NeurIPS 2025 Papers

Pre-trained language models represented by the Transformer have been proven to possess strong base capabilities, and the representative self-attention mechanism in the Transformer has become a classic in sequence modeling architectures. Different from the work of proposing sequence modeling architecture to improve the efficiency of attention mechanism, this work focuses on the impact of sequence modeling architectures on base capabilities. Specifically, our concern is: How exactly do sequence modeling architectures affect the base capabilities of pre-trained language models? In this work, we first point out that the mixed domain pre-training setting commonly adopted in existing architecture design works fails to adequately reveal the differences in base capabilities among various architectures. To address this, we propose a limited domain pre-training setting with out-of-distribution testing, which successfully uncovers significant differences in base capabilities among architectures at an early stage. Next, we analyze the base capabilities of stateful sequence modeling architectures, and find that they exhibit significant degradation in base capabilities compared to the Transformer. Then, through a series of architecture component analysis, we summarize a key architecture design principle: A sequence modeling architecture need possess full-sequence arbitrary selection capability to avoid degradation in base capabilities. Finally, we empirically validate this principle using an extremely simple Top-1 element selection architecture and further generalize it to a more practical Top-1 chunk selection architecture. Experimental results demonstrate our proposed sequence modeling architecture design principle and suggest that our work can serve as a valuable reference for future architecture improvements and novel designs.

Anomaly detection is a crucial task in data mining, focusing on identifying data points that deviate significantly from the main patterns in the data. This paper introduces Anomaly Detection by an Ensemble of Random Pairs of Hyperspheres (ADERH), a new isolation-based technique leveraging two key observations: (i) anomalies are comparatively rare, and (ii) they typically deviate more strongly from general patterns than normal data points. Drawing on a delta-separation argument, ADERH constructs an ensemble of multi-scale hyperspheres built upon randomly paired data points to identify anomalies. To address inevitable overlaps between anomalous and normal regions in the feature space, ADERH integrates two complementary concepts: Pitch, which highlights points near hypersphere boundaries, and NDensity, which down-weights hyperspheres centered on sparse (and often anomalous) regions. By averaging these local, density-adjusted ``isolation'' indicators across many random subsets, ADERH yields robust anomaly scores that clearly separate normal from abnormal samples. Extensive experiments on diverse real-world datasets show that ADERH consistently outperforms state-of-the-art methods while maintaining linear runtime scalability and stable performance across varying hyperparameter settings.

Gradient-based optimization methods have shown remarkable empirical success, yet their theoretical generalization properties remain only partially understood. In this paper, we establish a generalization bound for gradient flow that aligns with the classical Rademacher complexity bounds for kernel methods—specifically those based on the RKHS norm and kernel trace—through a data-dependent kernel called the loss path kernel (LPK). Unlike static kernels such as NTK, the LPK captures the entire training trajectory, adapting to both data and optimization dynamics, leading to tighter and more informative generalization guarantees. Moreover, the bound highlights how the norm of the training loss gradients along the optimization trajectory influences the final generalization performance. The key technical ingredients in our proof combine stability analysis of gradient flow with uniform convergence via Rademacher complexity. Our bound recovers existing kernel regression bounds for overparameterized neural networks and shows the feature learning capability of neural networks compared to kernel methods. Numerical experiments on real-world datasets validate that our bounds correlate well with the true generalization gap.

This paper studies the *optimistic* variant of Fictitious Play for learning in two-player zero-sum games. While it is known that Optimistic FTRL -- a *regularized* algorithm with a bounded stepsize parameter -- obtains constant regret in this setting, we show for the first time that similar, optimal rates are also achievable *without* regularization: we prove for two-strategy games that Optimistic Fictitious Play (using *any* tiebreaking rule) obtains only *constant regret*, providing surprising new evidence on the ability of *non*-no-regret algorithms for fast learning in games. Our proof technique leverages a geometric view of Optimistic Fictitious Play in the dual space of payoff vectors, where we show a certain energy function of the iterates remains bounded over time. Additionally, we also prove a regret *lower bound* of $\Omega(\sqrt{T})$ for *Alternating* Fictitious Play. In the unregularized regime, this separates the ability of optimism and alternation in achieving $o(\sqrt{T})$ regret.

Zero-shot optimization aims to achieve both generalization and performance gains on solving previously unseen black-box optimization problems over SOTA methods without task-specific tuning. Pre-trained optimization models (POMs) address this challenge by learning a general mapping from task features to optimization strategies, enabling direct deployment on new tasks. In this paper, we identify three essential components that determine the effectiveness of POMs: (1) task feature modeling, which captures structural properties of optimization problems; (2) optimization strategy representation, which defines how new candidate solutions are generated; and (3) the feature-to-strategy mapping mechanism learned during pre-training. However, existing POMs often suffer from weak feature representations, rigid strategy modeling, and unstable training. To address these limitations, we propose EPOM, an enhanced framework for pre-trained optimization. EPOM enriches task representations using a cross-attention-based tokenizer, improves strategy diversity through deformable attention, and stabilizes training by replacing non-differentiable operations with a differentiable crossover mechanism. Together, these enhancements yield better generalization, faster convergence, and more reliable performance in zero-shot black-box optimization.

Novelty detection in large scientific datasets faces two key challenges: the noisy and high-dimensional nature of experimental data, and the necessity of making *statistically robust* statements about any observed outliers. While there is a wealth of literature on anomaly detection via dimensionality reduction, most methods do not produce outputs compatible with quantifiable claims of scientific discovery. In this work we directly address these challenges, presenting the first step towards a unified pipeline for novelty detection adapted for the rigorous statistical demands of science. We introduce AutoSciDACT (Automated Scientific Discovery with Anomalous Contrastive Testing), a general-purpose pipeline for detecting novelty in scientific data. AutoSciDACT begins by creating expressive low-dimensional data representations using a contrastive pre-training, leveraging the abundance of high-quality simulated data in many scientific domains alongside expertise that can guide principled data augmentation strategies. These compact embeddings then enable an extremely sensitive machine learning-based two-sample test using the New Physics Learning Machine (NPLM) framework, which identifies and statistically quantifies deviations in observed data relative to a reference distribution (null hypothesis). We perform experiments across a range of astronomical, physical, biological, image, and synthetic datasets, demonstrating strong sensitivity to small injections of anomalous data across all domains.

Bilevel optimization has garnered significant attention in the machine learning community recently, particularly regarding the development of efficient numerical methods. While substantial progress has been made in developing efficient algorithms for optimistic bilevel optimization, the study of methods for solving Pessimistic Bilevel Optimization (PBO) remains relatively less explored, especially the design of fully first-order, single-loop gradient-based algorithms. This paper aims to bridge this research gap. We first propose a novel smooth approximation to the PBO problem, using penalization and regularization techniques. Building upon this approximation, we then propose SiPBA (Single-loop Pessimistic Bilevel Algorithm), a new gradient-based method specifically designed for PBO which avoids second-order derivative information or inner-loop iterations for subproblem solving. We provide theoretical validation for the proposed smooth approximation scheme and establish theoretical convergence for the algorithm SiPBA. Numerical experiments on synthetic examples and practical applications demonstrate the effectiveness and efficiency of SiPBA.

Vector-Quantized Variational Autoencoders (VQVAEs) have enabled strong performance in generative modeling by mapping continuous data to learnable codes. In this work, we identify a surprising yet consistent phenomenon that we term \emph{dimensional collapse}: despite using high-dimensional embeddings, VQVAEs tend to compress their representations into a much smaller subspace, typically only 4 to 10 dimensions. We provide an in-depth analysis of this phenomenon and reveal its relation to model performance and learning dynamics. Interestingly, VQVAEs naturally gravitate toward this low-dimensional regime, and enforcing higher-dimensional usage (e.g., via rank regularization) could lead to degraded performance. To overcome this low-dimensionality limitation, we propose \textbf{Divide-and-Conquer VQ (DCVQ)}, which partitions the latent space into multiple low-dimensional subspaces, each quantized independently. By design, each subspace respects the model’s preference for low dimensionality, while their combination expands the overall capacity. Our results show that DCVQ overcomes the inherent dimensional bottleneck and achieves improved reconstruction quality across image datasets.

Large Reasoning Models (LRMs) have become powerful tools for complex problem solving, but their structured reasoning pathways can lead to unsafe outputs when exposed to harmful prompts. Existing safety alignment methods reduce harmful outputs but can degrade reasoning depth, leading to significant trade-offs in complex, multi-step tasks, and remain vulnerable to sophisticated jailbreak attacks. To address this, we introduce SAFEPATH, a lightweight alignment method that fine-tunes LRMs to emit a short, 8-token Safety Primer at the start of their reasoning, in response to harmful prompts, while leaving the rest of the reasoning process unsupervised. Empirical results across multiple benchmarks indicate that SAFEPATH effectively reduces harmful outputs while maintaining reasoning performance. Specifically, SAFEPATH reduces harmful responses by up to 90.0\% and blocks 83.3\% of jailbreak attempts in the DeepSeek-R1-Distill-Llama-8B model, while requiring 295.9x less compute than Direct Refusal and 314.1x less than SafeChain. We further introduce a zero-shot variant that requires no fine-tuning. In addition, we provide a comprehensive analysis of how existing methods in LLMs generalize, or fail, when applied to reasoning-centric models, revealing critical gaps and new directions for safer AI.

Hemodynamics has a substantial influence on normal cardiovascular growth and disease formation, but requires time-consuming simulations to obtain. Deep Learning algorithms to rapidly predict hemodynamics parameters can be very useful, but their development is hindered by the lack of large dataset on anatomic geometries and associated fluid dynamics. This paper presents a new large-scale dataset of intracranial aneurysm (IA) geometries and hemodynamics to support the development of neural operators to solve geometry-dependent flow governing partial differential equations. The dataset includes 14,000 steady-flow cases and 200 pulsatile-flow cases simulated with computational fluid dynamics. All cases are computed using a laminar flow setup with more than 3 million cells. Boundary conditions are defined as a parabolic velocity profile with a realistic waveform over time at the inlet, and geometry-dependent mass flow split ratios at the two downstream outlets. The geometries are generated by a deep generative model trained on a cohort of 109 real IAs located at the middle cerebral artery bifurcation, capturing a wide range of geometric variations in both aneurysm sacs and parent vessels. Simulation results shows substantial influence of geometry on fluid forces and flow patterns. In addition to surface mesh files, the dataset provides volume data of velocity, pressure, and wall shear stresses (WSS). For transient cases, spatial and temporal gradients of velocity and pressure are also included. The dataset is tested with PointNet and graph U-Nets for WSS prediction, which showed relative L2 loss of 4.67\% for normalized WSS pattern.

World Models have vastly permeated the field of Reinforcement Learning. Their ability to model the transition dynamics of an environment have led to tremendous improvements in sample efficiency for online RL. Among them, the most notorious example is Dreamer, a model that learns to act in a diverse set of image-based environments. In this paper, we leverage similarity search and stochastic representations to approximate a world model without a training procedure. We establish a comparison with PlaNet, a well-established world model of the Dreamer family. We evaluate the models on the quality of latent reconstruction and on the perceived similarity of the reconstructed image, on both next-step and long horizon dynamics prediction. The results of our study demonstrate that a search-based world model is comparable to a training based one in both cases. Notably, our model shows stronger performance in long-horizon prediction with respect to the baseline on a range of visually different environments.

Recent work in Bayesian Experiment Design (BED) has shown the value of using Deep Learning (DL) to obtain highly efficient adaptive experiment designs. In this paper, we argue that a central bottleneck of DL training for BED is belief explosion. Specifically, as an agent progresses deeper into an experiment, the effective number of realisable beliefs grows enormously, placing significant sampling burdens on offline training schemes in an effort to gather experience from all regions of belief space. We argue that choosing an appropriate inductive bias for actor/critic networks is a critical component in mitigating the effects of belief explosion and has so far been overlooked in the BED literature. We show how Graph Neural Networks are particularly well-suited for BED DL training due to their domain permutation equivariance properties, resulting in multiple orders of magnitude improvement to sample efficiency compared to naive parameterizations.

The meteoric rise of Artificial Intelligence (AI), with its rapidly expanding market capitalization, presents both transformative opportunities and critical challenges. Chief among these is the urgent need for a new, unified paradigm for trustworthy evaluation, as current benchmarks increasingly reveal critical vulnerabilities. Issues like data contamination and selective reporting by model developers fuel hype, while inadequate data quality control can lead to biased evaluations that, even if unintentionally, may favor specific approaches. As a flood of participants enters the AI space, this "Wild West" of assessment makes distinguishing genuine progress from exaggerated claims exceptionally difficult. Such ambiguity blurs scientific signals and erodes public confidence, much as unchecked claims would destabilize financial markets reliant on credible oversight from agencies like Moody's. In high-stakes human examinations (e.g., SAT, GRE), substantial effort is devoted to ensuring fairness and credibility; why settle for less in evaluating AI, especially given its profound societal impact? This position paper argues that a laissez-faire approach is untenable. For true and sustainable AI advancement, we call for a paradigm shift to a unified, live, and quality-controlled benchmarking framework—robust by construction rather than reliant on courtesy or goodwill. Accordingly, we dissect the systemic flaws undermining today’s evaluation ecosystem and distill the essential requirements for next-generation assessments. To concretize this position, we introduce the idea of PeerBench, a community-governed, proctored evaluation blueprint that seeks to improve security and credibility through sealed execution, item banking with rolling renewal, and delayed transparency. PeerBench is presented as a complementary, certificate-grade layer alongside open benchmarks, not a replacement. We discuss trade-offs and limits and call for further research on mechanism design, governance, and reliability guarantees. Our goal is to lay the groundwork for evaluations that restore integrity and deliver genuinely trustworthy measures of AI progress.

Offline safe reinforcement learning aims to learn policies that satisfy predefined safety constraints from static datasets. Existing sequence-model-based methods condition action generation on symmetric input tokens for return-to-go and cost-to-go, neglecting their intrinsic asymmetry: RTG serves as a flexible performance target, while CTG should represent a rigid safety boundary. This symmetric conditioning leads to unreliable constraint satisfaction, especially when encountering out-of-distribution cost trajectories. To address this, we propose Boundary-to-Region (B2R), a framework that enables asymmetric conditioning through cost signal realignment . B2R redefines CTG as a boundary constraint under a fixed safety budget, unifying the cost distribution of all feasible trajectories while preserving reward structures. Combined with rotary positional embeddings , it enhances exploration within the safe region. Experimental results show that B2R satisfies safety constraints in 35 out of 38 safety-critical tasks while achieving superior reward performance over baseline methods. This work highlights the limitations of symmetric token conditioning and establishes a new theoretical and practical approach for applying sequence models to safe RL.

Gradient clipping is increasingly important in centralized learning (CL) and federated learning (FL). Many works focus on its optimization properties under strong assumptions involving Gaussian noise and standard smoothness. However, practical machine learning tasks often only satisfy weaker conditions, such as heavy-tailed noise and $(L_0, L_1)$-smoothness. To bridge this gap, we propose a high-probability analysis for clipped Stochastic Gradient Descent (SGD) under these weaker assumptions. Our findings show a better convergence rate than existing ones can be achieved, and our high-probability analysis does not rely on the bounded gradient assumption. Moreover, we extend our analysis to FL, where a gap remains between expected and high-probability convergence, which the naive clipped SGD cannot bridge. Thus, we design a new \underline{Fed}erated \underline{C}lipped \underline{B}atched \underline{G}radient (FedCBG) algorithm, and prove the convergence and generalization bounds with high probability for the first time. Our analysis reveals the trade-offs between the optimization and generalization performance. Extensive experiments demonstrate that \methodname{} can generalize better to unseen client distributions than state-of-the-art baselines.

Recent advances in artificial intelligence have significantly impacted image retrieval tasks, yet Patent-Product Image Retrieval (PPIR) has received limited attention. PPIR, which retrieves patent images based on product images to identify potential infringements, presents unique challenges: (1) both product and patent images often contain numerous categories of artificial objects, but models pre-trained on standard datasets exhibit limited discriminative power to recognize some of those unseen objects; and (2) the significant domain gap between binary patent line drawings and colorful RGB product images further complicates similarity comparisons for product-patent pairs. To address these challenges, we formulate it as an open-set image retrieval task and introduce a comprehensive Patent-Product Image Retrieval Dataset (PPIRD) including a test set with 439 product-patent pairs, a retrieval pool of 727,921 patents, and an unlabeled pre-training set of 3,799,695 images. We further propose a novel Intermediate Domain Alignment and Morphology Analogy (IDAMA) strategy. IDAMA maps both image types to an intermediate sketch domain using edge detection to minimize the domain discrepancy, and employs a Morphology Analogy Filter to select discriminative patent images based on visual features via analogical reasoning. Extensive experiments on PPIRD demonstrate that IDAMA significantly outperforms baseline methods (+7.58 mAR) and offers valuable insights into domain mapping and representation learning for PPIR. (The PPIRD dataset is available at: \href{https://loslorien.github.io/idama-project/}{https://loslorien.github.io/idama-project/})

The rapid advancements in large language models (LLMs), particularly in their reasoning capabilities, hold transformative potential for addressing complex challenges and boosting scientific discovery in atmospheric science. However, leveraging LLMs effectively in this domain requires a robust and comprehensive evaluation benchmark. Toward this end, we present AtmosSci-Bench, a novel benchmark designed to systematically assess LLM performance across five core categories of atmospheric science problems: hydrology, atmospheric dynamics, atmospheric physics, geophysics, and physical oceanography. AtmosSci-Bench features a dual-format design comprising both multiple-choice questions (MCQs) and open-ended questions (OEQs), enabling scalable automated evaluation alongside deeper analysis of conceptual understanding. We employ a template-based MCQ generation framework to create diverse, graduate-level problems with symbolic perturbation, while OEQs are used to probe open-ended reasoning. We conduct a comprehensive evaluation of representative LLMs, categorized into four groups: instruction-tuned models, advanced reasoning models, math-augmented models, and domain-specific climate models. Our analysis provides some interesting insights into the reasoning and problem-solving capabilities of LLMs in atmospheric science. We believe AtmosSci-Bench can serve as a critical step toward advancing LLM applications in climate services by offering a standard and rigorous evaluation framework. The source code of AtmosSci-Bench is available at https://github.com/Relaxed-System-Lab/AtmosSci-Bench.

Monocular 3D tracking aims to capture the long-term motion of pixels in 3D space from a single monocular video and has witnessed rapid progress in recent years. However, we argue that the existing monocular 3D tracking methods still fall short in separating the camera motion from foreground dynamic motion and cannot densely track newly emerging dynamic subjects in the videos. To address these two limitations, we propose TrackingWorld, a novel pipeline for dense 3D tracking of almost all pixels within a world-centric 3D coordinate system. First, we introduce a tracking upsampler that efficiently lifts the arbitrary sparse 2D tracks into dense 2D tracks. Then, to generalize the current tracking methods to newly emerging objects, we apply the upsampler to all frames and reduce the redundancy of 2D tracks by eliminating the tracks in overlapped regions. Finally, we present an efficient optimization-based framework to back-project dense 2D tracks into world-centric 3D trajectories by estimating the camera poses and the 3D coordinates of these 2D tracks. Extensive evaluations on both synthetic and real-world datasets demonstrate that our system achieves accurate and dense 3D tracking in a world-centric coordinate frame.

Hallucinations present a significant challenge in the development and evaluation of large language models (LLMs), directly affecting their reliability and accuracy. While notable advancements have been made in research on textual and visual hallucinations, there is still a lack of a comprehensive benchmark for evaluating auditory hallucinations in large audio language models (LALMs). To fill this gap, we introduce **AHa-Bench**, a systematic and comprehensive benchmark for audio hallucinations. Audio data, in particular, uniquely combines the multi-attribute complexity of visual data with the semantic richness of textual data, leading to auditory hallucinations that share characteristics with both visual and textual hallucinations. Based on the source of these hallucinations, AHa-Bench categorizes them into semantic hallucinations, acoustic hallucinations, and semantic-acoustic confusion hallucinations. In addition, we systematically evaluate seven open-source local perception language models (LALMs), demonstrating the challenges these models face in audio understanding, especially when it comes to jointly understanding semantic and acoustic information. Through the development of a comprehensive evaluation framework, AHa-Bench aims to enhance the robustness and stability of LALMs, fostering more reliable and nuanced audio understanding in LALMs. The benchmark dataset is available at \url{https://huggingface.co/datasets/ahabench/AHa-Bench}.