Machine Learning
Adaptive Moments are Surprisingly Effective for Plug-and-Play Diffusion Sampling
Christian Belardi, Justin Lovelace, Kilian Q. Weinberger et al. (+1 more)
Guided
diffusion sampling relies on approximating often intractable likelihood scores, which introduces significant noise into the sampling dynamics. We propose using adaptive moment estimation to stabilize these noisy likelihood scores during sampling. Despite its simplicity, our approach achieves state-of-the-art results on image restoration and class-conditional generation tasks, outperforming more complicated methods, which are often computationally more expensive. We provide empirical analysis of our method on both synthetic and real data, demonstrating that mitigating gradient noise through adaptive moments offers an effective way to improve
alignment.
Read the full abstract ▾
All authors Christian Belardi, Justin Lovelace, Kilian Q. Weinberger, Carla P. Gomes
Machine Learning
Conservative Continuous-Time Treatment Optimization
Nora Schneider, Georg Manten and Niki Kilbertus
Finding optimal treatment plans in continuous time is crucial for personalized medicine. Existing methods, however, do not explicitly optimize treatment plans and instead focus solely on predicting treatment effects. Naive optimization over these treatment-response models can exploit model errors, resulting in plans that appear optimal under the learned model but perform poorly in practice. To mitigate model exploitation, we propose SigCTO (Signature-based Conservative Treatment Optimization), a conservative framework for continuous-time treatment optimization. We derive a path-space regularizer based on signature-kernel maximum mean discrepancy that penalizes treatment plans for which predicted and true responses differ. Under suitable assumptions, the regularizer allows for consistent estimation, and the regularized objective upper bounds the true treatment cost. Experiments on pharmacometric simulators show that SigCTO finds treatment plans with lower ground-truth costs than state-of-the-art baselines. On real-world ICU data, SigCTO predicts treatment cost more accurately than baselines, and the regularizer detects prediction errors.
Read the full abstract ▾
Machine Learning
SOMP: Scalable Gradient Inversion for Large Language Models via Subspace-Guided Orthogonal Matching Pursuit
Yibo Li and Qiongxiu Li
Gradient inversion attacks reveal that private training text can be reconstructed from shared gradients, posing a
privacy risk to large
language models (LLMs). While prior methods perform well in small-batch settings, scaling to larger batch sizes and longer sequences remains challenging due to severe signal mixing, high computational cost, and degraded fidelity. We present SOMP (Subspace-Guided Orthogonal Matching Pursuit), a scalable gradient inversion framework that casts text recovery from aggregated gradients as a sparse signal recovery problem. Our key insight is that aggregated
transformer gradients retain exploitable head-wise
geometric structure together with sample-level sparsity. SOMP leverages these properties to progressively narrow the search space and disentangle mixed signals without exhaustive search. Experiments across multiple LLM families, model scales, and five
languages show that SOMP consistently outperforms prior methods in the aggregated-gradient regime.For long sequences at batch size B=16, SOMP achieves substantially higher reconstruction fidelity than strong baselines, while remaining computationally competitive. Even under extreme aggregation (up to B=128), SOMP still recovers meaningful text, suggesting that
privacy leakage can persist in regimes where prior attacks become much less effective.
Read the full abstract ▾
Machine Learning
pADAM: A Plug-and-Play All-in-One Diffusion Architecture for Multi-Physics Learning
Amirhossein Mollaali, Bongseok Kim, Christian Moya et al. (+1 more)
Generalizing across disparate physical laws remains a fundamental challenge for artificial intelligence in science. Existing deep-learning solvers are largely confined to single-equation settings, limiting transfer across physical regimes and inference tasks. Here we introduce pADAM, a unified generative framework that learns a shared probabilistic prior across heterogeneous partial differential equation families. Through a learned joint distribution of system states and, where applicable, physical parameters, pADAM supports forward prediction and inverse inference within a single architecture without retraining. Across
benchmarks ranging from scalar
diffusion to nonlinear Navier–Stokes equations, pADAM achieves accurate inference even under sparse observations. Combined with conformal prediction, it also provides reliable uncertainty quantification with coverage guarantees. In addition, pADAM performs probabilistic model selection from only two sparse snapshots, identifying governing laws through its learned generative representation. These results highlight the potential of generative multi-physics modeling for unified and uncertainty-aware scientific inference.
Read the full abstract ▾
All authors Amirhossein Mollaali, Bongseok Kim, Christian Moya, Guang Lin
Machine Learning
A Practical Algorithm for Feature-Rich, Non-Stationary Bandit Problems
Wei Min Loh, Sajib Kumer Sinha, Ankur Agarwal et al. (+1 more)
Contextual bandits are incredibly useful in many practical problems. We go one step further by devising a more realistic problem that combines: (1) contextual bandits with dense arm features, (2) non-linear reward functions, and (3) a generalization of correlated bandits where reward distributions change over time but the degree of correlation maintains. This formulation lends itself to a wider set of applications such as recommendation tasks. To solve this problem, we introduce conditionally coupled contextual C3 Thompson sampling for Bernoulli bandits. It combines an improved Nadaraya-Watson
estimator on an
embedding space with Thompson sampling that allows online learning without retraining. Empirical results show that C3 outperforms the next best
algorithm by 5.7% lower average cumulative regret on four OpenML tabular
datasets as well as demonstrating a 12.4% click lift on Microsoft News
Dataset (MIND) compared to other
algorithms.
Read the full abstract ▾
All authors Wei Min Loh, Sajib Kumer Sinha, Ankur Agarwal, Pascal Poupart
Machine Learning · Cross-listed from cs.GT
Finding Common Ground in a Sea of Alternatives
Jay Chooi, Paul Gölz, Ariel D. Procaccia et al. (+2 more)
We study the problem of selecting a statement that finds common ground across diverse population preferences. Generative AI is uniquely suited for this task because it can access a practically infinite set of statements, but AI systems like the Habermas machine leave the choice of generated statement to a voting rule. What it means for this rule to find common ground, however, is not well-defined. In this work, we propose a formal model for finding common ground in the infinite alternative setting based on the proportional veto core from social choice. To provide guarantees relative to these infinitely many alternatives and a large population, we wish to satisfy a notion of proportional veto core using only query access to the unknown distribution of alternatives and voters. We design an efficient sampling-based
algorithm that returns an alternative in the (approximate) proportional veto core with high probability and prove matching lower bounds, which show that no
algorithm can do the same using fewer queries. On a synthetic
dataset of preferences over text, we confirm the effectiveness of our sampling-based
algorithm and compare other social choice methods as well as LLM-based methods in terms of how reliably they produce statements in the proportional veto core.
Read the full abstract ▾
All authors Jay Chooi, Paul Gölz, Ariel D. Procaccia, Benjamin Schiffer, Shirley Zhang
Machine Learning · Cross-listed from cs.CL
Probing Cultural Signals in Large Language Models through Author Profiling
Valentin Lafargue, Ariel Guerra-Adames, Emmanuelle Claeys et al. (+2 more)
Large
language models (LLMs) are increasingly deployed in applications with societal impact, raising concerns about the cultural biases they encode. We probe these representations by evaluating whether LLMs can perform author profiling from song lyrics in a zero-shot setting, inferring singers' gender and ethnicity without task-specific fine-tuning. Across several open-source models evaluated on more than 10,000 lyrics, we find that LLMs achieve non-trivial profiling performance but demonstrate systematic cultural
alignment: most models default toward North American ethnicity, while DeepSeek-1.5B aligns more strongly with Asian ethnicity. This finding emerges from both the models' prediction distributions and an analysis of their generated rationales. To quantify these disparities, we introduce two
fairness metrics, Modality Accuracy Divergence (MAD) and Recall Divergence (RD), and show that Ministral-8B displays the strongest ethnicity bias among the evaluated models, whereas Gemma-12B shows the most balanced behavior. Our code is available on [GitHub](https://github.com/ValentinLafargue/CulturalProbingLLM) and results on [HuggingFace](https://huggingface.co/
datasets/ValentinLAFARGUE/AuthorProfilingResults).
Read the full abstract ▾
All authors Valentin Lafargue, Ariel Guerra-Adames, Emmanuelle Claeys, Elouan Vuichard, Jean-Michel Loubes
Machine Learning · Cross-listed from math.DS
Data-driven forced response analysis with min-max representations of nonlinear restoring forces
Akira Saito and Hiromu Fujita
This paper discusses a novel data-driven nonlinearity identification method for mechanical systems with nonlinear restoring forces such as
polynomial, piecewise-linear, and general displacement-dependent nonlinearities. The proposed method is built upon the universal approximation
theorem that states that a nonlinear function can be approximated by a linear combination of activation functions in artificial neural network framework. The proposed approach utilizes piecewise linear springs with initial gaps to act as the activation functions of the
neurons of artificial neural networks. A library of piecewise linear springs with initial gaps are constructed, and the contributions of the springs on the nonlinear restoring force are determined by solving the linear regression problems. The piecewise linear springs are realized by combinations of min and max functions with biases. The proposed method is applied to a Duffing oscillator with cubic stiffness, and a piecewise linear oscillator with a gap and their nonlinearities are successfully determined from their free responses. The obtained models are then used for conducting forced response analysis and the results match well with those of the original system. The method is then applied to experimentally-obtained free response data of a cantilevered plate that is subjected to
magnetic restoring force, and successfully finds the piecewise linear representation of the
magnetic force. It is also shown that the obtained model is capable of accurately capturing the steady-state response of the system subject to harmonic base excitation.
Read the full abstract ▾
Machine Learning
Bayesian Inference of Psychometric Variables From Brain and Behavior in Implicit Association Tests
Christian A. Kothe, Sean Mullen, Michael V. Bronstein et al. (+7 more)
Objective. We establish a principled method for inferring mental health related psychometric variables from neural and behavioral data using the Implicit Association Test (IAT) as the data generation engine, aiming to overcome the limited predictive performance (typically under 0.7 AUC) of the gold-standard D-score method, which relies solely on reaction times. Approach. We propose a sparse hierarchical
Bayesian model that leverages multi-modal data to predict experiences related to mental illness symptoms in new participants. The model is a multivariate generalization of the D-score with trainable parameters, engineered for parameter efficiency in the small-cohort regime typical of IAT studies. Data from two IAT variants were analyzed: a suicidality-related E-IAT ($n=39$) and a psychosis-related PSY-IAT ($n=34$). Main Results. Our approach overcomes a high inter-individual variability and low within-session effect size in the
dataset, reaching AUCs of 0.73 (E-IAT) and 0.76 (PSY-IAT) in the best modality configurations, though corrected 95% confidence intervals are wide ($\pm 0.18$) and results are marginally significant after FDR correction ($q=0.10$). Restricting the E-IAT to MDD participants improves AUC to 0.79 $[0.62, 0.97]$ (significant at $q=0.05$). Performance is on par with the best reference methods (shrinkage LDA and EEGNet) for each task, even when the latter were adapted to the task, while the proposed method was not. Accuracy was substantially above near-chance D-scores (0.50-0.53 AUC) in both tasks, with more consistent cross-task performance than any single reference method. Significance. Our framework shows promise for enhancing IAT-based assessment of experiences related to entrapment and psychosis, and potentially other mental health conditions, though further validation on larger and independent cohorts will be needed to establish clinical utility.
Read the full abstract ▾
All authors Christian A. Kothe, Sean Mullen, Michael V. Bronstein, Grant Hanada, Marcelo Cicconet, Aaron N. McInnes, Tim Mullen, Marc Aafjes, Scott R. Sponheim, Alik S. Widge
Machine Learning
SpecMoE: Spectral Mixture-of-Experts Foundation Model for Cross-Species EEG Decoding
Davy Darankoum, Chloé Habermacher, Julien Volle et al. (+1 more)
Decoding the orchestration of neural activity in electroencephalography (
EEG) signals is a central challenge in bridging neuroscience with artificial intelligence. Foundation models have made strides in generalized
EEG decoding, yet many existing frameworks primarily relying on separate temporal and spectral masking of raw signals during self-supervised pretraining. Such strategies often tend to bias learning toward high-frequency oscillations, as low-frequency rhythmic patterns can be easily inferred from the unmasked signal. We introduce a foundation model that utilizes a novel Gaussian-smoothed masking scheme applied to short-time
Fourier transform (STFT) maps. By jointly applying time, frequency, and time-frequency Gaussian masks, we make the reconstruction task much more challenging, forcing the model to learn intricate neural patterns across both high- and low-frequency domains. To effectively recover signals under this aggressive masking strategy, we design SpecHi-Net, a U-shaped hierarchical architecture with multiple encoding and decoding stages. To accelerate large-scale pretraining, we partition the data into three subsets, each used to train an independent expert model. We then combine these models through SpecMoE, a mixture of experts framework guided by a learned spectral gating mechanism. SpecMoE achieves state-of-the-art performance across a diverse set of
EEG decoding tasks, including sleep staging,
emotion recognition, motor imagery classification, abnormal signal detection, and drug effect prediction. Importantly, the model demonstrates strong cross-species and cross-subject generalization, maintaining high accuracy on both human and murine
EEG datasets.
Read the full abstract ▾
All authors Davy Darankoum, Chloé Habermacher, Julien Volle, Sergei Grudinin
Machine Learning
Understanding Quantization of Optimizer States in LLM Pre-training: Dynamics of State Staleness and Effectiveness of State Resets
Kristi Topollai and Anna Choromanska
Quantizing optimizer states is becoming an important ingredient of
memory-efficient large-scale pre-training, but the resulting optimizer dynamics remain only partially understood. We study low-precision exponential moving average (EMA) optimizer states and show how quantization can cause many nominal updates to round back to the same stored value, making the state effectively stale and slowing adaptation beyond what the nominal decay would suggest. We then develop a simple predictive model of stalling that estimates one-step stalling probabilities and characterizes how stalling builds up over time after the initialization. This perspective provides a mechanistic explanation for why optimizer-state resets help in low precision: once a quantized EMA becomes effectively stale, resetting it can temporarily restore responsiveness. Motivated by this picture, we derive a simple theory-guided method for choosing useful reset periods, showing that in low precision the key question is not only whether resets help, but when they should be applied. Experiments in controlled simulations and LLM pre-training show that suitable reset schedules recover the performance lost to low-precision state storage while substantially reducing optimizer-state
memory.
Read the full abstract ▾
Machine Learning
GeMA: Learning Latent Manifold Frontiers for Benchmarking Complex Systems
Jia Ming Li, Anupriya and Daniel J. Graham
Benchmarking the performance of complex systems such as rail networks, renewable generation assets and national economies is central to transport planning, regulation and macroeconomic analysis. Classical frontier methods, notably Data Envelopment Analysis (DEA) and
Stochastic Frontier Analysis (SFA), estimate an efficient frontier in the observed input-output space and define efficiency as distance to this frontier, but rely on restrictive assumptions on the production set and only indirectly address heterogeneity and scale effects. We propose
Geometric Manifold Analysis (GeMA), a latent
manifold frontier framework implemented via a productivity-
manifold variational autoencoder (ProMan-VAE). Instead of specifying a frontier function in the observed space, GeMA represents the production set as the boundary of a low-dimensional
manifold embedded in the joint input-output space. A split-head encoder learns latent variables that capture technological structure and operational inefficiency. Efficiency is evaluated with respect to the learned
manifold, endogenous peer groups arise as clusters in latent technology space, a quotient construction supports scale-invariant benchmarking, and a local certification radius, derived from the decoder Jacobian and a Lipschitz bound, quantifies the
geometric robustness of efficiency scores. We validate GeMA on synthetic data with non-convex frontiers, heterogeneous technologies and scale bias, and on four real-world case studies: global urban rail systems (COMET), British rail operators (ORR), national economies (Penn World Table) and a high-frequency wind-farm
dataset. Across these domains GeMA behaves comparably to established methods when classical assumptions hold, and provides additional insight in settings with pronounced heterogeneity, non-convexity or size-related bias.
Read the full abstract ▾
Machine Learning
The Cost of Reasoning: Chain-of-Thought Induces Overconfidence in Vision-Language Models
Robert Welch, Emir Konuk and Kevin Smith
Vision-
language models (VLMs) are increasingly deployed in high-stakes settings where reliable uncertainty quantification (UQ) is as important as predictive accuracy. Extended
reasoning via chain-of-thought (CoT) prompting or
reasoning-trained models has become ubiquitous in modern VLM pipelines, yet its effect on UQ reliability remains poorly understood. Our results show that
reasoning tends to degrade the quality of many uncertainty estimates, even when it improves task accuracy. We identify implicit answer conditioning as the primary mechanism: as
reasoning traces converge on a conclusion before the final answer is generated,
token probabilities increasingly reflect consistency with the model's own
reasoning trace rather than uncertainty about correctness. In effect, the model becomes overconfident in its answer. In contrast, agreement-based consistency remains robust and often improves under
reasoning, making it a practical choice for uncertainty estimation in
reasoning-enabled VLMs.
Read the full abstract ▾
Machine Learning
Federated Learning with Multi-Partner OneFlorida+ Consortium Data for Predicting Major Postoperative Complications
Yuanfang Ren, Varun Sai Vemuri, Zhenhong Hu et al. (+6 more)
Background: This study aims to develop and validate
federated learning models for predicting major postoperative complications and mortality using a large multicenter
dataset from the OneFlorida Data Trust. We hypothesize that
federated learning models will offer robust generalizability while preserving data
privacy and security. Methods: This retrospective, longitudinal, multicenter cohort study included 358,644 adult patients admitted to five healthcare institutions, who underwent 494,163 inpatient major
surgical procedures from 2012-2023. We developed and internally and externally validated
federated learning models to predict the postoperative risk of intensive care unit (ICU) admission, mechanical ventilation (MV) therapy, acute kidney injury (AKI), and in-hospital mortality. These models were compared with local models trained on data from a single center and central models trained on a pooled
dataset from all centers. Performance was primarily evaluated using area under the receiver operating characteristics curve (AUROC) and the area under the precision-recall curve (AUPRC) values. Results: Our
federated learning models demonstrated strong predictive performance, with AUROC scores consistently comparable or superior performance in terms of AUROC and AUPRC across all outcomes and sites. Our
federated learning models also demonstrated strong generalizability, with comparable or superior performance in terms of both AUROC and AUPRC compared to the best local learning model at each site. Conclusions: By leveraging multicenter data, we developed robust, generalizable, and
privacy-preserving predictive models for major postoperative complications and mortality. These findings support the feasibility of
federated learning in clinical decision support systems.
Read the full abstract ▾
All authors Yuanfang Ren, Varun Sai Vemuri, Zhenhong Hu, Benjamin Shickel, Ziyuan Guan, Tyler J. Loftus, Parisa Rashidi, Tezcan Ozrazgat-Baslanti, Azra Bihorac
Machine Learning
Novelty-Driven Target-Space Discovery in Automated Electron and Scanning Probe Microscopy
Utkarsh Pratiush, Kamyar Barakati, Boris N. Slautin et al. (+4 more)
Modern automated microscopy faces a fundamental discovery challenge: in many systems, the most important scientific information does not reside in the immediately visible image features, but in the target space of sequentially acquired spectra or functional responses, making it essential to develop strategies that can actively search for new behaviors rather than simply optimize known objectives. Here, we developed a deep-kernel-learning BEACON framework that is explicitly designed to guide discovery in the target space by learning structure-property relationships during the experiment and using that evolving model to seek diverse response regimes. We first established the method through demonstration workflows built on pre-acquired ground-truth
datasets, which enabled direct benchmarking against classical acquisition strategies and allowed us to define a set of monitoring functions for comparing exploration quality, target-space coverage, and surrogate-model behavior in a transparent and reproducible manner. This benchmarking framework provides a practical basis for evaluating discovery-driven
algorithms, not just optimization performance. We then operationalized and deployed the workflow on STEM, showing that the approach can transition from offline validation to real experimental implementation. To support adoption and extension by the broader community, the associated notebooks are available, allowing users to reproduce the workflows, test the
benchmarks, and adapt the method to their own instruments and
datasets.
Read the full abstract ▾
All authors Utkarsh Pratiush, Kamyar Barakati, Boris N. Slautin, Catherine C. Bodinger, Christopher D. Lowe, Brandi M. Cossairt, Sergei V. Kalinin
Machine Learning · Cross-listed from math.ST
High-dimensional estimation with missing data: Statistical and computational limits
Kabir Aladin Verchand, Ankit Pensia, Saminul Haque et al. (+1 more)
We consider computationally-efficient estimation of population parameters when observations are subject to missing data. In particular, we consider estimation under the realizable contamination model of missing data in which an $\epsilon$ fraction of the observations are subject to an arbitrary (and unknown) missing not at random (MNAR) mechanism. When the true data is Gaussian, we provide evidence towards statistical-computational gaps in several problems. For mean estimation in $\ell_2$ norm, we show that in order to obtain error at most $\rho$, for any constant contamination $\epsilon \in (0, 1)$, (roughly) $n \gtrsim d e^{1/\rho^2}$ samples are necessary and that there is a computationally-inefficient
algorithm which achieves this error. On the other hand, we show that any computationally-efficient method within certain popular families of
algorithms requires a much larger sample complexity of (roughly) $n \gtrsim d^{1/\rho^2}$ and that there exists a
polynomial time
algorithm based on sum-of-squares which (nearly) achieves this lower bound. For covariance estimation in relative operator norm, we show that a parallel development holds. Finally, we turn to linear regression with missing observations and show that such a gap does not persist. Indeed, in this setting we show that minimizing a simple, strongly convex empirical risk nearly achieves the information-theoretic lower bound in
polynomial time.
Read the full abstract ▾
All authors Kabir Aladin Verchand, Ankit Pensia, Saminul Haque, Rohith Kuditipudi
Machine Learning
Learning Lineage-guided Geodesics with Finsler Geometry
Aaron Zweig, Mingxuan Zhang, David A. Knowles et al. (+1 more)
Trajectory inference investigates how to interpolate paths between observed timepoints of dynamical systems, such as temporally resolved population distributions, with the goal of inferring trajectories at unseen times and better understanding system dynamics. Previous work has focused on continuous
geometric priors, utilizing data-dependent spatial features to define a Riemannian metric. In many applications, there exists discrete, directed prior knowledge over admissible transitions (e.g. lineage trees in developmental biology). We introduce a Finsler metric that combines geometry with classification and incorporate both types of priors in trajectory inference, yielding improved performance on interpolation tasks in synthetic and real-world data.
Read the full abstract ▾
All authors Aaron Zweig, Mingxuan Zhang, David A. Knowles, Elham Azizi
Machine Learning
Cost Trade-offs in Matrix Inversion Updates for Streaming Outlier Detection
Florian Grivet and Louise Travé-Massuyès
Outlier detection identifies data points that deviate significantly from expected patterns, revealing anomalies that may require special
attention. Incorporating online learning further improves accuracy by continuously updating the model to reflect the most recent data. When employing the Christoffel function as an outlier score, online learning requires updating the inverse of a
matrix following a rank-k update, given the initial inverse. Surprisingly, there is no consensus on the optimal method for this task. This technical note aims to compare three different updating methods: Direct Inversion (DI), Iterative Sherman-Morrison (ISM), and Woodbury
Matrix Identity (WMI), to identify the most suitable approach for different scenarios. We first derive the theoretical computational costs of each method and then validate these findings through comprehensive Python simulations run on a CPU. These results allow us to propose a simple, quantitative, and easy-to-remember rule that can be stated qualitatively as follows: ISM is optimal for rank-1 updates, WMI excels for small updates relative to
matrix size, and DI is preferable otherwise. This technical note produces a general result for any problem involving a
matrix inversion update. In particular, it contributes to the ongoing development of efficient online outlier detection techniques.
Read the full abstract ▾
Machine Learning
Predictive Statistics Shape Emergent World Representations of Grid Walkers
Sasha Brenner, Thomas R. Knösche and Nico Scherf
Next-
token predictors often appear to develop internal representations of the latent world and its rules. The probabilistic nature of these models suggests a deep connection between the structure of the world and the geometry of probability distributions. In order to understand this link more precisely, we use a minimal
stochastic process as a controlled setting: constrained random walks on a two-dimensional
lattice that must reach a fixed endpoint after a predetermined number of steps. Optimal prediction of this process solely depends on a sufficient vector determined by the walker's position relative to the target and the remaining time horizon; in other words, the probability distributions are parametrized by the world's grid geometry. We train decoder-only
transformers and recurrent networks on prefixes sampled from the exact distribution of these walks and compare their hidden activations to sufficient statistics of prediction, by measuring
alignment and linear readability across layers. We find that the
transformer's computation factors into two stages: the first
attention block extracts the sufficient statistic from the input, and later layers transform it into the next-step predictive geometry. Across constraint variants the post-
attention representation is universal: a shared world-state of the
lattice that can be read directly as a world model, traced to the predictive geometry of the data. Later layers then specialize it to each variant's next-step distribution. Recurrent networks reach the same Bayes-optimal loss but do not isolate this world-state as a separate stage, showing that the world-model geometry also depends on architecture. Although demonstrated in a toy system, the results suggest that the geometry of the predictive distribution is a useful lens on how neural networks internalize the structure of their data.
Read the full abstract ▾
Machine Learning · Cross-listed from cs.RO
When Should a Robot Think? Resource-Aware Reasoning via Reinforcement Learning for Embodied Robotic Decision-Making
Jun Liu, Pu Zhao, Zhenglun Kong et al. (+12 more)
Embodied robotic systems increasingly rely on large
language model (LLM)-based
agents to support high-level
reasoning, planning, and decision-making during interactions with the environment. However, invoking LLM
reasoning introduces substantial computational latency and resource overhead, which can interrupt action execution and reduce system reliability. Excessive
reasoning may delay actions, while insufficient
reasoning often leads to incorrect decisions and task failures. This raises a fundamental question for
embodied agents: when should the
agent reason, and when should it act? In this work, we propose RARRL (Resource-Aware
Reasoning via Reinforcement Learning), a hierarchical framework for resource-aware orchestration of
embodied agents. Rather than learning low-level control policies, RARRL learns a high-level orchestration policy that operates at the
agent's decision-making layer. This policy enables the
agent to adaptively determine whether to invoke
reasoning, which
reasoning role to employ, and how much computational budget to allocate based on current observations, execution history, and remaining resources. Extensive experiments, including evaluations with empirical latency profiles derived from the ALFRED
benchmark, show that RARRL consistently improves task success rates while reducing execution latency and enhancing robustness compared with fixed or heuristic
reasoning strategies. These results demonstrate that adaptive
reasoning control is essential for building reliable and efficient
embodied robotic
agents.
Read the full abstract ▾
All authors Jun Liu, Pu Zhao, Zhenglun Kong, Xuan Shen, Peiyan Dong, Fan Yang, Lin Cui, Hao Tang, Geng Yuan, Wei Niu, Wenbin Zhang, Xue Lin, Gaowen Liu, Yanzhi Wang, Dong Huang
Machine Learning
Self-Aware Markov Models for Discrete Reasoning
Gregor Kornhardt, Jannis Chemseddine, Christian Wald et al. (+1 more)
Standard masked discrete
diffusion models face limitations in
reasoning tasks due to their inability to correct their own mistakes on the masking path. Since they rely on a fixed number of denoising steps, they are unable to adjust their computation to the complexity of a given problem. To address these limitations, we introduce a method based on learning a Markov transition kernel that is trained on its own outputs. This design enables
tokens to be remasked, allowing the model to correct its previous mistakes. Furthermore, we do not need a fixed time schedule but use a trained stopping criterion. This allows for adaptation of the number of function evaluations to the difficulty of the
reasoning problem. Our adaptation adds two lightweight prediction heads, enabling reuse and fine-tuning of existing pretrained models. On the Sudoku-Extreme
dataset we clearly outperform other flow based methods with a validity of 95%. For the Countdown-4 we only need in average of 10 steps to solve almost 96% of them correctly, while many problems can be solved already in 2 steps.
Read the full abstract ▾
All authors Gregor Kornhardt, Jannis Chemseddine, Christian Wald, Gabriele Steidl
Machine Learning · Cross-listed from cs.CL
Omanic: Towards Step-wise Evaluation of Multi-hop Reasoning in Large Language Models
Xiaojie Gu, Sherry T. Tong, Aosong Feng et al. (+8 more)
Evaluating the
reasoning abilities of large
language models (LLMs) solely from final answers can obscure failures in intermediate steps, especially in multi-hop QA
benchmarks without step-level annotations. To address this gap, we introduce Omanic, an open-domain 4-hop QA
benchmark designed not only to measure final-answer accuracy but also to diagnose where
reasoning breaks down. Omanic contains 10,296 machine-generated training examples (OmanicSynth) and 967 expert-reviewed human-annotated evaluation examples (OmanicBench), with each evaluation question decomposed into single-hop sub-questions, intermediate answers, and structured graph
topologies. Experiments with proprietary and open-source LLMs show that Omanic is challenging, while step-wise analysis reveals a later-hop bottleneck, factual knowledge floor, and error propagation along
reasoning chains. Fine-tuning on OmanicSynth transfers to six
reasoning and mathematics
benchmarks, yielding a 7.41-point average gain and validating its effectiveness as supervision for
reasoning-capability transfer. We release the data at https://huggingface.co/
datasets/li-lab/Omanic and the code at https://github.com/XiaojieGu/Omanic.
Read the full abstract ▾
All authors Xiaojie Gu, Sherry T. Tong, Aosong Feng, Sophia Simeng Han, Jinghui Lu, Yingjian Chen, Yusuke Iwasawa, Yutaka Matsuo, Chanjun Park, Rex Ying, Irene Li
Machine Learning
Simplex-to-Euclidean Bijection for Conjugate and Calibrated Multiclass Gaussian Process Classification
Bernardo Williams, Harsha Vardhan Tetali, Arto Klami et al. (+1 more)
We propose a conjugate and calibrated Gaussian process (GP) model for multi-class classification by exploiting the geometry of the probability simplex. Our approach uses Aitchison geometry to map simplex-valued class probabilities to an unconstrained Euclidean representation, turning classification into a GP regression problem with fewer latent dimensions than standard multi-class GP classifiers. This yields conjugate inference and reliable predictive probabilities without relying on distributional approximations in the model construction. The method is compatible with standard sparse GP regression techniques, enabling scalable inference on larger
datasets. Empirical results show well-calibrated and competitive performance across synthetic and real-world
datasets.
Read the full abstract ▾
All authors Bernardo Williams, Harsha Vardhan Tetali, Arto Klami, Marcelo Hartmann
Machine Learning
Trajectory-Optimized Time Reparameterization for Learning-Compatible Reduced-Order Modeling of Stiff Dynamical Systems
Joe Standridge, Daniel Livescu and Paul Cizmas
Stiff dynamical systems present a challenge for machine-learning reduced-order models (ML-ROMs), as explicit time integration becomes unstable in stiff regimes while implicit integration within learning loops is computationally expensive and often degrades training efficiency. Time reparameterization (TR) offers an alternative by transforming the independent variable so that rapid physical-time transients are spread over a stretched-time coordinate, enabling stable explicit integration on uniformly sampled grids. Although several TR strategies have been proposed, their effect on learnability in ML-ROMs remains incompletely understood. This work investigates time reparameterization as a stiffness-mitigation mechanism for neural ODE reduced-order modeling and introduces a trajectory-optimized TR (TOTR) formulation. The proposed approach casts time reparameterization as an optimization problem in arc-length coordinates, in which a traversal-speed profile is selected to penalize acceleration in stretched time. By targeting the smoothness of the training dynamics, this formulation produces reparameterized trajectories that are better conditioned and easier to learn than existing TR methods. TOTR is evaluated on three stiff problems: a parameterized stiff linear system, the van der Pol oscillator, and the HIRES chemical kinetics model. Across all cases, the proposed approach yields smoother reparameterizations and improved physical-time predictions under identical training regimens than other TR approaches. Quantitative results demonstrate loss reductions of one to two orders of magnitude compared to
benchmark algorithms. These results highlight that effective stiffness mitigation in ML-ROMs depends critically on the regularity and learnability of the time map itself, and that optimization-based TR provides a robust framework for explicit reduced-order modeling of multiscale dynamical systems.
Read the full abstract ▾
Machine Learning
When and Why Does Unsupervised RL Succeed in Mathematical Reasoning? A Manifold Envelopment Perspective
Zelin Zhang, Fei Cheng and Chenhui Chu
Although outcome-based reinforcement learning (RL) significantly advances the mathematical
reasoning capabilities of Large
Language Models (LLMs), its reliance on computationally expensive ground-truth annotations imposes a severe scalability bottleneck. Unsupervised RL guided by intrinsic rewards offers a scalable alternative, yet it suffers from opaque training dynamics and catastrophic instability, such as policy collapse and reward hacking. In this paper, we first design and evaluate a suite of intrinsic rewards that explicitly enforce concise and certain generation. Second, to discover the boundaries of this approach, we test base models across a spectrum of intrinsic
reasoning capabilities, revealing how a model's foundational logical prior dictates its success or failure. Finally, to demystify why certain configurations stabilize while others collapse, we introduce a novel
geometric diagnostic lens, showing that successful cases are enveloped by
manifolds. Ultimately, our work goes beyond merely demonstrating that enforcing concise and certain responses successfully boosts mathematical
reasoning; we reveal when this unsupervised approach breaks down and geometrically diagnose why.
Read the full abstract ▾
Machine Learning · Cross-listed from cs.CV
REFORGE: Multi-modal Attacks Reveal Vulnerable Concept Unlearning in Image Generation Models
Yong Zou, Haoran Li, Fanxiao Li et al. (+5 more)
Recent progress in image generation models (IGMs) enables high-fidelity content creation but also amplifies risks, including the reproduction of copyrighted content and the generation of offensive content. Image Generation Model Unlearning (IGMU) mitigates these risks by removing harmful concepts without full retraining. Despite growing
attention, the robustness under
adversarial inputs, particularly image-side threats in black-box settings, remains underexplored. To bridge this gap, we present REFORGE, a black-box red-teaming framework that evaluates IGMU robustness via
adversarial image
prompts. REFORGE initializes stroke-based images and optimizes perturbations with a cross-
attention-guided masking strategy that allocates noise to concept-relevant regions, balancing attack efficacy and visual fidelity. Extensive experiments across representative unlearning tasks and defenses demonstrate that REFORGE significantly improves attack success rate while achieving stronger semantic
alignment and higher efficiency than involved baselines. These results expose persistent vulnerabilities in current IGMU methods and highlight the need for robustness-aware unlearning against multi-modal
adversarial attacks. Our code is at: https://github.com/Imfatnoily/REFORGE.
Read the full abstract ▾
All authors Yong Zou, Haoran Li, Fanxiao Li, Shenyang Wei, Yunyun Dong, Li Tang, Wei Zhou, Renyang Liu
Machine Learning
Deep Tabular Representation Corrector
Hangting Ye, Peng Wang, Wei Fan et al. (+4 more)
Tabular data have been playing a mostly important role in diverse real-world fields, such as healthcare, engineering, finance, etc. The recent success of deep learning has fostered many deep networks (e.g.,
Transformer, ResNet) based tabular learning methods. Generally, existing deep tabular machine learning methods are along with the two paradigms, i.e., in-learning and pre-learning. In-learning methods need to train networks from scratch or impose extra constraints to regulate the representations which nonetheless train multiple tasks simultaneously and make learning more difficult, while pre-learning methods design several pretext tasks for pre-training and then conduct task-specific fine-tuning, which however need much extra training effort with prior knowledge. In this paper, we introduce a novel deep Tabular Representation Corrector, TRC, to enhance any trained deep tabular model's representations without altering its parameters in a model-agnostic manner. Specifically, targeting the representation shift and representation redundancy that hinder prediction, we propose two tasks, i.e., (i) Tabular Representation Re-estimation, that involves training a shift
estimator to calculate the inherent shift of tabular representations to subsequently mitigate it, thereby re-estimating the representations and (ii) Tabular Space Mapping, that transforms the above re-estimated representations into a light-
embedding vector space via a coordinate
estimator while preserves crucial predictive information to minimize redundancy. The two tasks jointly enhance the representations of deep tabular models without touching on the original models thus enjoying high efficiency. Finally, we conduct extensive experiments on state-of-the-art deep tabular machine learning models coupled with TRC on various tabular
benchmarks which have shown consistent superiority.
Read the full abstract ▾
All authors Hangting Ye, Peng Wang, Wei Fan, Xiaozhuang Song, He Zhao, Dandan Gun, Yi Chang
Machine Learning
Manifold-Matching Autoencoders: Embedding Data into Chosen Geometric Structures
Laurent Cheret, Vincent Létourneau, Isar Nejadgholi et al. (+3 more)
Manipulating the structure of data in the latent space of an autoencoder is a long-standing goal of representation learning. We propose a simple regularization,
Manifold Matching, that imposes an explicit
geometric structure on that latent space. The idea is to take a reference representation R of the data, which need not be the raw input X fed to the network, and to minimize, batch-wise, the mean squared error between the pairwise distances in R and those in the latent space Z. Although the autoencoder still learns to reconstruct X, the shape of its latent space comes to mimic that of R. The objective requires no labels, only reference distances, and because those are scalars, the dimensionality of R is decoupled from that of Z: any reference can guide any bottleneck. Using this single instrument we study how different
geometric structures can be imposed on a latent space, comparing against autoencoders built on
geometric or topological regularizers. Across fifteen
datasets under a shared protocol with statistical testing, the resulting
Manifold-Matching Autoencoder (MMAE) matches the global-structure quality of persistence-based autoencoders, statistically tying RTD-AE on distance correlation and triplet accuracy, and stays competitive on the local and topological metrics that other methods lead, at the computational cost of a plain autoencoder: it uses only a batched distance loss, with no persistence computation. We frame MMAE not as a universal winner but as an efficient and controllable regularizer whose reference selects which structure the latent inherits, and which turns any fixed
embedding into a reusable out-of-sample encoder. All experiments use a two-dimensional bottleneck.
Read the full abstract ▾
All authors Laurent Cheret, Vincent Létourneau, Isar Nejadgholi, Chris Drummond, Hussein Al Osman, Maia Fraser
Machine Learning · Cross-listed from cs.CV
Bridging the Simulation-to-Reality Gap in Electron Microscope Calibration via VAE-EM Estimation
Jilles S. van Hulst, W.P.M.H. Heemels and Duarte J. Antunes
Electron microscopy has enabled many scientific breakthroughs across multiple fields. A key challenge is the tuning of microscope parameters based on images to overcome optical aberrations that deteriorate image quality. This calibration problem is challenging due to the high-dimensional and noisy nature of the diagnostic images, and the fact that optimal parameters cannot be identified from a single image. We tackle the calibration problem for Scanning Transmission Electron Microscopes (STEM) by employing variational autoencoders (VAEs), trained on simulated data, to learn low-dimensional representations of images, whereas most existing methods extract only scalar values. We then simultaneously estimate the model that maps calibration parameters to encoded representations and the optimal calibration parameters using an expectation maximization (EM) approach. This joint estimation explicitly addresses the simulation-to-reality gap inherent in data-driven methods that train on simulated data from a digital twin. We leverage the known
symmetry property of the optical system to establish global identifiability of the joint estimation problem, ensuring that a unique optimum exists. We demonstrate that our approach is substantially faster and more consistent than existing methods on a real STEM, achieving a 2x reduction in estimation error while requiring fewer observations. This represents a notable advance in automated STEM calibration and demonstrates the potential of VAEs for information compression in images. Beyond microscopy, the VAE-EM framework applies to inverse problems where simulated training data introduces a reality gap and where non-injective mappings would otherwise prevent unique solutions.
Read the full abstract ▾
Machine Learning
SympFormer: Accelerated attention blocks via Inertial Dynamics on Density Manifolds
Viktor Stein, Wuchen Li and Gabriele Steidl
Transformers owe much of their empirical success in natural
language processing to the self-
attention blocks. Recent perspectives interpret
attention blocks as interacting particle systems, whose mean-field limits correspond to gradient flows of interaction energy functionals on probability density spaces equipped with Wasserstein-$2$-type metrics. We extend this viewpoint by introducing accelerated
attention blocks derived from inertial Nesterov-type dynamics on density spaces. In our proposed architecture,
tokens carry both spatial (feature) and velocity variables. The time discretization and the approximation of accelerated density dynamics yield
Hamiltonian momentum
attention blocks, which constitute the proposed accelerated
attention architectures. In particular, for linear self-
attention, we show that the
attention blocks approximate a Stein variational gradient flow, using a bilinear kernel, of a potential energy. In this setting, we prove that elliptically contoured probability distributions are preserved by the accelerated
attention blocks. We present implementable particle-based
algorithms and demonstrate that the proposed accelerated
attention blocks converge faster than the classical
attention blocks while preserving the number of oracle calls.
Read the full abstract ▾