Machine Learning · Cross-listed from stat.ML
SGD Provably Prioritizes a Shortcut Spurious Feature in the XOR Model
Tyler LaBonte and Vidya Muthukumar
Neural networks are known to be susceptible to over-reliance on spurious correlations. However, the precise mechanism by which models exploit shortcut features is not fully understood, and
algorithms to mitigate this behavior rely on as yet unjustified assumptions about the learned representations. In this work, we provide the first end-to-end theoretical characterization of spurious feature learning for two-layer ReLU neural networks trained by online minibatch SGD on the logistic loss. We consider data drawn from the high-dimensional Boolean hypercube with a quadratic signal function (namely XOR) and a linear spurious correlation. We show that SGD learns the spurious feature first, and exponentially fast. Moreover, the optimization dynamics couple the spurious and signal features, with a stronger spurious component inhibiting signal feature learning. Our analysis reveals precise phase transitions in the learning dynamics. In the first phase,
alignment between the signs of the spurious feature and second-layer weight drives rapid growth of the spurious feature. In the second phase, large majority group margin slows learning and the signal feature remains suppressed. When the spurious correlation is maximally strong, we show theoretically that the spurious feature dominates even at the sample complexity threshold where XOR would be learned in isolation (i.e., if the spurious feature was absent). In contrast, when the correlation strength is constant, we provide preliminary empirical evidence that the model can eventually learn the XOR signal, although the spurious feature is not forgotten.
Read the full abstract ▾
Machine Learning
Transformer Architectures as Complete Bayes Processes: A Formal Proof in the Measure-Theoretic Kernel Framework
Haobo Yang
We present a complete formal proof that
transformer architectures, when their internal update mechanisms satisfy a Bayes joint-distribution condition, implement exact
Bayesian posterior inference. Working within the measure-theoretic kernel framework, we define a hierarchy of abstractions – from the core
Bayesian transformer, through semantic
transformers with explicit update kernels, to full
transformer blocks with QKV/
attention/residual/MLP pipelines, and finally multilayer stacks – and prove at each level that the Bayes joint semantics implies the update kernel equals the posterior almost everywhere. For the block-level architecture, we derive the explicit Bayes formula through Radon-Nikodym differentiation and prove its normalization. We additionally prove that the softmax
attention mechanism induces a valid probability distribution over keys, establishing the bridge between the abstract kernel framework and concrete
attention implementations. The framework makes no architectural assumptions beyond the Markov kernel structure and exposes explicit conditions under which a
transformer block is provably
Bayesian. In essence, when this joint distribution condition is satisfied, the forward computation of a
Transformer is formally equivalent to a rigorous
Bayesian posterior update.
Read the full abstract ▾
Machine Learning · Cross-listed from cs.CR
CAN We Trust Your Results? A Cross-Dataset Study of Automotive IDS Evaluation
Beatrix Koltai, Gergely Acs and Andras Gazdag
The increasing connectivity of modern vehicles has made securing in-vehicle communication networks a critical challenge. Intrusion Detection Systems (IDS) have been widely studied as a defense mechanism for detecting malicious activities on the Controller Area Network (CAN) bus. However, the evaluation of CAN IDS methods remains difficult due to inconsistencies in experimental setups and the lack of standardized benchmarking frameworks. As a result, reported performance often depends on
dataset-specific characteristics and may not reflect how detection methods behave in different environments. This work introduces a benchmarking framework for consistent evaluation of CAN IDSs across multiple
datasets. Using the proposed framework, we integrate seven publicly available CAN IDS
datasets collected under different experimental conditions and perform cross-
dataset evaluation of five conceptually different IDS approaches. Our results highlight how detection performance can vary significantly across
datasets, demonstrating the importance of cross-
dataset benchmarking for assessing the robustness and generalization capabilities of CAN IDS methods.
Read the full abstract ▾
Machine Learning
Arko-T: A Foundation Model for Text-to-Structured 3D Generation
Liang Wang, Zhaoyang Xi, Zekai Xiang et al. (+5 more)
Text-to-3D systems can now synthesize a model from a single sentence, yet the result is a shape to render, not a design to edit. We present Arko-T, a 4B-parameter text-to-design model that maps natural-
language intent directly into executable, parametric CAD programs. Rather than optimizing for code executability alone, Arko-T aligns every stage of the pipeline to a formal notion of design state, so that data curation, code normalization, and execution-grounded supervision all work to preserve the features, parameters, and construction logic that make a CAD artifact editable. Benchmarked against seven frontier LLMs across 12 metrics, Arko-T attains the best score on 8 and the second-best on 3 more, at roughly one-tenth the per-
benchmark cost. The results suggest that targeted design-level training at moderate scale can match frontier general-purpose models on structured CAD generation.
Read the full abstract ▾
All authors Liang Wang, Zhaoyang Xi, Zekai Xiang, Heng Meng, Qishan Zhang, Pingyi Zhou, Jin Liu, Litao Chen
Machine Learning
Proofs of Ownership for Machine Learning Models
Ran Canetti, Shafi Goldwasser and Or Zamir
With the increasing adoption of Machine Learning, protecting model ownership has become an essential challenge. We initiate a formal study of Proof of Ownership for machine learning models: under what conditions can one prove that a stolen model originated from a particular creator? We model proofs of ownership as a game among three parties: a model owner, a thief, and a judge. The owner transforms the original model into a slightly perturbed model together with a proof of ownership. The thief then obtains the transformed model and attempts to minimally modify it so that it remains useful but escapes detection as owned by the model owner. Finally, the judge receives a model and a proof of ownership, and must decide whether the given model is a modified version of some model created by the model owner, or else the given model was developed independently. Our main result is a dichotomy for classifiers in the black-box setting: Under standard cryptographic assumptions, ownership of models for some concept class can be proven in the above sense if and only if the concept class is not self-correctable, in a sense close to that of Blum, Luby and Rubinfeld, STOC'90. The result is constructive and extends, with some variations, to a number of related settings.
Read the full abstract ▾
Machine Learning
Experience Augmented Policy Optimization for LLM Reasoning
Jinda Lu, Kexin Huang, Junkang Wu et al. (+7 more)
Reinforcement Learning with Verifiable Rewards (RLVR) is a powerful paradigm for improving the
reasoning capabilities of large
language models (LLMs). However, existing RLVR methods typically rely on on-policy optimization from scratch, resulting in high sampling costs and inefficient utilization of accumulated experience. As model capabilities and policy behaviors evolve during training, recent attempts to reuse experience via fixed
reasoning trajectories further suffer from policy mismatch. Motivated by these limitations, we argue that experience in RLVR should not be reused as fixed
reasoning trajectories, but instead expressed in a policy-adaptive manner. In this work, we propose Experience-Augmented Policy Optimization (EAPO), which leverages a prior RL-optimized policy as an action-level experience prior and selectively injects experience at critical decision points during rollout. To ensure stable and unbiased learning from experience-augmented rollouts, EAPO further incorporates an adapted importance sampling scheme. Experiments on using Qwen-2.5-math 7b and Qwen-3-8B on five different
benchmarks demonstrate that EAPO consistently improves
reasoning performance over state-of-the-art RLVR methods.
Read the full abstract ▾
All authors Jinda Lu, Kexin Huang, Junkang Wu, Shuo Yang, Jinghan Li, Chiyu Ma, Shaohang Wei, Xiang Wang, Guoyin Wang, Jingren Zhou
Machine Learning
Diffusion Fine-tuning with Rewarded Moment Matching Distillation
Alexis Jacq, Guillaume Couairon, Valentin De Bortoli et al. (+3 more)
Distillation and Reinforcement Learning (RL) fine-tuning are the primary pillars of
diffusion post-training. While traditionally studied in isolation, the interaction between these phases remains poorly understood, and in particular how fine-tuning impacts the generative quality of distilled models. We introduce Rewarded Moment Matching
Distillation (RMMD), a novel framework that simultaneously distills
diffusion models and maximizes a reward function. RMMD preserves the high-fidelity “naturalness” characteristic of advanced
distillation (such as 8-step Moment Matching) by adapting the sampling loop for on-policy training and repurposing the
distillation loss as a proxy for integral KL regularization. By evaluating the FID-Reward Pareto fronts on ImageNet, we demonstrate that RMMD achieves superior trade-offs compared to single-step baselines (DI++) and multi-step competitors (DRaFT, HyperNoise). Finally, we apply RMMD to GenCast, a state-of-the-art weather
forecasting model, to distill it while optimizing the Continuous Ranked Probability Score (CRPS) metric. The resulting distilled model achieves a 7.5x speedup while outperforming the teacher model on 93% of target weather variables, and being better calibrated. This proves that RMMD scales to complex, high-dimensional scientific domains.
Read the full abstract ▾
All authors Alexis Jacq, Guillaume Couairon, Valentin De Bortoli, Quentin Berthet, Arnaud Doucet, Romuald Elie
Machine Learning
Beyond IID: How General Are Tabular Foundation Models, Really?
Lennart Purucker, Andrej Tschalzev, Nick Erickson et al. (+7 more)
Foundation models for predictive machine learning on tabular data have recently gained significant traction in academia and industry. Research communities across disciplines are increasingly evaluating tabular foundation models on diverse
datasets and tasks. However, these task- and discipline-specific evaluations remain largely inaccessible to model researchers because
benchmark software and evaluation protocols are fragmented. As a result, model researchers rely on standard
benchmarks, which are mostly defined for tasks where tabular foundation models already excel. The most challenging scenarios are excluded, limiting meaningful progress in the field by focusing on marginal improvements on IID data rather than on broader, more demanding challenges. To overcome this, we introduce BeyondArena, the first unified holistic
benchmark for tabular data that supports diverse task types (IID, temporal, grouped), across sample size and feature dimensionality scales, with diverse feature types (with text, with high cardinality) from a broad range of disciplines. To enable unified benchmarking beyond standard
benchmarks, we introduce DataFoundry, a Python framework and metadata schema for curating tabular
datasets for predictive machine learning. Our results across 11 models and 142 curated
datasets show that current open-source tabular foundation models are competitive or excel on
datasets up to 100k rows, while traditional tree-based and deep learning models still dominate on grouped or large data. BeyondArena guides model research for the most demanding challenges in tabular data, enabling progress towards truly foundational tabular models.
Read the full abstract ▾
All authors Lennart Purucker, Andrej Tschalzev, Nick Erickson, Gioia Blayer, David Holzmüller, Alan Arazi, Alexander Pfefferle, Mustafa Tajjar, Gaël Varoquaux, Frank Hutter
Machine Learning · Cross-listed from cs.CL
MOPD: Multi-Teacher On-Policy Distillation for Capability Integration in LLM Post-Training
Wenhan Ma, Jianyu Wei, Liang Zhao et al. (+10 more)
Modern large
language models (LLMs) rely on reinforcement learning during post-training to push specific capabilities, yet integrating multiple capabilities into one model remains hard. Existing methods, such as Off-Policy Finetune and Mix-RL, are either inefficient or lose performance. In this work, we propose Multi-teacher On-Policy
Distillation (MOPD), a post-training paradigm for combining the capabilities of multiple domain RL teachers: we first run per-domain specialised RL to obtain a set of domain teachers, then distill these teachers into the student on its own rollouts. This eliminates exposure bias and provides a dense optimization signal. On Qwen3-30B-A3B, MOPD outperforms Mix-RL, Cascade RL, Off-Policy Finetune, and Param-Merge baselines, inheriting nearly all of each teacher's capability. MOPD also enables parallel, independent development of domain teachers, removing the cross-domain coupling typical of multi-domain post-training. MOPD has been deployed in the post-training of
MiMo-V2-Flash, an industrial-scale frontier model, demonstrating its practical value for capability integration in frontier-scale LLMs.
Read the full abstract ▾
All authors Wenhan Ma, Jianyu Wei, Liang Zhao, Hailin Zhang, Bangjun Xiao, Lei Li, Qibin Yang, Bofei Gao, Yudong Wang, Rang Li, Jinhao Dong, Zhifang Sui, Fuli Luo
Machine Learning · Cross-listed from cs.AI
ENC-ODE: Event-level Neurodegenerative Modeling in Continuous Time with Neural ODEs
Yujee Song, Seunghun Baek, Guorong Wu et al. (+1 more)
Accurately predicting the temporal
evolution of clinical biomarkers is crucial for the early diagnosis and management of neurodegenerative diseases such as Alzheimer's disease. However, this relies on longitudinal data to capture biomarker changes over time, which is often sparse and irregular due to the high cost, labor-intensive nature, and patient burden. To address these challenges, we propose ENC-ODE, an Event-level Neurodegenerative modeling in Continuous time with neural Ordinary Differential Equations. ENC-ODE predicts future biomarker
evolution by modeling clinical events through diagnosis-conditioned continuous dynamics. A target-conditioned
attention mechanism weights and aggregates event-level predictions for the target time and modality without history compression. Extensive experiments on Alzheimer's Disease Neuroimaging Initiative (ADNI)
dataset demonstrate that ENC-ODE outperforms representative sequence models while offering a scalable and neuroscientifically grounded solution for clinical support. The code is available at https://github.com/JardinDelSol/enc-ode.
Read the full abstract ▾
All authors Yujee Song, Seunghun Baek, Guorong Wu, Won Hwa Kim
Machine Learning
Predict, Reuse, and Repair: Accelerating Dynamic Sparse Attention for Long-Context LLM Decoding
Tianyu Wang, Gourav Rattihalli, Aditya Dhakal et al. (+4 more)
Dynamic sparse
attention (DSA) accelerates long-context LLM decoding by attending to only the top-K KV blocks relevant to each query, but it introduces a serialized selection-to-
attention dependency that emerges as a new latency bottleneck. We present PRR, a speculate-reuse-repair runtime that exploits temporal locality in DSA selections to predict likely blocks, speculate the
attention over them while selection is in flight, and incrementally repair missed blocks once the true selected set is known. PRR uses a lightweight EMA-based predictor, a profiling-guided speculation budget that keeps speculative work off the critical path, and a FlashAttention-based repair kernel that folds missed blocks into the partial
attention state using online-softmax statistics. Across long-context
benchmarks and representative DSA methods, PRR reduces per-
token decoding latency by up to 40% while preserving downstream task accuracy. Github: https://github.com/Tianyu9748/Incremental_FlashAttention
Read the full abstract ▾
All authors Tianyu Wang, Gourav Rattihalli, Aditya Dhakal, Junbo Li, Zhiwei Ren, Dejan Milojicic, Longfei Shangguan
Machine Learning · Cross-listed from stat.ML
Róisín Luo, Christian Gagné, Jonas Ngnawé et al. (+2 more)
Delayed generalization (\ie~grokking) refers to the phenomenon in which a neural network fits its training data early in training but only begins to generalize after a prolonged delay, often through an abrupt transition. Despite extensive empirical study, its underlying mechanism remains poorly understood. In this work, we first theoretically characterize a shell–core topological configuration of the reachable solution space induced by Adam's optimization dynamics with weight-shrinkage regularization, supported by empirical evidence. This optimization-induced topological configuration gives rise to grokking. In model's parameter space, random initialization solutions concentrate on a thin outer spherical shell, enclosing another spherical shell of memorization solutions, which in turn contains a core corresponding to the generalization solutions. Leveraging stopping-time theory, we then analyze the geometry of this topological configuration and the solution transition time at which optimization trajectories escape the memorization
manifold and first reach the boundary of the generalization
manifold. Our theoretical analysis derives grokking scaling laws for the learning rate, batch size, and $\ell_2$ regularization coefficient, which are further validated through experiments and shown to recover results from prior literature.
Read the full abstract ▾
All authors Róisín Luo, Christian Gagné, Jonas Ngnawé, Ihsan Ullah, Karyn Morrissey
Machine Learning
Scalar Representations of Neural Network Training Dynamics
Pedro Jiménez-González, Miguel C. Soriano and Lucas Lacasa
Training in artificial neural networks can be viewed as a trajectory evolving through a high-dimensional loss landscape. However, the large number of trainable parameters makes the direct analysis of these dynamics challenging. In this work, we treat such training trajectories as temporal networks and apply recently proposed strategies for the scalar
embedding of temporal networks. We investigate whether such a scalar
embedding provides a meaningful low-dimensional representation of neural network training dynamics. Using a multilayer perceptron trained on the MNIST classification task, we show that the
embedding preserves the main dynamical features observed in the original parameter space, including the emergence of sensitivity to initial conditions for specific learning rate regimes and an accurate reconstruction of the network's maximum Lyapunov exponent. We then use the embedded scalar trajectory to define a characteristic time, analogous to a Lyapunov time, after which the exponential separation between initially close embedded trajectories saturates. This characteristic time captures the typical decorrelation time between initially close network trajectories in the original high-dimensional system. Finally, we investigate the statistical organization of asymptotic training states through a spacing observable defined in the embedded space. We find that the distributions of rescaled asymptotic spacings collapse onto a common form across initial conditions and are compatible with a skew lognormal distribution. Altogether, our results suggest that scalar low-dimensional
embeddings provide a useful framework for studying and visualizing the dynamical properties of neural network optimization trajectories.
Read the full abstract ▾
Machine Learning · Cross-listed from cs.GR
RenderFormer++: Scalable and Physics-Informed Feed-Forward Neural Rendering
Huangsheng Du, Haoran Zhu, Youcheng Cai et al. (+2 more)
We present RenderFormer++, a scalable and physics-informed feed-forward neural rendering framework for global illumination in mesh scenes. Existing
Transformer-based neural rendering methods such as RenderFormer achieve promising cross-scene generalization, but lack explicit transport priors and scale poorly due to quadratic triangle-level
attention. To address these issues, we introduce Physics-Informed Transport Guidance (PITG), which embeds rendering-equation-inspired inductive biases into the
attention mechanism and introduces a transport consistency loss, encouraging physics-informed light transport modeling. We further propose Hierarchical Object-Centric Tokenization (HOCT), which aggregates triangle-level features into compact object-level
tokens via cross-
attention with learnable queries, substantially reducing computational and
memory costs. Extensive experiments demonstrate that RenderFormer++ achieves scalable and generalizable feed-forward global illumination rendering across complex large-scale scenes with competitive rendering quality and substantially improved efficiency over RenderFormer. The code will be made publicly available upon acceptance.
Read the full abstract ▾
All authors Huangsheng Du, Haoran Zhu, Youcheng Cai, Jingyang Meng, Ligang Liu
Machine Learning
FlowAWR: Online Adaptive Flow Reinforcement via Advantage-Weighted Rectification
Zheming Fu, Ruizhe He, Wei Shang et al. (+5 more)
Aligning generative flow models on continuous spaces via online reinforcement learning is constrained by intractable trajectory likelihoods. Existing density-approximated policy gradient methods rely on
stochastic SDE samplers to construct tractable transition kernels, which introduce training-inference inconsistencies and necessitates Classifier-Free Guidance (CFG). While implicit frameworks such as DiffusionNFT directly optimize forward-process velocity fields, its heuristic fixed-magnitude corrections prevent optimization strength from relative intra-group quality. We propose Flow Advantage-Weighted Rectification (FlowAWR), a paradigm that recasts continuous generative policy optimization as supervised regression toward a theoretically optimal velocity field. Starting from the optimal policy of a KL-constrained reward maximization, FlowAWR derives the optimal velocity field that admits a magnitude-aware, advantage-weighted rectification form, yielding SDE-free optimization and CFG-free generation. In comparative evaluations on SD3.5-Medium, FlowAWR achieves improved
alignment performance alongside a 2$\times$ to 5$\times$ convergence acceleration over DiffusionNFT (e.g., reaching a 24.12 PickScore in 1.2k steps, versus 23.82 in 2.0k steps for DiffusionNFT and 23.50 in $>$4k steps for FlowGRPO). Under multi-reward constraints, FlowAWR sustains generation quality, satisfying structural rules while maintaining stable out-of-domain performance. Code at \href{https://github.com/Fuzm7/FlowAWR}{FlowAWR}.
Read the full abstract ▾
All authors Zheming Fu, Ruizhe He, Wei Shang, Xiaoxiao Ma, Lianrui Mu, Lei Wang, Chang Liu, Siming Fu
Machine Learning · Cross-listed from cs.CV
Set-Inclusive Uncertainty Modeling for Robust Brain Tumor Segmentation
Seunghun Baek, Jihwan Park, Jaeyoon Sim et al. (+3 more)
Multimodal MRI is essential for accurate
brain tumor
segmentation. However, acquiring all modalities at inference is often challenging in practice, which causes intrinsic uncertainty due to unavoidable information loss. Without modeling this uncertainty, existing methods encode incomplete evidence into deterministic representations that appear plausible but lack reliability. In this regime, we propose a probabilistic representation framework that models representations as Gaussian distributions, where their mean captures task information and their variance measures uncertainty from missing evidence. To make variance reflect information deficiency, we regularize the mean from each partial configuration toward its full-modality counterpart, while scaling the variance with the discrepancy between their aligned means. We further introduce a set-inclusive strategy that exploits the hierarchical structure of modality subsets and enforces an ordering constraint to maintain their consistent uncertainty relationships. Extensive experiments on BraTS 2018 and 2020 demonstrate that our approach offers superior performance over baselines across diverse missing-modality scenarios. Code and model checkpoint are available at https://github.com/atlas-sky/SIUM.
Read the full abstract ▾
All authors Seunghun Baek, Jihwan Park, Jaeyoon Sim, Hoseok Lee, Seungjoo Lee, Won Hwa Kim
Machine Learning
On the Vulnerability of Parameter-Level Defenses to Model Merging
Kuangpu Guo, Qingyan Zheng, Jian Liang et al. (+4 more)
The training-free integration of expert models via model merging has exposed significant security risks, enabling free-riders to combine specialized models without authorization. Recent works propose parameter-level defenses that employ linear parameter transformations to neutralize this threat. In this paper, we systematically analyze such defenses and reveal that their protected task vectors are inherently small in magnitude. Consequently, the protected weights remain overwhelmingly dominated by the pretrained model. Based on this observation, we designate the pretrained model as a static reference anchor and propose the Anchor-Guided Attack (AGA) to circumvent existing safeguards. Specifically, AGA aligns the protected model with this anchor to recover the transformation
matrix analytically. Extensive evaluations validate that AGA consistently bypasses both individual and composite defenses under realistic defense-agnostic scenarios. Furthermore, we provide Anchor-Repulsive Fine-tuning (ARF), a defense method to mitigate the anchor dominance leveraged by AGA. Empirical results confirm that ARF effectively defeats the proposed attack. Our code is available at https://github.com/krumpguo/secure-merge-attack.
Read the full abstract ▾
All authors Kuangpu Guo, Qingyan Zheng, Jian Liang, Yongcan Yu, Zilei Wang, Ran He, Tieniu Tan
Machine Learning · Cross-listed from quant-ph
Learning the structure of open quantum systems
Laura Lewis, Ewin Tang and John Wright
We design an
algorithm for learning the coefficients of an $n$-
qubit constant-local Lindbladian to $\varepsilon$ error with $O(g d^2 \log(n) / \varepsilon^2)$ total
evolution time, where $g$ is the single-site energy and $d$ is the (approximate) degree of the interaction graph. Though Lindbladians present new challenges not present in the special case of
Hamiltonians, our
algorithm achieves the suite of desiderata attained by state-of-the-art
Hamiltonian learning
algorithms: (1) it uses non-adaptive, ancilla-free randomized Pauli measurement
circuits with a time resolution of only $\Theta(1/g)$; (2) it works without knowledge of the structure of the unknown Lindbladian; (3) it depends on a smooth form of degree, thereby supporting the learning of quasi-local and power-law Lindbladians. Moreover, we prove a lower bound showing that our
algorithm is optimal in each parameter up to logarithmic factors. Our
algorithm is a simple iterative method, where the objective function consists of
Fourier coefficients of the Lindbladian restricted to few-site regions. Its analysis identifies the difficulty unique to open systems, which we call "confusing" terms. For settings where the "confusion" is limited, the performance of the
algorithm improves. We demonstrate this for the case of structure learning of
Hamiltonians from access to real-time
evolution, where we obtain a new
algorithm that is significantly simpler than previous work. In addition, using the same iterative method, we design the first efficient
algorithm for structure learning
Hamiltonians from high-temperature Gibbs states.
Read the full abstract ▾
Machine Learning · Cross-listed from cs.CL
OLIVE: View-Augmented Latent Prediction with Waveform Reconstruction for Speech SSL
Karl El Hajal and Mathew Magimai.-Doss
We propose Online Latent prediction with Invariant Views and rEconstruction (OLIVE), a self-supervised
speech representation learning framework that jointly optimizes analysis and synthesis objectives. OLIVE couples masked latent prediction and waveform reconstruction through a functional separation within a single encoder: reconstruction constrains early encoder features to retain signal-level information, while masked latent prediction shapes later contextual representations for robust downstream performance, with waveform view augmentations defining the invariances they learn. To the best of our knowledge, OLIVE is the first
speech SSL model that directly reconstructs waveforms, retains its vocoder after pre-training, and remains competitive on discriminative tasks. Adding reconstruction largely preserves the
speech recognition performance of masked latent prediction alone, while improving waveform reconstruction and speaker tasks. Adding waveform mixup and gain augmentations provides a controllable way to shift the representations further toward speaker, generation, and reconstruction tasks, while remaining competitive on recognition and semantic tasks.
Read the full abstract ▾
Machine Learning
DRIFT: Difficulty Routing Self-DIstillation with Rhythm-Gated Exploration and Success BuFfer Training
Haisen Luo, Yiwei Liu, Haoning Wang et al. (+13 more)
Enabling large
language models to achieve stable self-improvement without external expert supervision remains a central challenge in complex
reasoning tasks. Existing self-
distillation and reinforcement learning methods lack explicit mechanisms for tracking problem-level learning progress and adapting optimization strategies accordingly. Consequently, training may over-optimize easy problems, receive weak supervision from hard problems, and fail to sufficiently explore borderline cases. To resolve these issues, we propose DRIFT, an online self-
evolution policy optimization framework for large
language models. DRIFT regulates the model's self-improvement process through the joint use of Difficulty Routing and Rhythm Gating. The former identifies the model's learning state at the problem level and dynamically allocates self-
distillation and reinforcement learning signals, while the latter refines policy updates at the
token level, concentrating exploration on critical
reasoning positions. By further incorporating a success buffer and a two-stage curriculum learning strategy, DRIFT preserves high-quality historical experience while progressively guiding the model from reliable behavior acquisition toward stable policy
evolution. Evaluated across five
benchmarks and three model scales, DRIFT surpasses the peak performance of both GRPO and SDPO across all evaluated metrics. On the average score over the five
benchmarks, DRIFT achieves 79.5$\%$, outperforming GRPO by 9.5$\%$ and SDPO by 7.5$\%$, establishing a new state-of-the-art result. Notably, on ToolUse, DRIFT reaches an accuracy of 79.2$\%$, improving over GRPO by 13.5$\%$ and SDPO by 10.7$\%$, setting a new state-of-the-art and substantially outperforming all concurrent methods.
Read the full abstract ▾
All authors Haisen Luo, Yiwei Liu, Haoning Wang, Dan Liu, Junxi Yin, Haotian Wang, Lei Zhang, Xiaoyu Tian, Shuaiting Chen, Yuansheng Song, Baoyan Guo, Xiongfei Yan, Bolan Yang, Chengwei Liu, Ming Cui, Jiong Chen
Machine Learning · Cross-listed from cs.CL
REAR: Test-time Preference Realignment through Reward Decomposition
Fuxiang Zhang, Pengcheng Wang, Chenran Li et al. (+6 more)
Aligning large
language models (LLMs) with diverse user preferences is a critical yet challenging task. While post-training methods can adapt models to specific needs, they often require costly data curation and additional training. Test-time scaling (TTS) presents an efficient, training-free alternative, but its application has been largely limited to verifiable domains like mathematics and coding, where response correctness is easily judged. To extend TTS to preference
alignment, we introduce a novel framework that models the task as a realignment problem, since the base model often fails to sufficiently align with the stated preference. Our key insight is to decompose the underlying reward function into two components: one related to the question and the other to preference information. This allows us to derive a REAlignment Reward (REAR) that selectively rescales the proportions of these two reward terms. We then show that REAR can be formulated as a linear combination of
token-level policy log-probabilities, making it computationally efficient and easy to integrate with various TTS
algorithms such as best-of-$N$ sampling and tree search. Experiments show that compared to other test-time baselines, REAR not only enables scalable test-time realignment for preference
alignment tasks under diverse user requirements, but also generalizes to mathematical and visual tasks under appropriate preference settings.
Read the full abstract ▾
All authors Fuxiang Zhang, Pengcheng Wang, Chenran Li, Yi-Chen Li, Yuxin Chen, Lang Feng, Chenfeng Xu, Masayoshi Tomizuka, Bo An
Machine Learning
FlexTab: A Flexible Encoder-Decoder Architecture for In-Context Learning Across Diverse Tabular Tasks
Marek Polewczyk, Maximilian Schambach, Marco Spinaci et al. (+2 more)
We introduce FlexTab, a flexible encoder-decoder architecture for in-context learning on tabular data that pairs a single, task-agnostic encoder with a suite of task-specific decoders. Unlike existing tabular in-context learners, which entangle feature representations with a specific prediction target, our design produces target-agnostic row
embeddings that can be leveraged across a wide range of downstream tasks within a table-native in-context learning setup. We demonstrate this flexibility on seven distinct problems: classification, regression, anomaly detection, clustering, entity matching,
retrieval, and entity classification in relational databases. Both the encoder and the task-specific decoders are trained on tasks generated from a large corpus of real-world, unlabeled tables. FlexTab achieves state-of-the-art performance on classification, regression, anomaly detection, entity matching, and
retrieval, while remaining competitive with specialized models on entity classification in a relational setting. These results demonstrate that a single shared encoder, paired with task-specific decoders, can serve as an effective general-purpose backbone for diverse tabular prediction problems. The inference code and checkpoints are publicly available at: https://github.com/SAP-samples/flextab .
Read the full abstract ▾
All authors Marek Polewczyk, Maximilian Schambach, Marco Spinaci, Sam Thelin, Johannes Höhne
Machine Learning · Cross-listed from math.OC
Local-Minima-Preserving Continuous Relaxation of Ising Problems
Debraj Banerjee, Santanu Mahapatra and Kunal N. Chaudhury
The generalized Ising problem captures a broad spectrum of hard combinatorial problems, including MAX-CUT, Number Partitioning (NPP), and Maximum Independent Set. In this work, we consider the notion of one-flip local minima for this problem. We construct a
polynomial relaxation and prove the landscape equivalence
theorem: there exists a one-to-one correspondence between the local minima of the relaxation and the one-flip minima of the original Ising problem. This guarantee reduces the Ising problem to finding the local minima of a smooth function, allowing us to leverage gradient-based optimizers such as ADAM. We demonstrate that our method is scalable and it achieves strong performance across challenging
benchmarks, including
spin-glass models, MAX-CUT, and NPP.
Read the full abstract ▾
Machine Learning · Cross-listed from stat.ML
Extrapolating from Regularised Solutions for Solving Ill-Conditioned Linear Systems in Machine Learning
Disha Hegde, Jon Cockayne and Chris. J. Oates
Rapid prototyping of
algorithms is a critical step in modern machine learning. Most
algorithms exploit linear
algebra, creating a need for lightweight numerical routines which – while potentially sub-optimal for the task at hand – can be rapidly implemented. For the numerical solution of ill-conditioned linear systems of equations, the standard solution for prototyping is Tikhonov-regularised inversion using a nugget. However, selection of the size of nugget is often difficult, and the use of data-adaptive procedures precludes automatic differentiation, introducing instabilities into end-to-end training. Further, while data-adaptive procedures perform multiple linear solves to select the size of nugget, only the result of one such solve is returned, which we argue is wasteful. This paper aims to circumvent the above difficulties, presenting autonugget; a Python package for automatic and stable numerical solution of linear systems suitable for rapid prototyping, and fully compatible with automatic differentiation using JAX. autonugget combines multiple linear solves using Richardson extrapolation to determine the solution of the ill-conditioned system, improving in accuracy over approximations based on a single nugget.
Read the full abstract ▾
Machine Learning
Hybrid Active-Online Learning Framework for Label-Efficient Concept Drift Adaptation in Optical Network Failure Detection
Yousuf Moiz Ali, Jaroslaw E. Prilepsky, João Pedro et al. (+4 more)
We propose a hybrid active-online learning framework for label-efficient concept drift adaptation in optical network failure detection. Using margin-based selective labeling, our method achieves nearceiling accuracy and AUC scores while querying only 3.4% of streaming samples, with negligible latency overhead compared to static inference.
Read the full abstract ▾
All authors Yousuf Moiz Ali, Jaroslaw E. Prilepsky, João Pedro, Sasipim Srivallapanondh, Antonio Napoli, Sergei K. Turitsyn, Pedro Freire
Machine Learning · Cross-listed from cs.CV
BrainJanus: A Unified Model for Understanding and Generation across Brain, Vision, and Language
Haitao Wu, Qirui Zhang, Zhouheng Yao et al. (+8 more)
Modeling the bidirectional correspondence between external sensory stimuli and internal neural activity has emerged as a critical frontier in neuroscience. However, existing approaches predominantly treat
brain encoding and decoding as isolated tasks, relying heavily on unimodal
alignment and external priors while overlooking the
brain's intrinsic nature as a
multimodal integration system. To address these limitations, we propose BrainJanus, the first unified
brain model that integrates
brain,
vision, and
language within a single framework. Specifically, we introduce a Unified
Brain Tokenizer to quantize continuous neural dynamics into discrete
tokens aligned with visual and linguistic representations in a shared Omni space. Building on this, we utilize an All-in-One autoregressive architecture that leverages next-
token prediction to enable seamless any-to-any generation, which encompasses image-to-
brain and text-to-
brain encoding, and
brain-to-image and
brain-to-text decoding. Extensive experiments demonstrate that BrainJanus achieves superior performance across diverse
benchmarks. Furthermore, our framework exhibits zero-shot generalization and preserves interpretable biological topography, highlighting its potential as a general-purpose
brain modeling paradigm. The code is available at \href{https://github.com/HaitaoWuTJU/BrainJanus}{GitHub}.
Read the full abstract ▾
All authors Haitao Wu, Qirui Zhang, Zhouheng Yao, Shangquan Sun, Qihao Zheng, Mianxin Liu, Chi Zhang, Wanli Ouyang, Chunfeng Song, Changqing Zhang, Jiamin Wu
Machine Learning
Toward an Energy-Optimized Operation of Data Centers Located in Wind Farms Using Reinforcement Learning
Jan Stenner, Alexander Kilian, Sebastian Peitz et al. (+1 more)
This paper studies Reinforcement Learning as an online controller for curtailment-aware workload shifting in wind-turbine-integrated high-performance computing (HPC) data centers. We introduce a reproducible fixed-day simulation framework with synthetic wind and price signals and delayed completion feedback, designed to be extensible toward more complex scenarios. As a controlled benchmarking basis, we then focus on the minimal case with one wind turbine and one co-located data center. In this setting, pure Reinforcement Learning exhibits a pronounced credit-assignment problem and tends to underuse free wind energy early in the day. We therefore evaluate two complementary countermeasures: optimization-based Imitation Learning and potential-based Reward Shaping. Across multi-seed training and a 200-day test set, Proximal Policy Optimization (PPO) and a Soft Actor-Critic (SAC) variant with an additional on-policy update routine achieve strong empirical performance among learned policies, and both Imitation Learning and Reward Shaping provide improvements in relevant configurations. A performance gap to the optimizer remains, which is expected: the optimizer plans offline with full-day foresight, whereas Reinforcement Learning must decide online from current observations without future realizations. The
benchmark and ablation results provide a transparent basis for extending the approach toward richer multi-site and continuous-time scenarios.
Read the full abstract ▾
All authors Jan Stenner, Alexander Kilian, Sebastian Peitz, Hermann de Meer
Machine Learning · Cross-listed from cs.CV
TRACE: A Concept Bottleneck Model for Longitudinal 3D Glioblastoma Response Assessment
Alia Tarek, Hamsa Saberr, Hamza Elghonemy et al. (+6 more)
Longitudinal glioblastoma response assessment requires comparing subtle tumor changes across MRI time points using structured clinical criteria such as RANO. However, most deep learning methods predict response labels directly from imaging features, which limits clinical inspection, verification, and correction. We introduce TRACE, a RANO 2.0-aligned concept bottleneck model for interpretable 4-class glioblastoma response classification on longitudinal 3D MRI. TRACE processes paired baseline and follow-up
multimodal MRI scans with a shared 3D
vision encoder, predicts clinically meaningful tumor measurements as root concepts, computes downstream RANO-derived concepts through deterministic rules, and incorporates scan interval and new-lesion information as passthrough concepts. This design frames response assessment as structured concept
reasoning rather than direct image-to-label prediction. Using 5-fold patient-wise cross-validation on the LUMIERE
dataset, TRACE achieves a 4-class macro F1 of 0.4769 and a binary progression-versus-non-progression macro F1 of 0.7085. It improves over a concept bottleneck baseline and remains within the range of published non-interpretable deep learning approaches. Ablation studies show that the expert RANO graph and intervention-consistency training are important for performance, while intervention experiments demonstrate that correcting concepts can improve downstream predictions. These results suggest that structured concept bottlenecks offer a transparent and clinically aligned direction for longitudinal glioblastoma response assessment, while highlighting the need for larger protocol-aligned
datasets and external validation.
Read the full abstract ▾
All authors Alia Tarek, Hamsa Saberr, Hamza Elghonemy, Youssef Afify, Tamer Basha, Omair Shahzad Bhatti, Abdulrahman M. Selim, Hasan Md Tusfiqur Alam, Daniel Sonntag
Machine Learning · Cross-listed from stat.ML
Highly Data Parallelizable Estimation of the Sliced-Wasserstein Distance Using Cumulative Distribution Functions
Christophe Vauthier, Quentin Mérigot and Anna Korba
The Sliced Wasserstein (SW) distance has emerged as a computationally attractive alternative to the Wasserstein distance by leveraging one-dimensional optimal transport along random projections. Standard
estimators of the SW distance rely on Monte Carlo averages of one-dimensional Wasserstein distances computed via
quantile functions, which require sorting projected samples and access to full
datasets. In this work, we introduce a new class of
estimators for the Sliced Wasserstein distance based on cumulative distribution functions (CDFs) of projected measures, that avoid sorting and scale via massive
dataset parallelism. This class includes several
estimators, some of them being indexed by hyperparameters controlling their variance or smoothness. We show that they are especially well suited to scenarios in which CDFs are more tractable than
quantile functions, such as mixtures of Gaussians, and moreover that they are also naturally compatible with
federated learning, since CDFs of projected data can be computed and aggregated locally without requiring the exchange of raw samples.
Read the full abstract ▾
Machine Learning
DreamForge-World 0.1 Preview: A Low-Compute Real-Time Controllable World Model
Daniyel Ayupov and Artur Markov-Tsoy
We present DreamForge-World 0.1 Preview, a preview foundational world model for real-time interactive world simulation. The system adapts the LongLive 1 autoregressive
video stack, itself derived from Wan2.1-T2V-1.3B, with a residual action pathway inspired by the
Matrix-Game family. DreamForge-World 0.1 Preview focuses on a complementary axis to frontier-scale world simulators: low-compute adaptation, consumer-GPU runtime, and broad interactive capability coverage. It supports live keyboard and mouse control,
multimodal initialization, mid-stream reprompting, dual-view operation, and minute-scale interactive rollouts at native 480p resolution, reaching up to 14 to 15 FPS FPS on a single RTX 4090 with a low
memory footprint. By leveraging open
video backbones and applying targeted adaptation runs, we build the preview system with high cost-efficiency. DF-World 0.1 Preview is not yet a
memory-complete or frontier-quality world simulator, but demonstrates a practical low-compute route toward real-time controllable world-model previews on consumer GPUs.
Read the full abstract ▾