Machine Learning · Cross-listed from stat.ME
Correcting Variable Importance Scored by Random Forests
Guancheng Zhou, Haiping Xu, Jason Liu et al. (+1 more)
Variable importance produced by Random Forests (RF) is used widely in statistical data analysis, and has played an important role in a variety of tasks such as assisting model interpretation, model selection and diagnosis, and cost-bounded learning etc. However, the calculation of variable importance in RF does not take into account of the correlations among variables, and variables that are correlated to many other variables tend to receive a lower importance index or being completely masked (i.e., with an importance index near zero) by other strongly correlated variables. To prevent influence from unwanted correlated variables in calculating variable importance, we propose to group variables by their conditional correlations (conditional on the response variable). We explore two computationally efficient options, with one grouping variables individually, and then separates the variable of interest from all correlated variables, while the other uses clustering to group variables according to their pair-wise conditional correlations. Our experiments show that both lead to sensible corrections to the importance of variables.
Read the full abstract ▾
All authors Guancheng Zhou, Haiping Xu, Jason Liu, Donghui Yan
Machine Learning
N-GRPO: Embedding-Level Neighbor Mixing for Enhanced Policy Optimization
Xukun Zhu, Hang Yu, Peng Di et al. (+1 more)
The success of Large
Language Models in mathematical
reasoning relies heavily on the generation of diverse and valid solution paths during the rollout phase. However, current rollout techniques face a fundamental trade-off:
token-level sampling often yields redundant trajectories that differ only in rephrasing, while
embedding-level methods utilizing random noise frequently disrupt semantic consistency. To resolve this, we introduce N-GRPO, a novel exploration strategy integrated into the Group Relative Policy Optimization (GRPO) framework. Rather than relying on
token-level sampling or native
embedding-level noise, our approach leverages Semantic Neighbor Mixing. This mechanism dynamically constructs input representations by mixing the
embeddings of an anchor
token and its nearest semantic neighbors, thereby injecting diversity while strictly adhering to the local semantic
manifold. Experimental evaluations on the DeepSeek-R1-Distill-Qwen models across different sizes show that N-GRPO not only achieves consistent improvements over strong baselines on math
reasoning benchmarks but also exhibits robust generalization capabilities on out-of-distribution tasks.
Read the full abstract ▾
All authors Xukun Zhu, Hang Yu, Peng Di, Linchao Zhu
Machine Learning · Cross-listed from cs.CR
Yv Zhang, Hao Sun, Hao Fang et al. (+5 more)
External
memory has become a core component of modern web
agents, enabling long-horizon
reasoning through the
retrieval of past experiences. However, this paradigm introduces a critical vulnerability: malicious content injected into
memory can be persistently recalled and repeatedly influence
agent behavior. In this work, we identify and systematically study
multimodal memory poisoning, an overlooked yet practical attack surface in web-
agent systems. We propose MemVenom, a unified black-box attack framework that poisons graph-structured external
memory with coordinated text-image evidence. Our method consists of a two-stage design: (1) a trigger-conditioned
retrieval attack that ensures high-probability recall of malicious
memory, and (2) a post-
retrieval attack induction that leverages
adversarial perturbations and stealthy OCR injection to override the original user objective. Unlike prior attacks that operate on
prompts or text-only
memory, our approach enables persistent, reusable, and goal-agnostic attacks without modifying model parameters or re-optimizing malicious tasks. Experiments across multiple web-
agent frameworks and
vision-
language models demonstrate that MemVenom achieves strong end-to-end attack success with minimal impact on benign performance, reaching up to 99.15% on GPT-5-family web
agents, while transferring effectively across architectures and model scales.
Read the full abstract ▾
All authors Yv Zhang, Hao Sun, Hao Fang, Kuofeng Gao, Fan Mo, Bin Chen, Shu-Tao Xia, Yaowei Wang
Machine Learning · Cross-listed from cs.AI
When the Chain of Thought Knows Better: Failure Modes in Multi-Turn Reasoning Models
Sai Kartheek Reddy Kasu, Nils Lukas and Samuele Poppi
Failures in multi-turn
reasoning models are largely invisible to terminal-score evaluation. A model can lock onto an unsafe stance early in a long dialogue, yet its final-turn refusal rate may appear indistinguishable from a robustly aligned baseline. To expose these hidden temporal dynamics, we propose a trace-level diagnostic - the CoT-Output 2x2 safety
matrix. This framework labels every turn along two independent axes (internal
reasoning and visible output), yielding four operationally defined failure
cells: robust
alignment,
alignment faking, overt jailbreak, and a distinct failure mode we term context-injection failure (where the CoT maintains safe
reasoning, but the visible output produces harm, highlighting a multi-turn manifestation of
reasoning unfaithfulness). We evaluate three distilled
reasoning targets against a fixed attacker across five oversight conditions, collecting 6750 turn-level observations on the Information-Hazard scenario. Our analysis reveals two reproducible vulnerabilities: an oversight paradox where explicit monitoring cues paradoxically increase
alignment-faking rates rather than suppress them, and a context-injection failure where models lock onto unsafe external outputs despite safe internal states. We release the full
dataset of multi-turn dialogues and CoT traces to support follow-up trace-diagnostic research.
Read the full abstract ▾
Machine Learning
SPACR: Single-Pass Adaptive Training of Uncertainty-Aware Conformal Regressors
Soundouss Messoudi, Sylvain Rousseau and Sébastien Destercke
Conformal Prediction (CP) provides robust uncertainty guarantees for predictive models, but is typically applied post hoc, which misaligns model training with the conformal goal of producing efficient (i.e., narrow) intervals. We propose SPACR (Single-Pass Adaptive Conformal Regressor), a novel method for directly training uncertainty-aware regressors within a differentiable loss. SPACR jointly optimizes accuracy, efficiency, and validity without batch-splitting or a predefined confidence level during training. As a result, a single SPACR model yields valid prediction intervals at multiple confidence levels during inference, avoiding the costly retraining required by methods like Directly Optimized Inductive Conformal Regression (DOICR). Experiments on diverse tabular and image
datasets show that SPACR consistently gives tighter intervals and better coverage-efficiency trade-offs compared to standard CP and DOICR, while significantly reducing computational costs relative to retraining-dependent baselines.
Read the full abstract ▾
Machine Learning
Pre-AF 13: An Interpretable Atrial Fibrillation Risk Score Mined from Discharge Reports
Olga Shakhmatova, Dmitrii Kriukov, Daniil Larionov et al. (+10 more)
Background. Atrial fibrillation (AF) is the most prevalent cardiac arrhythmia and a major determinant of prognosis. Established AF risk scores rely on factors (older age, hypertension) nearly ubiquitous among patients with cardiovascular disease (CVD), offering limited stratification in this high-risk group. Most target long-term (5-10 year) rather than medium-term prediction. We developed interpretable ML models predicting AF risk over a 24-month and entire follow-up horizon in CVD patients using routinely collected hospital data. Methods. Single-center retrospective study of electronic health records from the National Research Cardiology Center (Russia) for patients aged >=18 with CVD but without pre-existing AF, hospitalized more than once between January 2012 and May 2019. A custom NLP pipeline transformed unstructured discharge reports into 73 structured features, combining a rule-based parser with
transformer-based NER. Using LightAutoML we built a full model (73 features), a simple model (reduced subset), and a linear model for a bedside risk score. Performance was assessed by ROC AUC, compared with CHARGE-AF, C2HEST, MHS, and HAVOC, and interpreted via SHAP. Results. Of 80,576 records from 45,000 patients, 17,562 met inclusion criteria; 1,438 (8.19%) developed AF. The full model reached ROC AUC 0.735 (24-month) and 0.696 (entire follow-up); the simple model was nearly identical (0.725, 0.696). All non-linear models outperformed the four clinical risk scores (ROC AUC 0.53-0.64). The simple model uses 13 features and is named Pre-AF 13. SHAP identified age and left atrial volume as dominant predictors. A linear risk score (Pre-AF 9) stratified observed 24-month AF incidence from ~7% to 36%. Conclusion. Interpretable ML models built from routinely collected EHR data identify high-AF-risk CVD patients, outperforming established clinical risk scores.
Read the full abstract ▾
All authors Olga Shakhmatova, Dmitrii Kriukov, Daniil Larionov, Nikita Khromov, Iaroslav Bespalov, Alexander Zolotarev, Kirill Grishchenkov, Ekaterina Ivanova, Miron Kuznetsov, Ilya Sochenkov, Elizaveta Panchenko, Artem Shelmanov, Dmitry V. Dylov
Machine Learning
Transformer Based Model for Spatiotemporal Feature Learning in EEG Emotion Recognition
Xinglong Cui and Dian Gu
Electroencephalography (
EEG) is a widely adopted technique for monitoring
brain activity, offering valuable insights into neurological states due to its high temporal resolution and cost-effectiveness. To enhance the analysis of complex
EEG data, we propose
EEG-TransNet, an architecture designed to capture temporal, regional, and synchronous features of
EEG signals.
EEG-TransNet introduces three key modules: 1) a preprocessing and feature extraction module leveraging ResNet and wavelet-based denoising, 2) a Local Self-
Attention Block for regional feature learning, and 3) a Fuzzy-
Attention Synchronous
Transformer (FAST) to model spatiotemporal dependencies. Through extensive experiments on three
EEG datasets (BETA, SEED, and DepEEG), the proposed model consistently outperforms other methods in terms of classification accuracy and robustness across varying signal lengths. Ablation studies confirm the contribution of the Local Self-
Attention Block in improving performance, and the inclusion of depthwise separable convolutions in the decoder reduces computational complexity while maintaining high accuracy.
EEG-TransNet's ability to generalize across subjects with minimal performance variation highlights its potential as a robust tool for
EEG-based
brain activity classification and
emotion recognition tasks.
Read the full abstract ▾
Machine Learning · Cross-listed from eess.IV
++nnU-Net: Scaling nnU-Net with Prefix-Based Data Augmentation
Ana Sofia Santos, André Ferreira, Gijs Luijten et al. (+6 more)
The nnU-Net has demonstrated continuous success in medical
segmentation tasks, which heavily rely on the availability and diversity of annotated biomedical data. However, assembling medical imaging cohorts remains challenging due to numerous factors such as
privacy regulations and annotation costs. As a result, data augmentation plays a crucial role in increasing data availability while maintaining anatomical feasibility. Hence, we propose the ++nnU-Net, a novel data augmentation module based on image registration that operates prior to preprocessing and training take place. Our framework was evaluated across five different 2D
datasets. In this workflow, image data go through a two-stage registration process, generating new warped images. The transformations are then applied to the respective
segmentation. In addition, the pipeline computes available disk space, generates supplementary binary synthetic masks and generates checkpoints. We demonstrate that the ++nnU-Net outperforms the nnU-Net baseline, yielding improvements in Dice Similarity Coefficient scores. In the most prominent cases, we observe performance gains of approximately 22%. These findings highlight the effectiveness of registration-based data augmentation, particularly for 2D medical imaging
datasets and suggest that the ++nnU-Net provides a practical and scalable approach for enhancing
segmentation performance in data-limited settings. The source code for the ++nnU-Net is available at: https://github.com/sofia-adelie/plusplusnnunet.git
Read the full abstract ▾
All authors Ana Sofia Santos, André Ferreira, Gijs Luijten, Naida Solak, Lisle Faray de Paiva, Behrus Hinrichs-Puladi, Jens Kleesiek, Jan Egger, Victor Alves
Machine Learning
Unifying Data, Memory, and Compute Efficiency in LLM training: A Survey
Vanessa Schmidt, Huy Hoang Nguyen, Cédric Jung et al. (+2 more)
Resource constraints increasingly determine what can be trained, fine-tuned, and deployed in large
language models (LLMs), yet efficiency is often studied through isolated techniques rather than as an interacting system of limits. This survey adopts a constraint-centric perspective and organizes recent progress around three coupled bottlenecks: data efficiency (what to train on),
memory efficiency (how to fit training), and compute budget awareness (when and where to spend FLOPs). On the data axis, we review selection and
pruning methods that maximize learning per
token, ranging from scalable proxy signals based on learning dynamics to gradient- and influence-based scoring, as well as difficulty-aware and curriculum-style strategies. We highlight emerging evidence that different notions of good data dominate in different regimes, implying that optimal subsets depend on the task objective and resource budget rather than being universal. On the systems side, we show that GPU
memory, not raw compute, is often the dominant bottleneck in fine-tuning, and that effective scaling requires jointly reducing weight storage, optimizer states, and activation
memory rather than optimizing any single component in isolation. Beyond
memory, we frame training and inference as compute-governed processes in which optimization, data selection, and decoding must explicitly account for finite FLOP budgets. We review evidence for compute-optimal allocation and stopping rules, where computation should be halted or reallocated once marginal performance gains fall below a budget-dependent threshold. Together, these results unify compute-aware data selection, scaling laws, and adaptive inference under a common principle of resource-conditioned decision-making.
Read the full abstract ▾
All authors Vanessa Schmidt, Huy Hoang Nguyen, Cédric Jung, Shirin Salehi, Anke Schmeink
Machine Learning
Event-Driven Reinforcement Learning Enables Long-Horizon Control in Semiconductor Fabrication
Yavar Yeganeh, Mahsa Shekari, Nicla Frigerio et al. (+2 more)
Reinforcement learning promises to optimize sequential decisions in large-scale systems.
Semiconductor manufacturing systems are
stochastic and highly constrained environments where heterogeneous wafers traverse hundreds of processing steps across extensive equipment networks. These characteristics yield complex, high-dimensional decision problems with delayed feedback and long-horizon requirements, complicating production planning and control. We propose a deep reinforcement learning framework for multi-objective policy optimization at this scale. Specifically, we formulate control as a centralized-
agent problem, where a core policy coordinates system-wide decisions, while system
evolution is represented as an interconnected temporal process driven by discrete events. Accordingly, we develop a tailored event-driven temporal-difference formulation that remains general and can be integrated with various policy optimization methods under relevant training settings. We investigate several core model-free
algorithms incorporated into this framework and evaluate their effectiveness using high-fidelity simulations of diverse, industry-real operating scenarios. Across extensive validation experiments,
agents trained in both offline and online settings show significant and consistent gains in throughput and utilization. We further evaluate performance and generalization across training phases, clarifying the relative strengths of alternative reinforcement learning formulations and
algorithms. Overall, the results support the scalability, generality, and transferability of the proposed framework for controlling event-driven complex adaptive systems.
Read the full abstract ▾
All authors Yavar Yeganeh, Mahsa Shekari, Nicla Frigerio, Daniele Pagano, Andrea Matta
Machine Learning
From Observation to Intervention: A Causal Audit of Expert Importance in Mixture-of-Experts Models
Leonard Engmann, Christian Medeiros Adriano and Holger Giese
Interpretability methods routinely use population-level summary statistics over observed model behaviour to license claims about the effects of targeted interventions on specific computations; in Pearl's terms, they treat rung-1 associational evidence as if it supported rung-2 interventional conclusions, a move whose validity is rarely tested. We examine one concrete instance: the use of routing statistics in Mixture-of-Experts (MoE)
pruning, where utilization rates, activation norms, and routing weight distributions are treated as predictors of which experts can be removed without functional cost. A
token-level interventional audit across three high-redundancy MoE architectures (OLMoE-1B-7B-0924, Qwen1.5-MoE-A2.7B, DeepSeek-V2-Lite) finds no observational metric predicts causal expert importance in any model: across all 60 metric-layer combinations effect sizes stay below Cohen's $d = 0.23$, and no metric is reliably positive under our corrected, dual-test criterion. A per-
token routing weight control, run with identical $n$, rules out insufficient power, recovering a signal whose CI excludes zero at OLMoE's final MoE layer ($d = +0.231$, 95% CI $[+0.09, +0.37]$, $p = 0.0013$). Existing
pruning methods succeed in this regime not by identifying dispensable experts but because early-layer redundancy renders most selection criteria interchangeable. Our results provide an explicit counterexample to the common inferential step from population-level observational summaries to
token-level interventional claims about expert importance, and illustrate how interventional audits can calibrate the evidential standards for interpretability claims.
Read the full abstract ▾
Machine Learning · Cross-listed from hep-ph
Efficient AI-Inspired Reduction of Feynman Integrals via Tube Seeding
Justin Berman, Francois Charton, Andres Luna et al. (+2 more)
In this paper, we use machine learning to discover a new seeding strategy for integration-by-parts reduction of Feynman integrals, which is a frequent bottleneck in state-of-the-art calculations in theoretical particle and gravitational-wave physics. Our strategy allows us to reduce multi-loop integrals with large numerator powers via essentially the standard Laporta
algorithm but with a sparse selection of seed integrals that grows only linearly with the numerator power, whereas existing strategies lead to growth with a
polynomial power that increases with the complexity of the integral being reduced. The seeds are restricted to a thin tube-like region that connects the target integral to the master integrals along a zigzag path. We demonstrate the power of our approach by reducing non-planar 2-loop 5-point integrals of rank 20 with numerical kinematics over a finite field, which is prohibitively difficult for the Laporta
algorithm with conventional seeding. Going beyond individual integrals, we further demonstrate the reduction of a complete set of top-level rank-10 integrals by dividing the target integrals into several chunks, each of which can be solved by our sparse seeding strategy with considerably less time and a significantly lower
memory footprint than other state-of-the-art strategies, making the approach well-suited for phenomenological applications. We provide a proof-of-principle implementation on GitHub at https://github.com/andreslunagodoy/tube_seeding.
Read the full abstract ▾
All authors Justin Berman, Francois Charton, Andres Luna, Matthias Wilhelm, Mao Zeng
Machine Learning · Cross-listed from cs.CR
Do LLMs Make Neural Distinguishers Wise?
Tatsuya Sakagami, Masashi Hisai and Naoto Yanai
Neural distinguishers are a cryptanalysis method for symmetric-key cryptography that trains machine learning models on pairs of plaintexts and ciphertexts with specific differences in order to recover a secret key. To the best of our knowledge, no existing work has explored the use of large
language models (LLMs) for neural distinguishers. In this paper, we propose LLM-based neural distinguishers through a
prompt design and conduct extensive experiments with them on SPECK-32/64 to investigate whether LLMs can strengthen neural distinguishers. We then found three key insights. First, by comparing the results of LLM-based neural distinguishers with ResNet in the existing work, we demonstrate that LLMs provide no observable improvement in the performance of neural distinguishers. Second, we confirm that, at high rounds, the choice of differences is no longer effective for LLM-based neural distinguishers as well as ResNet. Third, we show that the performance of LLM-based neural distinguishers can be significantly improved by incorporating only the XOR operation results as a
prompt design.
Read the full abstract ▾
Machine Learning · Cross-listed from physics.comp-ph
An adaptive framework for the axisymmetric pulsar magnetosphere using physics-informed Kolmogorov-Arnold networks
Spyros Rigas, Ioannis Contopoulos, Georgios Alexandridis et al. (+1 more)
The
pulsar magnetosphere has only recently been addressed using Physics-Informed Neural Networks (PINNs), by deploying a domain-decomposition approach and treating the separatrix and equatorial current sheet as infinitesimally thin discontinuities. However, this baseline requires extensive manual hyperparameter tuning, achieves limited final accuracy and demands several hours of training. We refine this framework by introducing domain-specific neural architectures based on Kolmogorov-Arnold networks, an automated adaptive training pipeline and a physics-based convergence criterion that eliminate the need for manual calibration. The proposed methodology delivers self-consistent axisymmetric magnetosphere solutions with mean squared errors of the PDE residuals at O(1e-6) in double precision - an improvement of two orders of magnitude over the baseline - while achieving convergence in under 20 minutes in single precision. Importantly, the method reliably resolves
stellar radii reduced by up to 80% compared to the baseline, overcoming the severe spatial scale disparities that also challenge traditional solvers. Furthermore, by varying the flux that opens to infinity, we provide a correction to the equation that connects it to the equatorial T-point's position. The complete framework is released as the open-source library PulsarX.
Read the full abstract ▾
All authors Spyros Rigas, Ioannis Contopoulos, Georgios Alexandridis, Antonios Nathanail
Machine Learning
Cross-Agent Learning Signals Enable Coordinated Role-Decomposed LLM Training
Jaewan Park, Solbee Cho and Jay-Yoon Lee
Agentic search systems must coordinate evidence acquisition and response generation, yet existing approaches either couple both roles under a single
agent objective or decompose them without disentangling their respective contributions to the final outcome. We introduce DAC (Divide and Cooperate), a role-decomposed training framework that, given task-specific external verification signals, trains a searcher and a generator with role-specific cross-verification rewards. DAC allows the generator to abstain when the retrieved evidence appears insufficient, and uses this decision together with externally evaluated search sufficiency to assign appropriate credit to each role. To prevent degenerate over-abstention, we further introduce hard-positive evidence augmentation, which discourages abstaining on sufficient but hard evidence. Across seven general and multi-hop QA
benchmarks and two model backbones, DAC consistently outperforms strong single-
agent and multi-
agent baselines. Controlled evaluations show that these gains arise from coordinated improvements in both search and generation. Our results highlight the importance of explicitly and properly assigning credit across interacting roles when training agentic search systems.
Read the full abstract ▾
Machine Learning
PL-KKT-hPINN: Enforcing Nonlinear Equality Constraints on Neural Networks via Piecewise-Linear Projection
Fateme Mohammad Mohammadi, Hector Budman and Joshua L. Pulsipher
While physics-informed neural networks (PINNs) have shown strong potential for process modeling, physical equations are only enforced as soft constraints during training, and thus, they do not guarantee constraint satisfaction at inference. We propose a framework, called piecewise-linear Karush–Kuhn–Tucker hard-constrained PINNs (PL-KKT-hPINNs), that strictly enforces nonlinear equality constraints through piecewise-linear projection. This extends the KKT-hPINN framewor, which exactly enforces linear equalities through the Karush–Kuhn–Tucker (KKT) conditions associated with orthogonally projecting neural network outputs onto the constraint feasible region. The method is demonstrated on a continuous stirred-tank reactor (CSTR) case study for both one and two inputs. Results show that PL-KKT-hPINN preserves predictive accuracy comparable to that of a standard neural network while achieving substantially lower constraint violations. In addition, the proposed model shows improved robustness in low-data regimes, yielding lower RMSE than the unconstrained neural network for limited training sample sizes. These results demonstrate that PL-KKT-hPINN provides a computationally efficient and physically consistent framework for surrogate modeling of nonlinear chemical engineering systems.
Read the full abstract ▾
Machine Learning
One Step Closer to Ground Truth: A Multi-Scale Residual-Aware Representation Learning Pipeline for Predicting Time Series Data
Amrijit Biswas, Mustafa Kamal, Robin Krambroeckers et al. (+4 more)
Transformer-based models have emerged as leading paradigms in time-series
forecasting in recent years, employing self-
attention mechanisms to capture long-range dependencies. Despite their success, these single-stage
forecasting architectures exhibit persistent systematic residual biases arising from structural discrepancies, unmodeled
stochastic components, or inadequate multi-scale temporal representations. This limitation persists when residuals are treated as irreducible noise, precluding adaptive correction of structured error patterns. To address this limitation, we introduce a two-stage, model-agnostic framework that explicitly decouples
forecasting and residual learning into distinct stages of representation learning. A base
transformer first generates the initial predictions. Subsequently, a dedicated meta-corrector dynamically models structured error patterns across multivariate channels, preserves cross-variable dependencies, and iteratively refines the residual bias of the base
transformer. By formalizing this pipeline as a hypothesis space expansion, our framework addresses approximation limitations inherent in single-stage architectures, removes reliance on restrictive assumptions, and enables end-to-end learning of complex error dynamics. Evaluated on eight popular
benchmark datasets using established protocols, our approach achieves state-of-the-art performance, with significant improvements in standard metrics (MSE, MAE). The results demonstrate the framework's ability to mitigate systematic biases and enhance robustness to complex temporal dynamics, advancing the practical applicability of
transformer-based
forecasting models.
Read the full abstract ▾
All authors Amrijit Biswas, Mustafa Kamal, Robin Krambroeckers, M. M. Lutfe Elahi, Sifat Momen, Nabeel Mohammed, Shafin Rahman
Machine Learning · Cross-listed from stat.OT
ClusBench: The Clustering Benchmark Data Resource You've All Been Waiting For (?)
David P. Hofmeyr
Although some very common test beds exist for assessing the performance of clustering methods, large scale benchmarking is typically limited to relatively simplistic simulation set-ups. Here we describe the production and curation of close to 3000 synthetic data sets, derived from more than 200 publicly available data sets; the majority of which arose from real-world applications. By fitting a flexible non-parametric distribution to each base data set we are able to retain much of the nuance in real-world data which is difficult to reproduce in standard simulations, while also producing data sets whose sizes are sometimes substantially greater than the data sets from which they are derived. The synthetic data sets, plus an accompanying R package, are available for download from https://github.com/DavidHofmeyr/ClusBench.
Read the full abstract ▾
Machine Learning
In Defense of Information Leakage in Concept-based Models
Mateo Espinosa Zarlenga
Concept-based models (CMs), deep neural networks that ground their predictions on representations aligned with human-understandable concepts (e.g., "round", "stripes", etc.), have been shown to learn representations that leak concept-irrelevant information. As the traditional narrative goes, this leakage is undesirable and should be eradicated as it leads to uninterpretable models. In this paper, we posit that this conventional view of leakage in CMs is not only ill-posed, as the evidence of how leakage makes a model less interpretable is often inconclusive, but also bound to lead to impractical CMs under common real-world constraints. Specifically, we argue that in real-world settings where concept incompleteness is the norm, some leakage is often necessary for constructing accurate and intervenable CMs. To this end, we propose that there is such a thing as benign leakage and show that, by optimizing a reframing of the typical CM training objective, CMs can encourage and exploit this form of leakage without sacrificing accuracy or intervenability.
Read the full abstract ▾
Machine Learning
How Does Reasoning Flow? Tracing Attention-Induced Information Flow for Targeted RL in LLMs
Zhichen Dong, Yang Li, Yuhan Sun et al. (+9 more)
Token-level credit assignment remains a key obstacle for reinforcement learning (RL) in large
language models (LLMs), where RL recipes typically treat all
tokens equally, failing to distinguish decisive
reasoning steps from routine formatting or fluent filler. Recent attempts leverage model-internal signals to assign finer-grained credit, but these are often point-wise heuristics that ignore the global structure of information propagation. We propose FlowTracer, an RL framework that traces answer-targeted
reasoning flow on an
attention-induced directed acyclic graph in which nodes correspond to
tokens and edge capacities come from aggregated
attention weights and derives
token credit from this global structure. The edge capacities are reweighted to retain only the influence that can reach the answer region, while enforcing local flow conservation so intermediate
tokens neither lose nor gain effective mass due to path length or irrelevant branches. On this graph, FlowTracer extracts an information-flow backbone connecting the question to the answer and scores
tokens by flow throughput, revealing high-impact hubs and aggregation checkpoints that mediate long-range dependencies. These derived importances are used to shape
token-level rewards, enabling learning signals to focus precisely on the
tokens that route information toward (or away from) correct answers and delivering consistent performance gains across a range of
reasoning tasks.
Read the full abstract ▾
All authors Zhichen Dong, Yang Li, Yuhan Sun, Weixun Wang, Yijia Luo, Zinian Peng, Taiheng Ye, Chao Yang, Wenbo Su, Yu Cheng, Bo Zheng, Junchi Yan
Machine Learning
PhysMetrics.Weather: An Evaluation Framework for Physical Consistency in ML Weather Models
Emma Kasteleyn, Timo Maier, Axel Lauer et al. (+3 more)
Machine learning weather prediction (MLWP) models have achieved impressive
forecasting performance at a small fraction of the computational costs required for traditional physics-based methods. However, they are primarily (1) data-driven and (2) evaluated using pixel-wide error metrics (e.g., RMSE), so there are no guarantees that their forecasts are consistent with known physical laws. We introduce PhysMetrics$.$Weather, an evaluation framework that assesses the physical realism of MLWP models across three types of metrics: conservation, spectral, and dynamical. By quantifying physical realism, this tool guides the development of physics-informed architectures and helps evaluate whether MLWP models are reliable for operational use. Our framework is available on Github at https://github.com/Emmakast/PhysMetrics.Weather.
Read the full abstract ▾
All authors Emma Kasteleyn, Timo Maier, Axel Lauer, Veronika Eyring, Pierre Gentine, Ana Lucic
Machine Learning
Is Fairness Truly Fair? Towards Reliable Lipschitz Fairness in Multi-Task Learning via Fixed-\texorpdfstring{$\delta$}{delta} Alignment
Junbo Ding, Xin Zang, Chenchen Pan et al. (+3 more)
Lipschitz-style individual
fairness formalizes the idea that semantically similar examples should receive similar predictions, but its evaluation in multi-task learning (MTL) can be confounded by method-induced representation scales. This paper identifies threshold confounding: when the
auditing tolerance is derived from each model's own representation distances, different
algorithms are compared under different semantic thresholds. A threshold-drift analysis further shows how Bias rankings can change and identifies sufficient conditions for ranking preservation. We propose ReLiF, a reliability-aware framework that separates evaluation-time fixed-$\delta$
auditing from training-time controlled regularization. ReLiF uses a shared reference tolerance for comparable
auditing and a violation-rate feedback controller to keep the Lipschitz surrogate active without letting it dominate
stochastic training. This work also develops supporting analysis for threshold drift, reference-tolerance selection, and the relationship between the huberized training surrogate and its unsmoothed positive-margin counterpart. Experiments on clinical time-series
benchmarks and NYUv2 (NYU Depth V2) dense prediction show that fixed-$\delta$
auditing exposes utility–
fairness trade-offs that method-dependent thresholds can obscure. On NYUv2 with a ResNet50 backbone, ReLiF achieves competitive utility while substantially reducing aligned bias under shared fixed thresholds. On clinical
benchmarks, ReLiF yields controlled
fairness-regularized trade-offs, while fixed-$\delta$
auditing reveals that task-balancing baselines can sometimes achieve lower bias and that genuine utility–
fairness trade-offs persist. These results support fixed-$\delta$
auditing as a semantically consistent protocol for evaluating Lipschitz
fairness in MTL.
Read the full abstract ▾
All authors Junbo Ding, Xin Zang, Chenchen Pan, Donghao Song, Jiaxin Zhu, Danhuai Guo
Machine Learning · Cross-listed from cs.HC
Profy: Interpretable Visualization of Expertise-Dependent Motor Skills Toward Supporting Piano Practice
Kazuki Kawamura, Fujiki Nakamura, Hayato Nishioka et al. (+3 more)
The quality of piano performance depends on nuanced timing, articulation, and dynamic control, but practice feedback is often summary-based and hard to act on. We introduce Profy, a weakly supervised system that learns from take-level labels derived from aggregated listener ratings (expert-labeled vs. amateur-labeled) to produce time-aligned highlights for review during piano practice. We collected synchronized 1 kHz key-motion and audio from 73 pianists and used 1,083 valid takes for modeling and evaluation. The model outputs clip-level predictions together with evidence scores on a shared resampled model time base for visualization. On 20 amateur clips from short technique studies annotated by 21 expert pianists, the displayed highlight score aligns with passages that expert pianists marked for review despite training without localized labels (Pearson r=0.61, ROC-AUC 0.75). Rather than summarizing a take with a single global score, Profy helps learners decide where to inspect next by supporting scrubbing, looping, and focused replay of time-localized passages associated with expert-amateur differences.
Read the full abstract ▾
All authors Kazuki Kawamura, Fujiki Nakamura, Hayato Nishioka, Momoko Shioki, Shinichi Furuya, Jun Rekimoto
Machine Learning · Cross-listed from cs.RO
Dexterous Point Policy: Learning Point-based Dexterous Hand Policies from Human Demonstrations
Beomjun Kim, Seong Hyeon Park, Seunghoon Sim et al. (+3 more)
Robotic foundation models pre-trained on human demonstration
videos have shown promise, but a significant embodiment gap remains when the resulting policies are deployed on real
robots. A common remedy is to fine-tune these models on
robot-specific demonstrations. However,
robot data collection can be prohibitively expensive and time-consuming, which is particularly acute in dexterous manipulation, e.g., teleoperating a multi-fingered hand for even a single atomic task can take days. To address this, we introduce Dexterous Point Policy, a framework that learns dexterous manipulation policies directly from human
videos and requires no
robot demonstrations. Our core insight is that a unified 3D keypoint representation can bridge human and
robot embodiments when used for both observations and actions. Specifically, we extract 3D keypoints of task-relevant objects and human hands from raw
videos, and train an autoregressive
transformer over these keypoints. We observe that at the keypoint level, specifically the wrist and fingertips, human and
robot behaviors closely align, enabling direct policy transfer. On a suite of real-
robot tasks spanning pick-and-place and tool use, Dexterous Point Policy attains 75.0% success, whereas a state-of-the-art VLA baseline reaches only 1.0%. Furthermore, our method generalizes strongly to unseen scenarios, including multi-object environments and novel object categories.
Read the full abstract ▾
All authors Beomjun Kim, Seong Hyeon Park, Seunghoon Sim, Seungjun Moon, Sanghyeok Lee, Jinwoo Shin
Machine Learning
Fast and Highly Expressive Policy Learning for Offline Reinforcement Learning via Bootstrapped Flow Q-Learning
Thanh Nguyen, Tri Ton, Hongbin Choe et al. (+2 more)
Diffusion-based Q-learning has emerged as a powerful paradigm for offline reinforcement learning, but its reliance on multi-step denoising makes both training and inference computationally expensive and brittle. Recent efforts to accelerate
diffusion Q-learning toward single-step action generation typically introduce auxiliary networks, policy
distillation, or multi-phase training, which frequently compromise simplicity, stability, or performance. To address these limitations, we introduce Bootstrapped Flow Q-Learning (BFQ), a novel framework that enables accurate single-step action generation during both training and inference, without auxiliary networks or
distillation procedures. BFQ adopts a divide-and-conquer view of the displacement vector along the flow path: it begins by learning short-range displacements that can be accurately estimated from the Flow Matching marginal velocity, and bootstraps these components to directly learn a noise-to-action mapping in a single step. This formulation eliminates multi-step denoising, resulting in a learning procedure that is substantially faster, simpler, and more robust. Extensive D4RL evaluations show that BFQ improves performance while significantly reducing computational cost compared to multi-step
diffusion baselines, demonstrating that single-step action generation suffices for high-performance offline Reinforcement Learning.
Read the full abstract ▾
All authors Thanh Nguyen, Tri Ton, Hongbin Choe, Tung M. Luu, Chang D. Yoo
Machine Learning
Geometry-Aware Reinforcement Learning for 2D Irregular Nesting
Auguste Lehuger and Guillaume Henon-Just
Traditional heuristic solvers for the 2D irregular nesting problem share a fundamental limitation: they are blind to polygon geometry, relying on guided brute-force to navigate the continuous placement space with minimal geometrical guidance. In this paper, we argue that Reinforcement Learning is uniquely positioned to overcome this bottleneck. By pairing an optimization policy with a geometry-aware neural encoder, an
agent can automatically discover rich
geometric priors directly from data, utilizing these learned intuitions to strategically guide exploration. To realize this, we introduce the Polygons
Transformer (PoT), a novel architecture that encodes 2D continuous vector geometries while allowing cross-polygon
attention. We couple this novel architecture with a Combinatorial Optimization Reinforcement Learning (CORL) training framework to find optimal solutions. To support this paradigm, we release an open-source training
dataset derived from complex geographic contours alongside a dedicated evaluation
benchmark. Empirically, our
agent slightly exceeds Sparrow, the state-of-the-art heuristic, on small (4-polygon) instances, while a clear scaling gap remains on larger (8-polygon) instances.
Read the full abstract ▾
Machine Learning
Causal Ensemble Agent: Hierarchical Causal Discovery with LLM-guided Expert Reweighting
Xinyu Li, Yuanyuan Wang, Haoxuan Li et al. (+7 more)
Causal discovery aims to uncover causal structures from observational data, which is crucial for real-world decision-making. However, different causal discovery
algorithms can produce divergent results that conflict with each other, complicating the identification of accurate causal graphs. Traditional approaches rely on numerical values and statistical assumptions, often ignoring rich domain-specific information, such as feature descriptions, which could also help structure learning. While recent works explore using Large
Language Models (LLMs) to infer causal relations via direct queries, such methods can be unreliable due to a lack of
alignment with the actual data. To address these limitations, we propose Causal Ensemble
Agent (CEA), a novel framework that aggregates structural insights from statistical discovery experts across different graph levels via linear opinion pooling, and uses an LLM as a meta-referee to dynamically reweight experts when the aggregated confidence is close to the decision boundary, thereby composing an improved and more complete causal graph. Extensive experiments on both synthetic and real-world
datasets demonstrate that CEA achieves the strongest overall performance across a wide range of causal discovery methods, highlighting the effectiveness of using LLMs for meta-analysis in causal discovery.
Read the full abstract ▾
All authors Xinyu Li, Yuanyuan Wang, Haoxuan Li, Chuan Zhou, Erdun Gao, Bo Han, Tongliang Liu, Kun Zhang, Howard Bondell, Mingming Gong
Machine Learning · Cross-listed from math.NA
Dmsh: A Multi-Agent Reinforcement Learning Framework for All-Quad Mesh Generation
Anirudh Kalyan, Cosmin Anitescu, Xiaoying Zhuang et al. (+3 more)
Generating high-quality meshes for arbitrary geometries remains a fundamental bottleneck in computational engineering, often demanding heuristic tuning and semi-manual workflows. In this paper, we introduce Dmsh, a first fully automated reinforcement learning pipeline that unifies
geometric decomposition and quadrilateral mesh generation within a single learning-based framework. Dmsh decomposes the problem through three coordinated
agents handling
topology simplification,
geometric regularization, and mesh generation. The meshing process is formulated as a Markov Decision Process and solved using a parametric Soft Actor-Critic architecture with decoupled critics, enabling efficient exploration of a hybrid discrete-continuous action space. A curriculum learning strategy ensures scalability from simple domains to highly complex geometries, suppressing seed variance. By design, the recursive decomposition enables parallel meshing of subregions, yielding globally conforming all-quadrilateral meshes without post hoc correction. Across a wide range of
benchmarks, Dmsh consistently outperforms existing methods in automation, robustness, and mesh quality, establishing a new paradigm for learning-based mesh generation.
Read the full abstract ▾
All authors Anirudh Kalyan, Cosmin Anitescu, Xiaoying Zhuang, Timon Rabczuk, Somdatta Goswami, Sundararajan Natarajan
Machine Learning · Cross-listed from eess.SY
Toward Proactive RF Charging Scheduling: Generative AI for Decision Support
Amirhossein Azarbahram, Osmel M. Rosabal, David Ernesto Ruiz-Guirola et al. (+3 more)
Radio frequency wireless power transfer (RF-WPT) is an enabling technology for supporting uninterrupted communications in future Internet of Things systems by reducing the need for battery replacement and mitigating battery-waste-related issues. For large-scale RF-WPT deployment, one of the main challenges is the scheduler-level resource allocation. Specifically, the RF charger must decide how much energy to deliver, when, and to whom, under limited charging resources, incomplete receiver-side information, and uncertain near-future charging conditions. This article positions generative artificial intelligence (GenAI) as a promising tool for this setting because it can foresee multiple plausible charging scenarios conditioned on coarse operational context and receiver-side information. We propose GenAI to act as an uncertainty-aware support layer for the RF-WPT scheduler rather than as a standalone
forecasting or decision-making tool. To this end, we first revisit the main challenges of RF-WPT scheduling, and discuss how major GenAI families can support uncertainty-aware charging decisions by generating scenario-based inputs for downstream tasks. We then present a case study showing that distribution-aware prediction can improve robust charging decisions over deterministic, ensemble, and non-learning baselines, particularly under risk-sensitive objectives. Finally, we outline key open challenges and future research directions.
Read the full abstract ▾
All authors Amirhossein Azarbahram, Osmel M. Rosabal, David Ernesto Ruiz-Guirola, Melike Erol-Kantarci, Kaibin Huang, Onel L. A. López
Machine Learning
Embedding Hybrid Systems into Continuous Latent Vector Fields
Sangli Teng, Hang Liu and Koushil Sreenath
This work proves that an $n$-dimensional hybrid system can be embedded into an $m$-dimensional Euclidean space equipped with a continuous vector field on its embedded image whenever $m>2n$. This result suggests that an intrinsically discontinuous hybrid system generically admits a continuous extrinsic representation that is well-posed for differentiable optimization. Building on this existence
theorem, we show that a latent Neural ODE with consistency loss in both the latent and state space can accurately recover the flow of hybrid systems. Extensive experiments suggest the proposed method outperforms the existing method in learning hybrid systems with varying geometries from only time series data.
Read the full abstract ▾