Machine Learning · Cross-listed from cs.AI
CrowdMath: A Dataset of Crowdsourced Mathematical Research Discussions
Sherin Muckatira, Jesse Geneson, Slava Gerovitch et al. (+3 more)
Large
language models have made substantial progress on mathematical
reasoning, but existing
benchmarks typically evaluate well-specified problems with final answers, step-by-step solutions, or complete proofs. They do not capture collaborative open-problem solving: a setting in which participants propose partial arguments, identify gaps or errors in prior steps, repair flawed
reasoning, and gradually synthesize incremental contributions into a proof. We introduce CrowdMath, a
dataset of 164 expert-annotated progress chains from the MIT
PRIMES–Art of Problem Solving (AoPS) CrowdMath program (2016-2025), a collaborative research initiative whose discussions have led to peer-reviewed publications. Each chain traces a multi-participant forum discussion from an open-problem statement to a completed proof. Posts are labeled by their functional roles in the evolving solution process, including partial progress, proof completion, erroneous
reasoning, and error identification. We define evaluation tasks and
benchmark six frontier models. Models achieve 83-88% accuracy on next-post prediction, suggesting that they can follow the local flow of mathematical discussion. However, they struggle to identify the functional significance of individual contributions with the best model achieving only 0.42 macro-F1 on post-role classification. CrowdMath exposes a gap between solving well-specified mathematical problems and understanding collaborative mathematical progress as it unfolds.
Read the full abstract ▾
All authors Sherin Muckatira, Jesse Geneson, Slava Gerovitch, Pavel Etingof, Mikhail Gronas, Anna Rumshisky
Machine Learning · Cross-listed from eess.IV
Advanced Flood Prediction with Physics-Guided Deep Learning: Combining UNet, FNO, and SAR/Optical Imagery
Tewodros Syum Gebre, Jagrati Talreja and Leila Hashemi-Beni
Accurate and scalable flood mapping remains challenging due to limited ground observations, heterogeneous terrain conditions, and the difficulty of enforcing hydrodynamic consistency within data-driven models. This work introduces a physics-guided deep learning framework that integrates multi-modal remote sensing (Sentinel-1 SAR, Sentinel-2 optical imagery, and DEM-derived terrain features) with constraints from the depth-averaged shallow water equations (SWE). The proposed hybrid architecture combines a UNet to capture fine-scale spatial details with a
Fourier Neural Operator (FNO) to model basin-scale hydraulic interactions, while physics-informed residual losses ensure mass and momentum consistency. Evaluated across diverse floodplain settings, the hybrid model achieves an Intersection over Union of 0.82 and an F1 score of 0.90 for flood extent prediction, outperforming UNet-only and FNO-only baselines. Using hydrodynamic simulations as reference data, the model achieves an RMSE of 0.21 m for water depth and 0.15 m/s for flow velocity. Physics consistency is maintained, with low residuals and mass imbalance below 2.1%. Ablation studies confirm that removing physicsbased regularization significantly degrades performance, underscoring the value of physical constraints for stability and generalization. These results demonstrate that
embedding hydrodynamic principles into deep learning yields more accurate, reliable, and physically coherent flood predictions, offering strong potential for operational monitoring and large-scale deployment.
Read the full abstract ▾
Machine Learning · Cross-listed from cs.AI
Lean4Agent: Formal Modeling and Verification for Agent Workflow and Trajectory
Ruida Wang, Jerry Huang, Pengcheng Wang et al. (+3 more)
Equipping Large
Language Models (LLMs) to execute reliable multi-step workflows has become a central challenge in artificial intelligence. Despite recent advances in LLMs' agentic capabilities, most
agent systems still lack formal methods for specifying, verifying, and debugging their workflow and execution trajectories. This challenge mirrors a long-standing problem in mathematics, where the ambiguity of natural
languages (NLs) motivates the development of formal
languages (FLs). Inspired by this paradigm, we propose **Lean4
Agent**, to the best of our knowledge, the first framework that uses Lean4, a dependent-type FL to model and verify
agent behavior. **Lean4
Agent** launches **FormalAgentLib**, an extensible Lean4 library for formally modeling and verifying
agent workflows' semantic consistency under explicit assumptions, and enabling localization of execution-time failures revealed by trajectories. Building on **FormalAgentLib**, we further develop **LeanEvolve**, which applies results in **FormalAgentLib** to revise workflows to enhance its capability. Extensive experiments on a hard problem subset of SWE-Bench-Verified and a subset of ELAIP-Bench across 5 leading LLMs indicate that the verification-passing workflows outperform the failing ones by an average of **11.94%**, and **LeanEvolve** further improves SWE performance by **7.47%** on average. Furthermore, **Lean4
Agent** establishes a foundation for a new field of using expressive dependent-type FL to formally model and verify
agent behavior.
Read the full abstract ▾
All authors Ruida Wang, Jerry Huang, Pengcheng Wang, Xuanqing Liu, Luyang Kong, Tong Zhang
Machine Learning · Cross-listed from cs.AR
P-Cast Precision in FP8 Attention: Sink-Induced Collapse and the Optimality of S=2^8
Reed Lau
FP8 (E4M3) acceleration for
attention computation offers significant throughput gains, but the 3-bit mantissa introduces precision challenges when the softmax probability
matrix~$P$ is cast to FP8 before the $P \cdot V$
matrix multiplication. We analyze two implementation choices that affect output precision under the
Attention Sink phenomenon: (1)~the KV block iteration order, and (2) the static scaling factor applied to $P$ before casting. We show that forward KV iteration causes P-collapse – to leading order a fraction $\Phi(\Delta + \delta_k - 6.93 - \ln S)$ of non-sink $P$ values underflow to zero, where the small shift $\delta_k \approx 1$ (for $k_{\text{sink}}{=}4$) is the expected within-sink-block score maximum – and that reverse iteration removes it, with a zero-underflow guarantee when reverse is combined with $S{=}256$. We further give a constructive characterization of $S = 256 = 2^8$ as the static scale that simultaneously satisfies (i)~bit-exact IEEE 754 scaling, (ii) the lower envelope of a sawtooth function $dp(S)$ over the E4M3 number line ($dp = 2^{-4}$, the minimum worst-case quantization step), and (iii)~the maximum normal-range coverage \emph{among bit-exact ($2^k$) scales} (a non-bit-exact scale such as $448$ attains slightly higher coverage; sec.5}). Both optimizations are already deployed in FlashAttention-3/4 on engineering grounds; our contribution is a quantitative account of why these choices are good and a closed-form threshold $\Delta_c = 6.93 + \ln S - \delta_k$ for predicting kernel-level precision loss. Kernel-faithful experiments ($Q, K, V$ in FP32 to isolate the P-cast effect) show $3$-$10\times$ MSE improvement at moderate sink strengths, and paired tests confirm both fixes saturate to the same precision floor when combined – which motivated updating the hpc-ops kernel from $S{=}1$ to $S{=}256$.
Read the full abstract ▾
Machine Learning · Cross-listed from cs.AI
SafeGene: Reusable Adapters for Transferable Safety Alignment
Yanghan Wang, Zhiqiang Kou, Fu Feng et al. (+2 more)
Open-weight LLMs are increasingly fine-tuned into customized assistants, but downstream fine-tuning can weaken safety
alignment and make models more vulnerable to malicious
prompts, even when the training data is not intentionally harmful. This creates a recurring safety recovery problem as target models are repeatedly updated with new task data or user interactions. We propose SafeGene, a reusable safety-adapter module designed for cross-task reuse within each architecture-compatible model family. Rather than treating safety recovery as a model-specific repair step, SafeGene treats safety capability as an independent, reusable adapter representation decoupled from task-specific updates. This representation is obtained from aligned–degraded model discrepancies, refined into task-transferable safety vectors through data-aware layer selection, and expressed in each downstream task-adapted model via few-shot layer-wise coefficient recalibration. Experiments across multiple model families, downstream tasks, and safety judges show that SafeGene-enhanced models reduce harmful response rates while maintaining downstream performance, outperforming representative safe adaptation methods in safety–utility trade-off.
Read the full abstract ▾
All authors Yanghan Wang, Zhiqiang Kou, Fu Feng, Jing Wang, Xin Geng
Machine Learning · Cross-listed from cs.AI
DiBS: Diffusion-Informed Branch Selection
Bo Liu, Yuan Xie, Yuan Gao et al. (+4 more)
Sudoku is a representative constraint satisfaction problem that requires global structural
reasoning under strict discrete constraints. The existing works of solving Sudoku mainly focus on two dominant approaches, i.e., traditional heuristic and deep learning solver. However, they suffer from two complementary limitations: learning-based solvers lack hard correctness guarantees, while complete symbolic solvers are still prone to long-tail search. To address these shortcomings, we propose a novel
diffusion model-guided approach, termed as DiBS, for the branch selection search process. Specifically, DiBS keeps the symbolic solver complete and uses the
diffusion model as a branch-ordering guide. The core method is ranking candidate values under the current partial assignment and lightweight consistency signal. Furthermore, we provide an in-depth theoretical proof to reveal how it works and why it works. Experiments on the challenging Royle 17-clue Sudoku
benchmark show that our DiBS substantially reduces search cost relative to strong heuristic baselines, especially in nodes, backtracks, and long-tail percentiles. Besides, these results confirm that learned global guidance is effective on hard instances where branch-order mistakes are most expensive. All codes are available at https://github.com/shanxierdan/DiBS.
Read the full abstract ▾
All authors Bo Liu, Yuan Xie, Yuan Gao, Xiaolong Luo, Peng Ye, Tao Chen, Fujun Han
Machine Learning · Cross-listed from q-bio.QM
Probabilistic learning to perform pre-onset individualised prediction of disease severity: application to Veno Occlusive Disease
Dalia Chakrabarty, Kane Warrior, Chuqiao Zhang et al. (+2 more)
We advance a new probabilistic supervised learning approach that permits reliable, automated, and early individualised prediction of the severity with which a disease will develop in a prospective patient. The prediction capacity is illustrated via the pre-transplant prediction of the score of severity of Veno Occlusive Disease (or VOD) in the digital twin (DT) of the considered prospective patient, where this score parametrises the severity with which VOD will develop in this patient, after they undergo their Bone Marrow Transplant. The learning of the relationship between the pre-transplant variables, and a severity score variable is undertaken by modelling this relationship as a (random) function that is treated as a sample function of an adequately-chosen
stochastic process. The parameters of this underlying process are learnt using a training
dataset that is generated using the real-time
evolution of retrospective patients in a cohort, with this training
dataset subsequently augmented in size by a probabilistic inverse learning of the score of prospective patients. The augmented training set, then permits the learning of the function that capacitates - at the pre-transplant stage - automated prediction of the score of the severity of VOD that characterises the DT of a physical patient in their unique pre-transplant state. This score is subsequently fed back to the real prospective patient as the severity with which VOD will develop in them, after this patient undergoes their transplant. Such a score then permits the treating Haematologist-Oncologists to decide on the treatment regimen, which in this illustration reduces to deciding on treating the patient with Defibrotide. An AI facility is developed to undertake such automated prediction, with the physician inputting the data on the pre-transplant state that characterises the DT of the prospective patient under consideration.
Read the full abstract ▾
All authors Dalia Chakrabarty, Kane Warrior, Chuqiao Zhang, Akash Bhojgaria, Joydeep Chakrabartty
Machine Learning · Cross-listed from cs.AR
DxPTA: An Architecture Design Space Exploration with Optical Dataflow-guided Strategy for HW/SW Co-Design of Photonic Transformer Accelerators
Rachmad Vidya Wicaksana Putra, Solomon Micheal Serunjogi, Mahmoud Rasras et al. (+1 more)
Transformer-based networks have emerged as prominent AI models with state-of-the-art performance, which potentially pave the way toward artificial general intelligence (AGI). However, their large sizes still hinder their efficient implementation, thus highlighting the need for alternate solutions to enable their energy-efficient acceleration. Recently, state-of-the-art works propose photonic
transformer accelerators (PTAs) with significant speedup and energy efficiency improvements over the conventional electronic accelerators. However, their PTA architectures are developed without considering the application constraints (e.g., area, power, energy, and latency). Moreover, their manual design approach also requires huge design time to determine a suitable architecture for the targeted application, hence making this approach not scalable. To address these limitations, we propose DxPTA, a novel design space exploration methodology for enabling efficient hardware/software co-design of the appropriate PTA architecture that meets all constraints. It is achieved by (1) identifying the PTA architecture parameters based on the coherent optical dataflow; (2) analyzing the impact/significance of the parameters; and (3) leveraging this analysis for devising a constraint-aware architecture search
algorithm. Experimental results show that, our DxPTA can find the appropriate PTA architectures for different
transformer-based models (i.e., DeiT-T/S/B and BERT-B/L). It achieves up to 26mm^2 area, 4.8W power, 39mJ energy, and 6ms latency, for constraints of 50mm^2 area, 5W power, 50mJ energy, and 10ms latency; with 15.2x faster searching time than the exhaustive approach. These results demonstrate the potential of DxPTA methodology for enabling efficient PTA designs for diverse AGI-based applications.
Read the full abstract ▾
All authors Rachmad Vidya Wicaksana Putra, Solomon Micheal Serunjogi, Mahmoud Rasras, Muhammad Shafique
Machine Learning · Cross-listed from cs.AI
Detecting and Mitigating Bias by Treating Fairness as a Symmetry Operation
Nishit Singh
Machine learning systems deployed in high stakes socioeconomic settings routinely display bias. We formalize bias as a
symmetry breaking operation: a classifier is fair if its outputs remain invariant under the counterfactual operation of switching a sensitive attribute, with merit features held fixed. We implement loss based regularization as a
symmetry restoring mechanism and evaluate the framework on four synthetic
datasets with varying levels of noise, correlation, and bias. The framework achieves upwards of 90% violation reduction, with accuracy costs around 5%. This framework does not require causal graph knowledge, is computationally lightweight, and generalizes to any sensitive attribute definable as a bit-flip, making it suitable for contexts where local sources of discrimination remain absent from mainstream
benchmarks.
Read the full abstract ▾
Machine Learning · Cross-listed from stat.ML
Central Description Length (CDL) Clustering Validation Index
Mahdi Shamsi and Soosan Beheshti
Selecting a clustering
algorithm and its hyperparameters without labels is a common difficulty in engineering machine learning pipelines that work with unsupervised analysis of sensor, image, or process data. Clustering validation indices (CVIs) provide internal scores for ranking candidate clusterings, but most popular CVIs are built from Euclidean compactness and separation terms and so tend to favour compact, convex partitions. Their performance is known to degrade on non convex, irregular, or variable density data, where kernel transformations or alternative distance measures are typically used at the cost of additional tuning and computation. This paper introduces the Central Description Length (CDL) clustering validation index. CDL uses the observed within cluster compactness, the estimated cluster centers, and the estimated cluster covariances to compute a probabilistic upper bound on the description length associated with the unobservable true cluster centers. The bound condenses intra cluster compactness and centroid displacement into a single computable quantity and is evaluated on the partition produced by any clustering
algorithm. The implementation uses only observable quantities (the data, the partition, the estimated centers, and the estimated covariances) and does not use ground truth labels. On synthetic
benchmarks with non convex and arbitrary shape clusters, CDL-CVI selected the reference number of clusters more often and reached higher Adjusted Rand Index (ARI) values than the conventional CVIs we tested, without an additional kernel preprocessing stage. On image
benchmarks (MNIST, CIFAR-10, STL-10) clustered from frozen unsupervised
embeddings, CDL-CVI returned cluster numbers close to the reference class counts across K-means, DBSCAN, and spectral clustering in the reported trials.
Read the full abstract ▾
Machine Learning · Cross-listed from q-bio.CB
Quantifying the biophysical properties of stomatocytes in health and disease
Zhaojie Chai, Jianlu Zheng, He Li et al. (+2 more)
Hereditary stomatocytosis (HS) comprises red blood
cell (RBC) disorders characterized by cup-shaped erythrocytes that respond oppositely to splenectomy: curative in overhydrated HS (OHS) but potentially thrombogenic in dehydrated HS (DHS/xerocytosis). This paradox persists because RBC biomechanics is governed by partly independent parameters–shear modulus, bending
rigidity, surface-to-volume ratio (S/V), and cytoplasmic viscosity–that existing assays capture only piecemeal. Here we combine dissipative particle dynamics (DPD) simulations with microfluidic imaging to construct a control discocyte and three stomatocyte models (ST-RBC1-3) at fixed membrane area and decreasing volume (109.7, 101.5, 89.8 fL), spanning the OHS-to-DHS range. Tracing this parameter set through five mechanically orthogonal assays, we find that interendothelial-slit (IES) traversal is geometry-dominated: overhydrated ST-RBC1 requires an order of magnitude higher critical pressure than healthy RBCs, whereas dehydrated ST-RBC3 passes freely. ST-RBC3 nonetheless suppresses membrane tank-treading and raises low-shear whole-blood viscosity by ~29% at physiological haematocrit, comparable to Gaucher-disease hyperviscosity. A funnel-obstacle chip amplifies these differences into a label-free centerline-offset signal predicted to separate all four RBC types (~4.5 standard deviations between extreme phenotypes). These results unite single-
cell mechanics, splenic filtration, and hemorheology in one framework, resolve the splenectomy paradox, and point toward microfluidic pre-operative risk stratification in HS.
Read the full abstract ▾
All authors Zhaojie Chai, Jianlu Zheng, He Li, Ming Dao, George Em Karniadakis
Machine Learning · Cross-listed from q-bio.QM
The Language of Elution: Autoregressive Prediction of the Next Feature in Untargeted LC-HRMS Lipidomics
Dayanjan S. Wijesinghe
Untargeted
liquid chromatography-high-resolution mass spectrometry (LC-HRMS) detects thousands of molecular features per sample, yet only 2-20% receive confident structural annotations. A root cause of this "dark metabolome" is that tandem MS/MS acquisition is reactive: instruments select precursors only after ions appear, blind to what elutes next. We reframe chromatographic elution as an autoregressive sequence prediction task. Because reversed-phase elution order is governed by hydrophobicity, successive features form a physically constrained sequence, like
tokens in
language. We discretize the mass-to-charge (m/z) axis into 110 bins and train long short-term
memory (LSTM) and
Transformer models to predict the next eluting m/z bin from five annotation-free per-
token features: m/z bin, mass defect, retention-time gap, polarity, and intensity rank. Trained on 15,242 features from four clinical lipidomics cohorts (342
plasma samples; SCIEX TripleTOF 6600+, Waters CSH C18), the LSTM reaches 98.4% top-1 accuracy (99.99% top-5; mean absolute error 3.6 Da) and the
Transformer 98.0%. Ablation shows autoregressive context accounts for 55.5 percentage points while no single feature contributes more than 0.2 pp: the sequential pattern, not molecular properties, drives prediction. Models transfer across instruments sharing the method (r=0.999 on an independent Agilent 6530
dataset) but fail under a different column chemistry (5.1% top-1) or polarity mode (2.6%), confirming method- and mode-specificity. Fine-tuning on as few as two to five quality-control injections recovers held-out accuracy from 2.6% to nearly 50%, so cross-condition deployment needs minimal calibration. These results establish that elution sequences are highly predictable and lay the groundwork for predictive MS/MS acquisition to improve annotation coverage in untargeted metabolomics.
Read the full abstract ▾
Machine Learning
PE-MHL: Physics-Encoded Modular Hybrid Layers for Scalable Learning of Complex Systems
Ismail Hassaballa and Mircea Lazar
Hybrid models that combine physics-based and data-driven components have shown strong potential for achieving accuracy and interpretability in control applications. While recent methods have made progress in incorporating physical consistency, challenges remain in scalability, robustness to noise, and control of model complexity. This paper proposes a Physics-Encoded Modular Hybrid Layer (PE-MHL) framework, in which a baseline physics-based model is incrementally refined through the addition of new sub-models, where each new component adds complexity while preserving what previous components have already learned. We establish a theoretical guarantee for this construction: with a least-squares initialization of each new sub-model, the training error is monotonically non-increasing in the number of sub-models and provably converges. Empirical evaluations on a nonlinear NARX
benchmark and the Quanser Aero 2 platform demonstrate that PE-MHL matches or outperforms its fully connected monolithic relaxation (over five times more parameters) in accuracy and generalization, while also providing more stable training dynamics and better preservation of underlying data structures.
Read the full abstract ▾
Machine Learning
Scaling Novel Graph Generation via Lightweight Structure-Guided Autoregressive Models
Alessio Barboni, Massimiliano Lupo Pasini, Bishal Lakha et al. (+1 more)
Generating realistic and diverse graphs is a key problem in machine learning, with applications in molecular discovery,
circuit design, cybersecurity, and beyond. However, current graph generative models remain limited by scalability and novelty.
Diffusion-based methods often require costly full-adjacency operations and long denoising chains, while many autoregressive and hybrid models have at least quadratic complexity. In addition, these models often imitate training graphs rather than generalize beyond them. We propose a lightweight autoregressive framework to address these issues. It uses a structure-guided topological ordering to serialize graphs into regular edge sequences, enabling near log-linear generation, and a two-phase training strategy that combines exploration-oriented augmentation with iterative refinement to reduce overfitting and promote controlled novelty. Experiments on molecular and non-molecular
benchmarks show that our approach improves novelty while preserving high validity and uniqueness. The framework also supports both LSTM and Mamba-style causal sequence backbones, with large-
memory accelerators enabling longer graph-sequence experiments beyond typical GPU limits.
Read the full abstract ▾
All authors Alessio Barboni, Massimiliano Lupo Pasini, Bishal Lakha, Edoardo Serra
Machine Learning
Interpretability Is Not Just a Tax: Sparsity and Diversity Yield Legible and Composable Reward Experts for Personalization
Yifan Wang, Jinyi Mu, Mayank Jobanputra et al. (+5 more)
Preference modeling aims to capture heterogeneous human values and increasingly relies on learning latent components to represent this variation without fine-grained supervision. What these components actually represent, however, is rarely examined. We audit Mixture-of-Experts (MoE) reward models along three rungs—semantic coherence, functional correspondence, and composability—in controlled and real-world settings, holding the architecture and training data fixed so that differences isolate the effect of the training objective. Our diagnosis is that the components conventional MoE reward models learn are entangled, which limits not only interpretability but also the personalization such models are built for. The
entanglement is not intrinsic to the architecture but a property of the training objective: off-the-shelf sparsity and diversity regularizers already substantially reduce it. The resulting configuration, which we call sparse MoE, recovers all underlying semantic categories without domain supervision in a controlled setting, reaching 84.8% expert purity against 60.8% for conventional MoE; on real-world data its experts admit human-interpretable task- and domain-level descriptions aligned with their functional strengths. Targeted interventions on routing weights produce predictable preference changes, showing that these interpretations are behaviorally grounded rather than merely descriptive. Crucially, the disentanglement induced by this objective also makes the experts composable: lightweight router updates yield a 25.81-point improvement in few-shot personalization with 50 examples, roughly 2.7 times the gain of the identical architecture and adaptation procedure applied to entangled experts. Together, these results show that MoE reward models can learn interpretable expert structure, and that the same structure can substantially improve personalization.
Read the full abstract ▾
All authors Yifan Wang, Jinyi Mu, Mayank Jobanputra, Yu Wang, Ji-Ung Lee, Soyoung Oh, Isabel Valera, Vera Demberg
Machine Learning
The Loss Is Not Enough: Sampling Conditions and Inductive Bias in Contrastive Representation Learning
Justinas Zaliaduonis, Patrick Putzky, Till Richter et al. (+1 more)
Contrastive learning has become a leading paradigm for self-supervised representation learning, yet the conditions under which it recovers meaningful latent geometry remain incompletely understood. We develop a measure-theoretic framework formalizing the diversity condition, a support requirement on positive-pair sampling that is necessary for isometric latent recovery. We show that the standard full-support von Mises-Fisher setting implies the satisfaction of the diversity condition and as a consequence global contrastive loss minimizers recover latent geometry up to orthogonal transformation, while restricted conditionals can make non-orthogonal maps attain strictly lower asymptotic contrastive loss. We introduce a support-corrected Information Noise Contrastive Estimation (InfoNCE) variant as a theoretical fix: this correction makes orthogonal latent space recovery achievable but does not uniquely select it. Experiments on synthetic
benchmarks validate the identifiability predictions, and CIFAR-10 experiments are consistent with the qualitative prediction that architectural inductive bias becomes more important when sampling diversity is limited. Together, our results clarify how sampling mechanisms and encoder inductive bias interact in contrastive representation learning.
Read the full abstract ▾
All authors Justinas Zaliaduonis, Patrick Putzky, Till Richter, Sergios Gatidis
Machine Learning
Derivative Informed Learning of Exchange-Correlation Functionals
Eike S. Eberhard, Luca A. Thiede, Abdul Aldossary et al. (+5 more)
Machine-learned (ML) exchange-correlation (XC) functionals aim to replace human-designed density functional approximations by learning directly from reference data, but they still do not consistently outperform traditional $\mathcal{O}(N^4)$-scaling hybrid functionals. We study a hybrid-
distillation setting in which $\mathcal{O}(N^3)$-scaling ML-XC functionals are trained to reproduce B3LYP/def2-SVP targets. We introduce Derivative Informed XC-Loss (DI-Loss), a loss that incorporates additional information from the reference hybrid functional by supervising first and second derivatives of the energy on the Grassmannian of admissible density
matrices. Rather than only matching the self-consistent fixed point, DI-Loss aligns the local first- and second-order response of the learned functional with that of the target functional. Across four evaluated architectures, DI-Loss consistently improves the main energy metrics. Averaged uniformly across architectures, the total-energy MAE decreases by 66% relative to energy and density supervision alone. The density-sensitive mean-field energy metric $E_\rho$ improves from $1.2$ to $0.8$ mEh on average, while dipole and $\mathcal{L}_2$ density errors do not improve uniformly. We further show that densities from the distilled functionals reduce hybrid-functional SCF iterations by up to 50%. In downstream TDDFT calculations, Hessian supervision improves excited-state predictions, with XCdiff reducing the mean excitation-energy MAE by 19 - 35%.
Read the full abstract ▾
All authors Eike S. Eberhard, Luca A. Thiede, Abdul Aldossary, Andreas Burger, Nicholas Gao, Vignesh Bhethanabotla, Alán Aspuru-Guzik, Stephan Günnemann
Machine Learning
From Ticks to Flows: Dynamics of Neural Reinforcement Learning in Continuous Environments
Saket Tiwari, Tejas Kotwal and George Konidaris
We present a novel theoretical framework for deep reinforcement learning (RL) in continuous environments by modeling the problem as a continuous-time
stochastic process, drawing on insights from
stochastic control. Building on previous work, we introduce a viable model of actor-critic
algorithm that incorporates both exploration and
stochastic transitions. For single-hidden-layer neural networks, we show that the state of the environment can be formulated as a two time scale process: the environment time and the gradient time. Within this formulation, we characterize how the time-dependent random variables that represent the environment's state and estimate of the cumulative discounted return evolve over gradient steps in the infinite width limit of two-layer networks. Using the theory of
stochastic differential equations, we derive, for the first time in continuous RL, an equation describing the infinitesimal change in the state distribution at each gradient step, under a vanishingly small learning rate. Overall, our work provides a novel nonparametric formulation for studying overparametrized neural actor-critic
algorithms. We empirically corroborate our theoretical result using a toy continuous control task.
Read the full abstract ▾
Machine Learning
RL Excursions during Pre-Training: Re-examining Policy Optimization for LLM training
Rachit Bansal, Clara Mohri, Tian Qin et al. (+2 more)
The standard LLM training pipeline applies reinforcement learning (RL) only after pre-training and supervised fine-tuning (SFT). We question this status quo by training a LLM from scratch and applying RL, SFT, and SFT followed by RL directly to intermediate pre-training checkpoints. We find that RL is effective very early, and often matches the full SFT$\to$RL pipeline early as well. Through experiments on harder problems, we find that targeted pre-training data composition is a strong lever for RL effectiveness, even more so than model scale. Beyond
reasoning accuracy, applying RL directly to base checkpoints expands the model's distribution; the sharpening effect reported in recent work arises only when RL follows SFT. The general capabilities of the model remain essentially unchanged by RL, while they degrade following SFT. Finally, we merge RL and SFT objectives by parallel averaging, which outperforms across all other training methods discussed, across metrics, while preserving general capabilities. Together, these results suggest that LLM training might benefit from an expanded use of RL.
Read the full abstract ▾
All authors Rachit Bansal, Clara Mohri, Tian Qin, David Alvarez-Melis, Sham Kakade
Machine Learning · Cross-listed from cs.CR
Long-Term and Short-Term Transistor Aging in Deep Neural Networks: Impact and Mitigation
Alireza Sarmadi, Virinchi Roy Surabhi, Prashanth Krishnamurthy et al. (+3 more)
Deep neural networks (DNNs) are used in a variety of real-world applications including, for example, image classification and
speech recognition. The inference accuracy of DNN implemented on hardware in integrated
circuits (ICs) degrades under phenomena such as transistor aging. Aging slows down the switching speed of transistors, resulting in system-level timing violations due to unsustainable clocks. To maintain reliability for the entire projected lifetime, designers add guardbands to prevent timing violations; however, adding large timing guardbands causes losses in performance (speed or throughput). This chapter provides a detailed discussion of the effects of long-term and short-term transistor aging on DNN inference accuracy. Furthermore, to mitigate aging effects on DNN's accuracy and keep them at bay, a methodology for aging-aware retraining is presented in order to generate a resilient DNN even when aggressive (i.e., smaller than required) guardbands are used. This improves the inference accuracy of the DNNs even in the presence of aging-induced degradation. These effects are discussed in this chapter along with mitigation strategies on a hardware implementation of a DNN for image classification on an off-the-shelf image
dataset. The application of short-term aging as an excitation mechanism for the detection of hardware Trojans in integrated
circuits is also briefly discussed.
Read the full abstract ▾
All authors Alireza Sarmadi, Virinchi Roy Surabhi, Prashanth Krishnamurthy, Hussam Amrouch, Ramesh Karri, Farshad Khorrami
Machine Learning · Cross-listed from math.OC
Nonlocal Mean Field Schrödinger Bridge with Learned Interactions
Daisuke Inoue, Dante Kalise and Mathieu Laurière
The Schrödinger Bridge Problem connects an initial distribution to a terminal one along a minimum-energy
stochastic process. Its mean-field extension, the Mean-Field Schrödinger Bridge, governs interacting populations whose dynamics and costs depend on the collective distribution. When these interactions are nonlocal, their direct evaluation scales quadratically with the population size, making large ensembles intractable within FBSDE-based solvers. We replace these terms with neural surrogates in state and time, trained on empirical interaction values along sampled trajectories and embedded in a four-stage alternating scheme that updates the forward and backward potentials and the surrogates in turn, while preserving forward–backward consistency and the prescribed endpoint marginals. We derive Grönwall-type stability bounds quantifying how surrogate errors propagate to the generated trajectories under a small-gain condition. On crowd-
navigation and high-dimensional opinion-dynamics
benchmarks, the surrogates reproduce the trajectories obtained with exact evaluation at reduced training cost. The advantage is most significant when the interaction is a nonlinear functional of the measure, such as the normalized bounded-confidence drift, for which random-batch subsampling is biased and unstable whereas the learned surrogate remains accurate.
Read the full abstract ▾
Machine Learning · Cross-listed from cs.AI
Can Generalist Agents Automate Data Curation?
Feiyang Kang, Hanze Li, Adam Nguyen et al. (+5 more)
Curating training data is among the most consequential yet labor-intensive parts of modern AI development: practitioners iteratively propose, implement, evaluate, and revise data policies against noisy
benchmark feedback. We ask whether generalist coding
agents can automate this data-curation loop. We introduce *Curation-Bench*, an
agent-centric
benchmark that fixes the model, training recipe, and evaluation suite while giving
agents command-line access to inspect data, implement policies, submit them to a fixed training/evaluation pipeline, and revise. In a
vision-
language instruction-tuning instantiation, out-of-the-box
agents reach strong published data-selection baselines within ten iterations. However, trajectory analysis reveals a persistent *execution-research gap*:
agents mainly tune local policy variants rather than explore new policy families, even when given strategy guides and paper references. Scaffolds requiring each iteration to cite, instantiate, and adapt a prior method shift
agents toward method-guided exploration. The scaffolded
agent autonomously composes – without human design input – a data-selection policy that outperforms strong published baselines at one-tenth their data budget. Overall, current
agents can run the curation loop, but reliable data research requires scaffolded method adaptation, not open-ended prompting alone. Code and
benchmark are open-sourced.
Read the full abstract ▾
All authors Feiyang Kang, Hanze Li, Adam Nguyen, Mahavir Dabas, Jiaqi W. Ma, Frederic Sala, Dawn Song, Ruoxi Jia
Machine Learning · Cross-listed from cs.AI
VAMPS: Visual-Assisted Mathematical Problem Solving Benchmark
Amirhossein Dabiriaghdam, Shayan Vassef, Mohammadreza Bakhtiari et al. (+5 more)
Multimodal large
language models are increasingly capable of complex
reasoning, yet their performance often degrades when they must externalize a problem through a tool and then reason over the tool's output, specifically when they rely on visual aids. This gap is especially important because real engineering and scientific workflows often rely on visualization tools for analysis, validation, and decision-making. To study this discrepancy, we introduce VAMPS (Visual-Assisted Mathematical Problem Solving), a
benchmark for graph-assisted mathematics. VAMPS contains 1,168
multimodal, bilingual multiple-choice question-answer pairs drawn from Iranian University Entrance Exam
algebra and calculus problems and expanded with human-reviewed LLM-generated synthetic variants, all selected so that plotting provides a natural solution strategy by revealing intersections, extrema, asymptotes, etc. Designed for both benchmarking and diagnosis, VAMPS goes beyond prior
multimodal benchmarks that primarily evaluate
reasoning over fixed visual inputs by testing whether a model can benefit from constructing a useful graph and grounding its answer in the resulting visualization. Overall, we found that across a diverse set of models, direct analytical solving surprisingly outperforms tool-enabled visual solving, even on problems where plotting is a natural strategy.
Read the full abstract ▾
All authors Amirhossein Dabiriaghdam, Shayan Vassef, Mohammadreza Bakhtiari, Yasamin Medghalchi, Ilker Hacihaliloglu, Mesrob Ohannessian, Lele Wang, Giuseppe Carenini
Machine Learning
Recover-LoRA for Aggressive Quantization: Reclaiming Accuracy in 2-Bit Language Models via Low-Rank Adaptation with Knowledge Distillation on Synthetic Data
Devleena Das, Rajeev Patwari, Elliott Delaye et al. (+1 more)
Aggressive weight quantization to 2-bit precision offers substantial throughput and
memory gains for large
language model (LLM) inference, but typically incurs severe accuracy degradation. These gains are particularly relevant for edge and on-device deployment, where
memory capacity and bandwidth are primary constraints. In this work, we extend Recover-LoRA – a lightweight, data-free accuracy recovery method originally developed for general model weight corruption – to the setting of ultra-low-bit quantization. We propose a selective mixed-precision strategy in which only gate and up projection layers of the MLP are quantized to 2-bit (W2), while all other linear layers remain at higher precision, yielding a mixed-precision GateUp configuration. We demonstrate via roofline analysis across three model families (4B–20B) and two hardware platforms that a W4/W2-GateUp deployment (4-bit base with 2-bit gate/up) delivers 7.5–23.3% TPS improvement over uniform W4 depending on model and context length, while confining quantization error to a predictable subset of layers. We then apply Recover-LoRA – training low-rank adapters on the quantized layers via logit
distillation with synthetic data – to recover accuracy lost from 2-bit quantization of the gate and up layers. In a case study on Qwen3-4B, Recover-LoRA achieves 80–95% accuracy recovery on 9 of 12
benchmarks, using only 10k synthetic training samples and no labeled data. We further demonstrate that synthetic data performs comparably to curated labeled data for
distillation-based recovery, and that recovery generalizes to out-of-distribution evaluation tasks. Our results present Recover-LoRA as a practical post-quantization accuracy recovery tool for aggressive weight compression in deployment settings.
Read the full abstract ▾
All authors Devleena Das, Rajeev Patwari, Elliott Delaye, Ashish Sirasao
Machine Learning · Cross-listed from cs.CL
Supportive Token Revealing for Fast Diffusion Language Model Decoding
Giries Abu Ayoub, Mario Barbara, Lluís Pastor-Pérez et al. (+4 more)
Discrete
diffusion language models can generate text efficiently by updating multiple masked positions in parallel, but this parallelism introduces a quality-latency trade-off. Aggressive decoding may commit mutually dependent
tokens too early, while conservative decoding requires many denoising steps. Existing methods address this tension by deciding which
tokens are safe to reveal using confidence or dependency criteria. However, avoiding unsafe commits does not necessarily make the remaining masked sequence easy to decode, since uncertain
tokens may depend on masked
tokens, creating a bottleneck for denoising steps. We propose AXON, a training-free module that can be added on top of existing parallel decoding strategies for
diffusion language models. Rather than replacing the base decoder, AXON monitors the remaining uncertain masked
tokens and intervenes only when their current state suggests that additional context is needed. It then shifts the criterion from which
tokens are safest to reveal to which confident reveals would best support later denoising. AXON selects anchors, confident masked
tokens that uncertain positions attend to, using
attention, uncertainty, and confidence signals. Experiments on
reasoning and code-generation
benchmarks across multiple
diffusion language models show that AXON improves the quality-latency trade-off of existing parallel decoders, often reducing the number of function evaluations while maintaining or improving accuracy.
Read the full abstract ▾
All authors Giries Abu Ayoub, Mario Barbara, Lluís Pastor-Pérez, Tanja Bien, Aneesh Barthakur, Alaa Maalouf, Loay Mualem
Machine Learning
Edge of Stability Selectively Shapes Learning Across the Data Distribution
Shauna Kwag, Anakha Ganesh, Tomaso Poggio et al. (+1 more)
Existing analyses of the edge of stability (EoS) treat it as a global property of optimization. We show that it is also selective: the stability constraint redistributes learning across subsets of the training distribution, amplifying progress on some groups while suppressing progress on others. Using a branching intervention that enters or exits the EoS regime from the same training state, we causally demonstrate this trade-off and identify two necessary conditions for a group to benefit. First, its aggregate gradient must align with the top Hessian eigenvector. We isolate this mechanism with a controlled perturbation that preserves distance but randomizes direction, destroying
alignment and eliminating the advantage. Second, the group must sustain non-vanishing gradient magnitude over time. Under cross-
entropy loss, gradient saturation decouples confidently classified groups, shifting the advantage to output-outliers, whose gradients persist. Together, these results show that EoS functions not only as a stability boundary, but as a mechanism governing the allocation of learning across the data distribution.
Read the full abstract ▾
All authors Shauna Kwag, Anakha Ganesh, Tomaso Poggio, Pierfrancesco Beneventano
Machine Learning · Cross-listed from eess.AS
Representation Matters in Randomized Smoothing for Audio Classification
Jong-Ik Park, Shreyas Chaudhari, José M. F. Moura et al. (+1 more)
Randomized smoothing (RS) certifies robustness in the vector space where Gaussian noise is added. In audio classification, this space is often not uniquely defined as standard pipelines normalize, range-control, and transform waveforms into log-mel or other spectral features. We show that direct RS is therefore under-specified unless the certified object and preprocessing policy are explicit. On two audio
benchmarks, keyword spotting and environmental-sound classification, we study waveform, feature-space, and post-processed smoothing. Our diagnostics show why representation-aware reporting is necessary: at the same smoothing level $\sigma=0.0025$, the two
datasets share the same median raw radius $.007996$, but different waveform energies yield different SNR-equivalent scales ($83.98$ vs. $90.97$ dB); log-mel smoothing gives higher positive-radius certified accuracy on environmental sounds ($68.42\%$ vs. $65.53\%$), certifying more examples with nonzero radius but over features rather than waveforms; and clipping or peak normalization changes the effective perturbation norm by roughly $230$–$351\times$. We therefore recommend that audio RS studies choose and report the task-specific certified object and perturbation model, including the perturbation location, gain policy, raw radius, and any post-noise geometry changes.
Read the full abstract ▾
All authors Jong-Ik Park, Shreyas Chaudhari, José M. F. Moura, Carlee Joe-Wong
Machine Learning
A Geometric View of Counterfactual Behavior: Interaction of Boundary Proximity and Local Support
Ioanna Gemou, Matteo Gamba, Randall Balestriero et al. (+1 more)
Counterfactual explanations seek small, semantically meaningful changes to an input that alter a model's prediction, and are widely used to interpret and audit machine learning systems. In modern
vision,
language, and
multimodal systems, pretrained encoders map inputs to representation spaces, and downstream classifier heads impose decision boundaries within those spaces. As a result, the feasibility and distance of nearby counterfactuals depend on boundary placement relative to the data. Yet models with similar predictive performance can differ substantially in whether such changes are achievable and how far representations must move. This work examines this variation using a standardized local search probe across several pretrained encoders and linear classifier heads. Results show that despite similar predictive performance, models differ substantially in their counterfactual behavior. Under fixed representations, varying only the classifier head alters counterfactual outcomes while leaving predictive performance largely unchanged. This variation is explained by the interaction of decision-boundary proximity and local data support, which jointly determine whether prediction changes are both feasible and lie in regions supported by the data, and can also improve counterfactual search within fixed models. Together, these findings identify counterfactual behavior as a distinct dimension beyond predictive performance and show that it can be altered without changing accuracy, with implications for model selection, robustness, and the reliability of counterfactual methods.
Read the full abstract ▾
All authors Ioanna Gemou, Matteo Gamba, Randall Balestriero, Ritambhara Singh
Machine Learning · Cross-listed from cs.MM
DetectZoo: A Unified Toolkit for AI-Generated Content Detection Across Text, Audio, and Image Modalities
Sajad Ebrahimi, Nima Jamali, Bardia Shirsalimian et al. (+8 more)
The growing popularity and capacity of generative models have eroded the distinction between human and machine-generated content, motivating a growing body of work on detection across text, images, and audio. Most available detectors are either commercial software or, if open-source, come with incompatible codebases with bespoke preprocessing, evaluation protocols, and evaluation metrics, which make their adoption, fair comparison, and reproduction quite difficult. To address this critical gap, we introduce DetectZoo, a first-of-its-kind, extensible toolkit designed to provide a unified interface for AI-generated content detection across text, audio, and image modalities. DetectZoo standardizes the complete empirical pipeline, from data ingestion and preprocessing to model assessment, offering researchers a cohesive framework to
benchmark state-of-the-art detectors systematically. By integrating diverse public
datasets and baseline detection
algorithms under a single, unified API, our toolkit facilitates rigorous and reproducible evaluation. DetectZoo provides reference implementations of 61 detectors, native loaders for 22
benchmark datasets, and a standardized evaluation pipeline that reports multiple metrics through a common interface. Each detector is self-contained yet accessible through the same interface, automatically caches pretrained weights, and reproduces the original published results. DetectZoo lowers the barrier to entry for multi-modal AI forensics, enabling researchers to identify performance gaps across domains and accelerating the development of robust, generalizable detection techniques. The open-source repository and comprehensive documentation are publicly available at https://github.com/sadjadeb/DetectZoo, and the package can be installed via pip install detectzoo.
Read the full abstract ▾
All authors Sajad Ebrahimi, Nima Jamali, Bardia Shirsalimian, Kelly McConvey, Wentao Zhang, Jalehsadat Mahdavimoghaddam, Maksym Taranukhin, Maura Grossman, Vered Shwartz, Yuntian Deng, Ebrahim Bagheri
Machine Learning · Cross-listed from cs.CL
Cross-Prompt Generalization in Detecting AI-Generated Fake News Using Interpretable Linguistic Features
Aya Vera-Jimenez, Samuel Jaeger, Calvin Ibenye et al. (+1 more)
The increasing use of large
language models has raised concerns about the spread of AI-generated fake news, particularly under varying prompting strategies. Most existing detection models are trained and evaluated under a single generation setting, leaving their ability to generalize across unseen
prompts unclear. In this study, we investigate cross-
prompt generalization in fake news detection using three
datasets of AI-generated articles produced under distinct
prompts, combined with real news articles. We extract interpretable linguistic features capturing lexical diversity, readability, and
emotion-based characteristics and evaluate a random forest classifier under a cross-
prompt framework, where models trained on one
prompt are tested on another. Across all six train-test combinations, performance remains consistently high, with AUC values ranging from 0.988 to 1.000. Analysis of feature distributions shows that AI-generated text exhibits increased lexical diversity, reduced readability, and substantially lower emotional intensity compared to the overall
dataset, with variations across
prompts. Despite these distributional shifts, the classifier maintains strong performance, indicating that these features capture stable properties of AI-generated text that generalize across prompting strategies. These findings suggest that feature-based approaches can provide robust detection of AI-generated fake news under
prompt variability.
Read the full abstract ▾
All authors Aya Vera-Jimenez, Samuel Jaeger, Calvin Ibenye, Dhrubajyoti Ghosh