses it as a surrogate simulator, improving data efficiency and enabling potential generalization beyond the dataset support. However, most existing offline MBRL methods follow a two-stage training procedure: first learning a world model by maximizing the likelihood of the observed transitions, then optimizing a policy to maximize its expected return under the learned model. This objective mismatch results in a world model that is not necessarily optimized for effective policy learning. Moreover, we observe that policies learned via offline MBRL often lack robustness during deployment, and small adversarial noise in the environment can lead to significant performance degradation. To address these, we propose a framework that dynamically adapts the world model alongside the policy under a unified learning objective aimed at improving robustness. At the core of our method is a maximin optimization problem, which we solve by innovatively utilizing Stackelberg learning dynamics. We provide theoretical analysis to support our design and introduce computationally efficient implementations. We benchmark our algorithm on twelve noisy D4RL MuJoCo tasks and three stochastic Tokamak Control tasks, demonstrating its state-of-the-art performance. Code is available at https://github.com/Agentic-Intelligence-Lab/ROMBRL.

While the shift toward unified foundation models has revolutionized many deep learning domains, sleep medicine remains largely restricted to task-specific models that focus on localized micro-structure features. These approaches often neglect the rich, multi-modal context of Polysomnography (PSG) and fail to capture the global macro-structure of a full night's sleep. To address this, we introduce SleepMaMi, a Sleep Foundation Model engineered to master both hour-long sleep architectures and fine-grained signal morphologies. Our framework utilizes a hierarchical dual-encoder design: a Macro-Encoder to model full-night temporal dependencies and a Micro-Encoder to capture short-term characteristics from biosignals. Macro-Encoder is trained via Demographic-Guided Contrastive Learning, which aligns overnight sleep patterns with objective subject metadata, such as age and sex, to refine global representations. Micro-Encoder is optimized via a hybrid Masked Autoencoder (MAE) and multi-modal contrastive objective. Pre-trained on a massive corpus of >20,000 PSG recordings (158K hours), SleepMaMi outperforms or matches existing foundation models across a diverse suite of downstream tasks, demonstrating superior generalizability and label-efficient adaptation for clinical sleep analysis.

Audio-Visual Large Language Models (AV-LLMs) face prohibitive computational costs of processing massive, redundant audio-visual tokens. Existing unimodal compression techniques fail to capture the heterogeneous and mutually influential information density of joint audio-visual signals. Furthermore, we identify a fundamental and overlooked theoretical bottleneck in sparse token reduction: positional aliasing. We demonstrate that aggressive sparse sampling on standard position-encoded sequences violates the Nyquist limit relative to the effective token interval, causing phase-wrapping collisions that corrupt temporal monotonicity. To address this, we introduce EchoingPixels, a framework for aliasing-resistant joint token reduction. Our Cross-Modal Semantic Sieve performs extractive selection on the synergistic audio-visual stream, dynamically allocating budgets based on joint-modality saliency rather than fixed per-modality ratios. To resolve positional aliasing, we derive Sync-RoPE, a spectral low-pass filter for Rotary Positional Embeddings that adapts encoding bandwidth to the sparse sampling rate, preserving monotonic temporal relationships in the reduced stream. Experiments show that EchoingPixels achieves performance comparable to full models using only 5-20% of original tokens, validating theoretically grounded sparse learning as a robust solution for efficient AV-LLMs. Code is available at https://github.com/CharlesGong12/EchoingPixels.

Recent work shows that flow-matching networks can be effective for value function estimation in reinforcement learning, but it remains unclear why they work well or whether flow-matching Q-functions differ fundamentally from standard critics. We show that their success is not explained by distributional RL: explicitly modeling return distributions often degrades performance. Instead, we argue that flow-matching Q-functions are effective because they couple a learned velocity field with an integration procedure that is used both during training and to read out Q-values at inference time. This coupling enables robust value prediction through \emph{test-time recovery} from imperfect intermediate estimates where errors dampen out as more integration steps are performed. This mechanism is absent in monolithic critics. Beyond test-time recovery, training with the integration procedure induces more \emph{plastic} representations, allowing critics to represent non-stationary future TD targets without overwriting previous features. We formalize these effects and validate them empirically, showing that flow-matching critics outperform monolithic critics by over $2\times$ in performance and achieve $5$–$10\times$ higher sample efficiency in high-UTD regimes.

Recent advances in online reinforcement learning (RL) for large language models (LLMs) have demonstrated promising performance in complex reasoning tasks. However, they often exhibit an imbalanced exploration–exploitation trade-off, resulting in unstable optimization and sub-optimal performance. We introduce IB-Score, a novel metric grounded in Information Bottleneck (IB) theory that evaluates policy’s exploration-exploitation balance by quantifying the trade-off between step-level reasoning diversity and mutual information shared with the correct answer. Analysis based on IB-Score shows that popular online RL approaches (e.g., GRPO) with common regularization methods fail to consistently maintain balance during training with suboptimal results. To address this, we propose Information Bottleneck-driven Tree-based Policy Optimization (IB-TPO), a principled framework that formulates IB-Score as a fine-grained optimization objective and utilizes a novel IB-guided tree sampling strategy that not only improves the efficiency of online sampling with 50\% more trajectories under same token budget, but also reuses the tree structure for effective IB-Score Monte Carlo estimation. Extensive experiments across standard benchmarks show that our method significantly outperforms GRPO baseline by 2.9% to 3.6% and also outperforms other state-of-the-art online RL approaches.

Recently, Diffusion Transformers (DiTs) have emerged in Real-World Image Super-Resolution (Real-ISR) to generate high-quality textures, yet their heavy inference burden hinders real-world deployment. While Post-Training Quantization (PTQ) is a promising solution for acceleration, existing methods in super-resolution mostly focus on U-Net architectures, whereas generic DiT quantization is typically designed for text-to-image tasks. Directly applying these methods to DiT-based super-resolution models leads to severe degradation of local textures. Therefore, we propose **Q-DiT4SR**, the first PTQ framework specifically tailored for DiT-based Real-ISR. We propose **H-SVD**, a hierarchical SVD that integrates a global low-rank branch with a local block-wise rank-1 branch under a matched parameter budget. We further propose **V**ariance-**a**ware **S**patio-**T**emporal **M**ixed **P**recision: **VaSMP** allocates cross-layer weight bit-widths in a data-free manner based on rate-distortion theory, while **VaTMP** schedules intra-layer activation precision across diffusion timesteps via dynamic programming (DP) with minimal calibration. Experiments on multiple real-world datasets demonstrate that our Q-DiT4SR achieves SOTA performance under both **W4A6** and **W4A4** settings. Notably, the W4A4 quantization configuration reduces model size by **5.8**$\times$ and computational operations by **6.14**$\times$. Our code and models will be available at https://github.com/xunzhang1128/Q-DiT4SR.

Neural Controlled Differential Equations (Neural CDEs) provide a powerful continuous-time framework for sequence modeling, yet the roughness of the driving control path often restricts their efficiency. Standard splines introduce high-frequency variations that force adaptive solvers to take excessively small steps, driving up the Number of Function Evaluations (NFE). We propose a novel approach to Neural CDE path construction that replaces exact interpolation with Kernel and Gaussian Process (GP) smoothing, enabling explicit control over trajectory regularity. To recover details lost during smoothing, we propose an attention-based Multi-View CDE (MV-CDE) and its convolutional extension (MVC-CDE), which employ learnable queries to inform path reconstruction. This framework allows the model to distribute representational capacity across multiple trajectories, each capturing distinct temporal patterns. Empirical results demonstrate that our method, MVC-CDE with GP, achieves state-of-the-art accuracy while significantly reducing NFEs and total inference time compared to spline-based baselines.

Hyperbolic geometry has recently emerged as a powerful representation space for multimodal learning, as it naturally captures hierarchical semantic structure across modalities. Despite this progress, how such representations behave under continual learning poses fundamentally different challenges that remain underexplored. This work provides a geometric perspective on this problem and establishes a theoretical foundation for representation preservation in hyperbolic space, showing that preventing forgetting requires cross-modal invariance under a shared hyperbolic isometry. We further show that forgetting in hyperbolic continual learning involves both semantic relation drift and hierarchy-related distortion, motivating preservation of both cross-modal relational structure and hierarchical geometry. Guided by these insights, a principled continual learning framework is derived that preserves essential geometric structure while allowing effective adaptation to new tasks. Experiments on continual multimodal benchmarks corroborate the effectiveness of the proposed approach.

Continuous ECG monitoring is clinically valuable, but scaling it beyond electrodes to comfortable long-term use motivates contactless mmWave sensing. In practice, mmWave-to-ECG reconstruction is severely constrained by the scarcity of high-quality synchronized recordings. Therefore, we propose \textbf{Cardio-mmFlow}, a Gaussian-prior-free physics-informed flow matching framework that synthesizes mmWave radar signals from clinical ECG. It learns a direct transport trajectory between the latent manifolds of ECG and radar. Considering subject-dependent propagation differences, we incorporate a simplified mass--spring--damper inspired modulation and inject it into the flow dynamics via feature-wise linear modulation for personalization. Extensive experiments show that our system generates high fidelity radar data in both signal and latent domains. It significantly improves zero-shot downstream mmWave to ECG task, and enable Atrial Fibrillation classification with synthetic data. Further analyses evaluate the model interpretability.

Modern language models are trained almost exclusively on token sequences produced by a fixed tokenizer, an external lossless compressor often over UTF‑8 byte sequences, thereby coupling the model to that compressor. This work introduces proxy compression, an alternative training scheme that preserves the efficiency benefits of compressed inputs while providing an end-to-end, raw-byte interface at inference time. During training, one language model is jointly trained on raw byte sequences and compressed views generated by external compressors; through the process, the model learns to internally align compressed sequences and raw bytes. This alignment enables strong transfer between the two formats, even when training predominantly on compressed inputs which are discarded at inference. Extensive experiments on code language modeling demonstrate that proxy compression substantially improves training efficiency and significantly outperforms pure byte-level baselines given fixed compute budgets. As model scale increases, these gains become more pronounced, and proxy-trained models eventually match or rival tokenizer approaches, all while operating solely on raw bytes and retaining the inherent robustness of byte-level modeling.

Reinforcement learning (RL) for large language models (LLMs) relies on imperfect reward supervision, necessitating constraints on policy updates to prevent overfitting. Nevertheless, the widely adopted KL constraint over-penalizes actions with low reference probabilities and lacks the sparsity to discard marginal policy shifts. In contrast, the L1-norm offers a distinct mechanism that is more tolerant of low-probability actions yet strictly suppresses minor probability perturbations. Motivated by this, we propose Magnitude-Regularized Policy Optimization (MRPO), which enforces an L1-norm constraint on policy updates. We demonstrate that MRPO permits substantial probability boosts for low-probability actions and induces sparse updates, ensuring invariance to noise that preserves the top-ranking order. Furthermore, MRPO admits a TRPO-style monotonic improvement bound under standard regularity assumptions and achieves a tighter approach to optimality than KL-based methods in single-step scenarios. Empirically, MRPO delivers exceptional results across diverse scenarios, notably doubling the performance gains of GRPO in preference alignment, outperforming DAPO in mathematical reasoning, and surpassing DPO in offline settings using only binary rewards.