Peer Revenue Fully funded

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Computer Science Machine Learning

cs.LG Watch Machine Learning

Tuesday 21 April

Machine Learning · Cross-listed from cs.CL

From Signal Degradation to Computation Collapse: Uncovering the Two Failure Modes of LLM Quantization

Chenxi Zhou, Pengfei Cao, Jiang Li et al. (+4 more)

Post-Training Quantization (PTQ) is critical for the efficient deployment of Large Language Models (LLMs). While 4-bit quantization is widely regarded as an optimal trade-off, reducing the precision to 2-bit usually triggers a catastrophic “performance cliff.” It remains unclear whether the underlying mechanisms differ fundamentally. Consequently, we conduct a systematic mechanistic analysis, revealing two qualitatively distinct failure modes: Signal Degradation, where the computational patterns remain intact but information precision is impaired by cumulative error; and Computation Collapse, where key components fail to function, preventing correct information processing and destroying the signal in the early layers. Guided by this diagnosis, we conduct mechanism-aware interventions, demonstrating that targeted, training-free repair can mitigate Signal Degradation, but remains ineffective for Computation Collapse. Our findings provide a systematic diagnostic framework for PTQ failures and suggest that addressing Computation Collapse requires structural reconstruction rather than mere compensation.
Read the full abstract

All authors Chenxi Zhou, Pengfei Cao, Jiang Li, Bohan Yu, Jinyu Ye, Jun Zhao, Kang Liu

Machine Learning

Super Apriel: One Checkpoint, Many Speeds

SLAM Labs: Oleksiy Ostapenko, Raymond Li, Torsten Scholak et al. (+11 more)

We release Super Apriel, a 15B-parameter supernet in which every decoder layer provides four trained mixer choices – Full Attention (FA), Sliding Window Attention (SWA), Kimi Delta Attention (KDA), and Gated DeltaNet (GDN). A placement selects one mixer per layer; placements can be switched between requests at serving time without reloading weights, enabling multiple speed presets from a single checkpoint. The shared checkpoint also enables speculative decoding without a separate draft model. The all-FA preset matches the Apriel 1.6 teacher on all reported benchmarks; recommended hybrid presets span $2.9\times$ to $10.7\times$ decode throughput at 96% to 77% quality retention, with throughput advantages that compound at longer context lengths. With four mixer types across 48 layers, the configuration space is vast. A surrogate that predicts placement quality from the per-layer mixer assignment makes the speed-quality landscape tractable and identifies the best tradeoffs at each speed level. We investigate whether the best configurations at each speed level can be identified early in training or only after convergence. Rankings stabilize quickly at 0.5B scale, but the most efficient configurations exhibit higher instability at 15B, cautioning against extrapolation from smaller models. Super Apriel is trained by stochastic distillation from a frozen Apriel 1.6 teacher, followed by supervised fine-tuning. We release the supernet weights, Fast-LLM training code, vLLM serving code, and a placement optimization toolkit.
Read the full abstract

All authors SLAM Labs: Oleksiy Ostapenko, Raymond Li, Torsten Scholak, Alireza Mousavi-Hosseini, Aman Tiwari, Denis Kocetkov, Joel Lamy Poirier, Kelechi Ogueji, Nanda H Krishna, Rafael Pardinas, Sathwik Tejaswi Madhusudhan, Shruthan Radhakrishna, Srinivas Sunkara, Valerie Becaert

2604.19877 · 21 Apr · Models: https://huggingface.co/ServiceNow-AI/SuperApriel-15B-Base and https://huggingface.co/ServiceNow-AI/SuperApriel-15B-Instruct . Dev model: https://huggingface.co/ServiceNow-AI/SuperApriel-0.5B-Base . Training code: https://github.com/ServiceNow/Fast-LLM . Async RL: https://github.com/ServiceNow/pipeline-rl . Training logs: https://wandb.ai/servicenow-team/Super_Apriel

Abstract PDF $ATTN ×3$BNCH ×1$DSTL ×1$STOC ×1

Machine Learning

DR-Venus: Towards Frontier Edge-Scale Deep Research Agents with Only 10K Open Data

Venus Team, Sunhao Dai, Yong Deng et al. (+10 more)

Edge-scale deep research agents based on small language models are attractive for real-world deployment due to their advantages in cost, latency, and privacy. In this work, we study how to train a strong small deep research agent under limited open-data by improving both data quality and data utilization. We present DR-Venus, a frontier 4B deep research agent for edge-scale deployment, built entirely on open data. Our training recipe consists of two stages. In the first stage, we use agentic supervised fine-tuning (SFT) to establish basic agentic capability, combining strict data cleaning with resampling of long-horizon trajectories to improve data quality and utilization. In the second stage, we apply agentic reinforcement learning (RL) to further improve execution reliability on long-horizon deep research tasks. To make RL effective for small agents in this setting, we build on IGPO and design turn-level rewards based on information gain and format-aware regularization, thereby enhancing supervision density and turn-level credit assignment. Built entirely on roughly 10K open-data, DR-Venus-4B significantly outperforms prior agentic models under 9B parameters on multiple deep research benchmarks, while also narrowing the gap to much larger 30B-class systems. Our further analysis shows that 4B agents already possess surprisingly strong performance potential, highlighting both the deployment promise of small models and the value of test-time scaling in this setting. We release our models, code, and key recipes to support reproducible research on edge-scale deep research agents.
Read the full abstract

All authors Venus Team, Sunhao Dai, Yong Deng, Jinzhen Lin, Yusheng Song, Guoqing Wang, Xiaofeng Wu, Yuqi Zhou, Shuo Yang, Zhenzhe Ying, Zhanwei Zhang, Changhua Meng, Weiqiang Wang

Machine Learning

Rethinking Reinforcement Fine-Tuning in LVLM: Convergence, Reward Decomposition, and Generalization

Carter Adams, Rafael Oliveira, Gabriel Almeida et al. (+1 more)

Reinforcement fine-tuning with verifiable rewards (RLVR) has emerged as a powerful paradigm for equipping large vision-language models (LVLMs) with agentic capabilities such as tool use and multi-step reasoning. Despite striking empirical successes, most notably Visual Agentic Reinforcement Fine-Tuning (Visual-ARFT), the theoretical underpinnings of this paradigm remain poorly understood. In particular, two critical questions lack rigorous answers: (i)~how does the composite structure of verifiable rewards (format compliance, answer accuracy, tool executability) affect the convergence of Group Relative Policy Optimization (GRPO), and (ii)~why does training on a small set of tool-augmented tasks transfer to out-of-distribution domains? We address these gaps by introducing the Tool-Augmented Markov Decision Process (TA-MDP), a formal framework that models multimodal agentic decision-making with bounded-depth tool calls. Within this framework, we establish three main results. First, we prove that GRPO under composite verifiable rewards converges to a first-order stationary point at rate $O(1/\sqrt{T})$ with explicit dependence on the number of reward components and group size (Theorem~1). Second, we derive a Reward Decomposition Theorem that bounds the sub-optimality gap between decomposed per-component optimization and joint optimization, providing a precise characterization of when reward decomposition is beneficial (Theorem~2). Third, we establish a PAC-Bayes generalization bound for tool-augmented policies that explains the strong out-of-distribution transfer observed in Visual-ARFT (Theorem~3).
Read the full abstract

All authors Carter Adams, Rafael Oliveira, Gabriel Almeida, Sofia Torres

Machine Learning · Cross-listed from cs.AR

ChipCraftBrain: Validation-First RTL Generation via Multi-Agent Orchestration

Cagri Eryilmaz

Large Language Models (LLMs) show promise for generating Register-Transfer Level (RTL) code from natural language specifications, but single-shot generation achieves only 60-65% functional correctness on standard benchmarks. Multi-agent approaches such as MAGE reach 95.9% on VerilogEval yet remain untested on harder industrial benchmarks such as NVIDIA's CVDP, lack synthesis awareness, and incur high API costs. We present ChipCraftBrain, a framework combining symbolic-neural reasoning with adaptive multi-agent orchestration for automated RTL generation. Four innovations drive the system: (1) adaptive orchestration over six specialized agents via a PPO policy over a 168-dim state (an alternative world-model MPC planner is also evaluated); (2) a hybrid symbolic-neural architecture that solves K-map and truth-table problems algorithmically while specialized agents handle waveform timing and general RTL; (3) knowledge-augmented generation from a 321-pattern base plus 971 open-source reference implementations with focus-aware retrieval; and (4) hierarchical specification decomposition into dependency-ordered sub-modules with interface synchronization. On VerilogEval-Human, ChipCraftBrain achieves 97.2% mean pass@1 (range 96.15-98.72% across 7 runs, best 154/156), on par with ChipAgents (97.4%, self-reported) and ahead of MAGE (95.9%). On a 302-problem non-agentic subset of CVDP spanning five task categories, we reach 94.7% mean pass@1 (286/302, averaged over 3 runs), a 36-60 percentage-point lift per category over the published single-shot baseline; we additionally lead three of four categories shared with NVIDIA's ACE-RTL despite using roughly 30x fewer per-problem attempts. A RISC-V SoC case study demonstrates hierarchical decomposition generating 8/8 lint-passing modules (689 LOC) validated on FPGA, where monolithic generation fails entirely.
Read the full abstract

Machine Learning · Cross-listed from hep-ex

Neural posterior estimation of the neutrino direction in IceCube using transformer-encoded normalizing flows on the sphere

R. Abbasi, M. Ackermann, J. Adams et al. (+411 more)

IceCube is a cubic-kilometer-scale neutrino detector located at the geographic South Pole. A precise directional reconstruction of IceCube neutrinos is vital for associations with astronomical objects. In this context, we discuss neural posterior estimation of the neutrino direction via a transformer encoder that maps to a normalizing flow on the 2-sphere. It achieves a new state-of-the-art angular resolution for the two main event morphologies in IceCube - tracks and showers - while being significantly faster than traditional B-spline-based likelihood reconstructions. All-sky scans can be performed within seconds rather than hours, and take constant computation time, regardless of whether the posterior extent is arc-minutes or spans the whole sky. We utilize a combination of $C^2$-smooth rational-quadratic splines, scale transformations and rotations to define a novel spherical normalizing-flow distribution whose parameters are predicted as a whole as the output of the transformer encoder. We test several structural choices diverting from the vanilla transformer architecture. In particular, we find dual residual streams, nonlinear QKV projection and a separate class token with its own cross-attention processing to boost test-time performance. The angular resolution for both showers and tracks improves substantially over the whole trained energy range from 100 GeV to 100 PeV. At 100 TeV deposited energy, for example, the median angular resolution improves by a factor of $1.3$ for throughgoing tracks, by a factor of $1.7$ for showers and by a factor of $2.5$ for starting tracks compared to state-of-the art likelihood reconstructions based on B-splines. While previous machine-learning (ML) efforts have managed to obtain competitive shower resolutions, this is the first time an ML-based method outperforms likelihood-based muon reconstructions above 100 GeV.
Read the full abstract

All authors R. Abbasi, M. Ackermann, J. Adams, J. A. Aguilar, M. Ahlers, J.M. Alameddine, S. Ali, N. M. Amin, K. Andeen, C. Argüelles, Y. Ashida, S. Athanasiadou, S. N. Axani, R. Babu, X. Bai, A. Balagopal V., S. W. Barwick, V. Basu, R. Bay, J. J. Beatty, J. Becker Tjus, P. Behrens, J. Beise, C. Bellenghi, S. Benkel, S. BenZvi, D. Berley, E. Bernardini, D. Z. Besson, E. Blaufuss, L. Bloom, S. Blot, F. Bontempo, J. Y. Book Motzkin, C. Boscolo Meneguolo, S. Böser, O. Botner, J. Böttcher, J. Braun, B. Brinson, Z. Brisson-Tsavoussis, R. T. Burley, D. Butterfield, K. Carloni, J. Carpio, N. Chau, Z. Chen, D. Chirkin, S. Choi, A. Chubarov, B. A. Clark, G. H. Collin, D. A. Coloma Borja, A. Connolly, J. M. Conrad, D. F. Cowen, C. De Clercq, J. J. DeLaunay, D. Delgado, T. Delmeulle, S. Deng, P. Desiati, K. D. de Vries, G. de Wasseige, T. DeYoung, J. C. Díaz-Vélez, S. DiKerby, T. Ding, M. Dittmer, A. Domi, L. Draper, L. Dueser, D. Durnford, K. Dutta, M. A. DuVernois, T. Ehrhardt, L. Eidenschink, A. Eimer, C. Eldridge, P. Eller, E. Ellinger, D. Elsässer, R. Engel, H. Erpenbeck, W. Esmail, S. Eulig, J. Evans, P. A. Evenson, K. L. Fan, K. Fang, K. Farrag, A. R. Fazely, A. Fedynitch, N. Feigl, C. Finley, D. Fox, A. Franckowiak, S. Fukami, P. Fürst, J. Gallagher, E. Ganster, A. Garcia, M. Garcia, E. Genton, L. Gerhardt, A. Ghadimi, C. Glaser, T. Glüsenkamp, J. G. Gonzalez, S. Goswami, A. Granados, D. Grant, S. J. Gray, S. Griffin, K. M. Groth, D. Guevel, C. Günther, P. Gutjahr, C. Ha, A. Hallgren, L. Halve, F. Halzen, L. Hamacher, M. Handt, K. Hanson, J. Hardin, A. A. Harnisch, P. Hatch, A. Haungs, J. Häußler, K. Helbing, J. Hellrung, B. Henke, L. Hennig, F. Henningsen, L. Heuermann, R. Hewett, N. Heyer, S. Hickford, A. Hidvegi, C. Hill, G. C. Hill, R. Hmaid, K. D. Hoffman, A. Hollnagel, D. Hooper, S. Hori, K. Hoshina, M. Hostert, W. Hou, M. Hrywniak, T. Huber, K. Hultqvist, K. Hymon, A. Ishihara, W. Iwakiri, M. Jacquart, S. Jain, O. Janik, M. Jansson, M. Jin, N. Kamp, D. Kang, W. Kang, A. Kappes, L. Kardum, T. Karg, A. Karle, A. Katil, M. Kauer, J. L. Kelley, M. Khanal, A. Khatee Zathul, A. Kheirandish, T. Kim, H. Kimku, F. Kirchner, J. Kiryluk, C. Klein, S. R. Klein, Y. Kobayashi, S. Koch, A. Kochocki, R. Koirala, H. Kolanoski, T. Kontrimas, L. Köpke, C. Kopper, D. J. Koskinen, P. Koundal, M. Kowalski, T. Kozynets, A. Kravka, N. Krieger, T. Krishnan, K. Kruiswijk, E. Krupczak, A. Kumar, E. Kun, N. Kurahashi, C. Lagunas Gualda, L. Lallement Arnaud, M. J. Larson, F. Lauber, J. P. Lazar, K. Leonard DeHolton, A. Leszczyńska, C. Li, J. Liao, C. Lin, Q. R. Liu, Y. T. Liu, M. Liubarska, C. Love, L. Lu, F. Lucarelli, W. Luszczak, Y. Lyu, M. Macdonald, E. Magnus, Y. Makino, E. Manao, S. Mancina, A. Mand, I. C. Mariş, S. Marka, Z. Marka, L. Marten, I. Martinez-Soler, R. Maruyama, J. Mauro, F. Mayhew, F. McNally, K. Meagher, A. Medina, M. Meier, Y. Merckx, L. Merten, J. Mitchell, L. Molchany, S. Mondal, T. Montaruli, R. W. Moore, Y. Morii, A. Mosbrugger, D. Mousadi, E. Moyaux, T. Mukherjee, M. Nakos, U. Naumann, J. Necker, L. Neste, M. Neumann, H. Niederhausen, M. U. Nisa, K. Noda, A. Noell, A. Novikov, A. Obertacke, V. O'Dell, A. Olivas, R. Orsoe, J. Osborn, E. O'Sullivan, B. Owens, V. Palusova, H. Pandya, A. Parenti, N. Park, V. Parrish, E. N. Paudel, L. Paul, C. Pérez de los Heros, T. Pernice, T. C. Petersen, J. Peterson, S. Pick, M. Plum, A. Pontén, V. Poojyam, B. Pries, R. Procter-Murphy, G. T. Przybylski, L. Pyras, C. Raab, J. Rack-Helleis, N. Rad, M. Ravn, K. Rawlins, Z. Rechav, A. Rehman, I. Reistroffer, E. Resconi, S. Reusch, C. D. Rho, W. Rhode, L. Ricca, B. Riedel, A. Rifaie, E. J. Roberts, S. Rodan, M. Rongen, A. Rosted, C. Rott, T. Ruhe, L. Ruohan, D. Ryckbosch, J. Saffer, D. Salazar-Gallegos, P. Sampathkumar, A. Sandrock, G. Sanger-Johnson, M. Santander, S. Sarkar, M. Scarnera, M. Schaufel, H. Schieler, S. Schindler, L. Schlickmann, B. Schlüter, F. Schlüter, N. Schmeisser, T. Schmidt, A. Scholz, F. G. Schröder, S. Schwirn, S. Sclafani, D. Seckel, L. Seen, M. Seikh, S. Seunarine, P. A. Sevle Myhr, R. Shah, S. Shah, S. Shefali, N. Shimizu, B. Skrzypek, R. Snihur, J. Soedingrekso, D. Soldin, P. Soldin, G. Sommani, C. Spannfellner, G. M. Spiczak, C. Spiering, J. Stachurska, M. Stamatikos, T. Stanev, T. Stezelberger, T. Stürwald, T. Stuttard, G. W. Sullivan, I. Taboada, S. Ter-Antonyan, A. Terliuk, A. Thakuri, M. Thiesmeyer, W. G. Thompson, J. Thwaites, S. Tilav, K. Tollefson, J. A. Torres, S. Toscano, D. Tosi, K. Upshaw, A. Vaidyanathan, N. Valtonen-Mattila, J. Valverde, J. Vandenbroucke, T. Van Eeden, N. van Eijndhoven, L. Van Rootselaar, J. van Santen, J. Vara, F. Varsi, M. Venugopal, M. Vereecken, S. Vergara Carrasco, S. Verpoest, D. Veske, A. Vijai, J. Villarreal, C. Walck, A. Wang, E. H. S. Warrick, C. Weaver, P. Weigel, A. Weindl, J. Weldert, A. Y. Wen, C. Wendt, J. Werthebach, M. Weyrauch, N. Whitehorn, C. H. Wiebusch, D. R. Williams, L. Witthaus, G. Wrede, X. W. Xu, J. P. Yanez, Y. Yao, E. Yildizci, S. Yoshida, R. Young, F. Yu, S. Yu, T. Yuan, S. Yun-Cárcamo, A. Zander Jurowitzki, A. Zegarelli, S. Zhang, Z. Zhang, P. Zhelnin, P. Zilberman

Machine Learning · Cross-listed from stat.AP

Spatio-temporal modelling of electric vehicle charging demand

Kaoutar Bouaachra, Yvenn Amara-Ouali, Yannig Goude et al. (+1 more)

Accurate forecasting of electric vehicle (EV) charging demand is critical for grid management and infrastructure planning. Yet the field continues to rely on legacy benchmarks; such as the Palo Alto (2020) dataset; that fail to reflect the scale and behavioral diversity of modern charging networks. To address this, we introduce a novel large-scale longitudinal dataset collected across Scotland (2022 2025), which release it as an open benchmark for the community. Building on this dataset, we formulate EV charging demand as a spatio-temporal latent Gaussian field and perform approximate Bayesian inference via Integrated Nested Laplace Approximation (INLA). The resulting model jointly captures spatial dependence, temporal dynamics, and covariate effects within a unified proba bilistic framework. On station-level forecasting tasks, our approach achieves competitive predictive accuracy against machine learning baselines, while additionally providing principled uncertainty quan tification and interpretable spatial and temporal decompositions properties that are essential for risk-aware infrastructure planning.
Read the full abstract

All authors Kaoutar Bouaachra, Yvenn Amara-Ouali, Yannig Goude, Raphaël Lachieze-Rey

Machine Learning

Graph-Theoretic Models for the Prediction of Molecular Measurements

Anna Niane and Prudence Djagba

Graph-theoretic approaches offer simplicity, interpretability, and low computational cost for molecular property prediction. Among these, the model proposed by Mukwembi and Nyabadza, based on the external activity $D(G)$ and internal activity $\zeta(G)$ indices, achieved strong results on a small flavonoid dataset. However, its ability to generalize to larger and chemically diverse datasets has not been tested. This study evaluates the baseline $D(G)$-$\zeta(G)$ polynomial model on five benchmark datasets from MoleculeNet, covering biological activity (BACE, 1,513 molecules), lipophilicity (LogP synthetic, 14,610 molecules; LogP experimental, 753 molecules), aqueous solubility (ESOL, 1,128 molecules), and hydration free energy (SAMPL, 642 molecules). The baseline model achieves an average $R^2 = 0.24$, confirming limited transferability. To address this, a systematic enhancement framework is proposed, progressively incorporating Ridge regularization, additional graph descriptors, physicochemical properties, ensemble learning with Gradient Boosting, Lasso feature selection, and a hybrid approach combining topological indices with Morgan fingerprints. The enhanced models raise the average best $R^2$ to 0.79, with individual improvements ranging from 165% to 274%. All improvements are statistically significant ($p < 0.001$). A direct comparison with a Graph Convolutional Network under identical experimental conditions shows that the enhanced classical models match or outperform deep learning on all five datasets. Comparison with the recent GNN+PGM hybrid of Djagba et al.\ further confirms competitiveness, with the enhanced models achieving the best results on two datasets and tying on one. The entire framework requires no GPU, trains in under five minutes, and uses only open-source tools, making it accessible for researchers in resource-limited settings.
Read the full abstract

Machine Learning

Expert Upcycling: Shifting the Compute-Efficient Frontier of Mixture-of-Experts

Chaitanya Dwivedi, Binxuan Huang, Himanshu Gupta et al. (+3 more)

Mixture-of-Experts (MoE) has become the dominant architecture for scaling large language models: frontier models routinely decouple total parameters from per-token computation through sparse expert routing. Scaling laws show that under fixed active computation, model quality scales predictably with total parameters, and MoEs realize this by increasing expert count. However, training large MoEs is expensive, as memory requirements and inter-device communication both scale with total parameter count. We propose expert upcycling, a method for progressively expanding MoE capacity by increasing the number of experts during continued pre-training (CPT). Given a trained E-expert model, the upcycling operator constructs an mE-expert model through expert duplication and router extension while holding top-K routing fixed, preserving per-token inference cost. Duplication provides a warm initialization: the expanded model inherits the source checkpoint's learned representations, starting from a substantially lower loss than random initialization. Subsequent CPT then breaks the symmetry among duplicated experts to drive specialization. We formalize the upcycling operator and develop a theoretical framework decomposing the quality gap into a capacity term and an initialization term. We further introduce utility-based expert selection, which uses gradient-based importance scores to guide non-uniform duplication, more than tripling gap closure when CPT is limited. In our 7B-13B total parameter experiments, the upcycled model matches the fixed-size baseline on validation loss while saving 32% of GPU hours. Comprehensive ablations across model scales, activation ratios, MoE architectures, and training budgets yield a practical recipe for deploying expert upcycling, establishing it as a principled, compute-efficient alternative to training large MoE models from scratch.
Read the full abstract

All authors Chaitanya Dwivedi, Binxuan Huang, Himanshu Gupta, Pratik Jayarao, Neeraj Varshney, Bing Yin

Machine Learning

Generalization at the Edge of Stability

Mario Tuci, Caner Korkmaz, Umut Şimşekli et al. (+1 more)

Training modern neural networks often relies on large learning rates, operating at the edge of stability, where the optimization dynamics exhibit oscillatory and chaotic behavior. Empirically, this regime often yields improved generalization performance, yet the underlying mechanism remains poorly understood. In this work, we represent stochastic optimizers as random dynamical systems, which often converge to a fractal attractor set (rather than a point) with a smaller intrinsic dimension. Building on this connection and inspired by Lyapunov dimension theory, we introduce a novel notion of dimension, coined the `sharpness dimension', and prove a generalization bound based on this dimension. Our results show that generalization in the chaotic regime depends on the complete Hessian spectrum and the structure of its partial determinants, highlighting a complexity that cannot be captured by the trace or spectral norm considered in prior work. Experiments across various MLPs and transformers validate our theory while also providing new insights into the recently observed phenomenon of grokking.
Read the full abstract

All authors Mario Tuci, Caner Korkmaz, Umut Şimşekli, Tolga Birdal

Machine Learning · Cross-listed from math.PR

Phase Transitions in the Fluctuations of Functionals of Random Neural Networks

Simmaco Di Lillo, Leonardo Maini and Domenico Marinucci

We establish central and non-central limit theorems for sequences of functionals of the Gaussian output of an infinitely-wide random neural network on the d-dimensional sphere . We show that the asymptotic behaviour of these functionals as the depth of the network increases depends crucially on the fixed points of the covariance function, resulting in three distinct limiting regimes: convergence to the same functional of a limiting Gaussian field, convergence to a Gaussian distribution, convergence to a distribution in the Qth Wiener chaos. Our proofs exploit tools that are now classical (Hermite expansions, Diagram Formula, Stein-Malliavin techniques), but also ideas which have never been used in similar contexts: in particular, the asymptotic behaviour is determined by the fixed-point structure of the iterative operator associated with the covariance, whose nature and stability governs the different limiting regimes.
Read the full abstract

Machine Learning

Safe Continual Reinforcement Learning in Non-stationary Environments

Austin Coursey, Abel Diaz-Gonzalez, Marcos Quinones-Grueiro et al. (+1 more)

Reinforcement learning (RL) offers a compelling data-driven paradigm for synthesizing controllers for complex systems when accurate physical models are unavailable; however, most existing control-oriented RL methods assume stationarity and, therefore, struggle in real-world non-stationary deployments where system dynamics and operating conditions can change unexpectedly. Moreover, RL controllers acting in physical environments must satisfy safety constraints throughout their learning and execution phases, rendering transient violations during adaptation unacceptable. Although continual RL and safe RL have each addressed non-stationarity and safety, respectively, their intersection remains comparatively unexplored, motivating the study of safe continual RL algorithms that can adapt over the system's lifetime while preserving safety. In this work, we systematically investigate safe continual reinforcement learning by introducing three benchmark environments that capture safety-critical continual adaptation and by evaluating representative approaches from safe RL, continual RL, and their combinations. Our empirical results reveal a fundamental tension between maintaining safety constraints and preventing catastrophic forgetting under non-stationary dynamics, with existing methods generally failing to achieve both objectives simultaneously. To address this shortcoming, we examine regularization-based strategies that partially mitigate this trade-off and characterize their benefits and limitations. Finally, we outline key open challenges and research directions toward developing safe, resilient learning-based controllers capable of sustained autonomous operation in changing environments.
Read the full abstract

All authors Austin Coursey, Abel Diaz-Gonzalez, Marcos Quinones-Grueiro, Gautam Biswas

Machine Learning

FASTER: Value-Guided Sampling for Fast RL

Perry Dong, Alexander Swerdlow, Dorsa Sadigh et al. (+1 more)

Some of the most performant reinforcement learning algorithms today can be prohibitively expensive as they use test-time scaling methods such as sampling multiple action candidates and selecting the best one. In this work, we propose FASTER, a method for getting the benefits of sampling-based test-time scaling of diffusion-based policies without the computational cost by tracing the performance gain of action samples back to earlier in the denoising process. Our key insight is that we can model the denoising of multiple action candidates and selecting the best one as a Markov Decision Process (MDP) where the goal is to progressively filter action candidates before denoising is complete. With this MDP, we can learn a policy and value function in the denoising space that predicts the downstream value of action candidates in the denoising process and filters them while maximizing returns. The result is a method that is lightweight and can be plugged into existing generative RL algorithms. Across challenging long-horizon manipulation tasks in online and batch-online RL, FASTER consistently improves the underlying policies and achieves the best overall performance among the compared methods. Applied to a pretrained VLA, FASTER achieves the same performance while substantially reducing training and inference compute requirements. Code is available at https://github.com/alexanderswerdlow/faster .
Read the full abstract

All authors Perry Dong, Alexander Swerdlow, Dorsa Sadigh, Chelsea Finn

Machine Learning

FB-NLL: A Feature-Based Approach to Tackle Noisy Labels in Personalized Federated Learning

Abdulmoneam Ali and Ahmed Arafa

Personalized Federated Learning (PFL) aims to learn multiple task-specific models rather than a single global model across heterogeneous data distributions. Existing PFL approaches typically rely on iterative optimization-such as model update trajectories-to cluster users that need to accomplish the same tasks together. However, these learning-dynamics-based methods are inherently vulnerable to low-quality data and noisy labels, as corrupted updates distort clustering decisions and degrade personalization performance. To tackle this, we propose FB-NLL, a feature-centric framework that decouples user clustering from iterative training dynamics. By exploiting the intrinsic heterogeneity of local feature spaces, FB-NLL characterizes each user through the spectral structure of the covariances of their feature representations and leverages subspace similarity to identify task-consistent user groupings. This geometry-aware clustering is label-agnostic and is performed in a one-shot manner prior to training, significantly reducing communication overhead and computational costs compared to iterative baselines. Complementing this, we introduce a feature-consistency-based detection and correction strategy to address noisy labels within clusters. By leveraging directional alignment in the learned feature space and assigning labels based on class-specific feature subspaces, our method mitigates corrupted supervision without requiring estimation of stochastic noise transition matrices. In addition, FB-NLL is model-independent and integrates seamlessly with existing noise-robust training techniques. Extensive experiments across diverse datasets and noise regimes demonstrate that our framework consistently outperforms state-of-the-art baselines in terms of average accuracy and performance stability.
Read the full abstract

Machine Learning · Cross-listed from cs.RO

VLA Foundry: A Unified Framework for Training Vision-Language-Action Models

Jean Mercat, Sedrick Keh, Kushal Arora et al. (+5 more)

We present VLA Foundry, an open-source framework that unifies LLM, VLM, and VLA training in a single codebase. Most open-source VLA efforts specialize on the action training stage, often stitching together incompatible pretraining pipelines. VLA Foundry instead provides a shared training stack with end-to-end control, from language pretraining to action-expert fine-tuning. VLA Foundry supports both from-scratch training and pretrained backbones from Hugging Face. To demonstrate the utility of our framework, we train and release two types of models: the first trained fully from scratch through our LLM–>VLM–>VLA pipeline and the second built on the pretrained Qwen3-VL backbone. We evaluate closed-loop policy performance of both models on LBM Eval, an open-data, open-source simulator. We also contribute usability improvements to the simulator and the STEP analysis tools for easier public use. In the nominal evaluation setting, our fully-open from-scratch model is on par with our prior closed-source work and substituting in the Qwen3-VL backbone leads to a strong multi-task table top manipulation policy outperforming our baseline by a wide margin. The VLA Foundry codebase is available at https://github.com/TRI-ML/vla_foundry and all multi-task model weights are released on https://huggingface.co/collections/TRI-ML/vla-foundry. Additional qualitative videos are available on the project website https://tri-ml.github.io/vla_foundry.
Read the full abstract

All authors Jean Mercat, Sedrick Keh, Kushal Arora, Isabella Huang, Paarth Shah, Haruki Nishimura, Shun Iwase, Katherine Liu

Machine Learning

Benign Overfitting in Adversarial Training for Vision Transformers

Jiaming Zhang, Meng Ding, Shaopeng Fu et al. (+2 more)

Despite the remarkable success of Vision Transformers (ViTs) across a wide range of vision tasks, recent studies have revealed that they remain vulnerable to adversarial examples, much like Convolutional Neural Networks (CNNs). A common empirical defense strategy is adversarial training, yet the theoretical underpinnings of its robustness in ViTs remain largely unexplored. In this work, we present the first theoretical analysis of adversarial training under simplified ViT architectures. We show that, when trained under a signal-to-noise ratio that satisfies a certain condition and within a moderate perturbation budget, adversarial training enables ViTs to achieve nearly zero robust training loss and robust generalization error under certain regimes. Remarkably, this leads to strong generalization even in the presence of overfitting, a phenomenon known as benign overfitting, previously only observed in CNNs (with adversarial training). Experiments on both synthetic and real-world datasets further validate our theoretical findings.
Read the full abstract

All authors Jiaming Zhang, Meng Ding, Shaopeng Fu, Jingfeng Zhang, Di Wang

Machine Learning

Adaptive MSD-Splitting: Enhancing C4.5 and Random Forests for Skewed Continuous Attributes

Jake Lee

The discretization of continuous numerical attributes remains a persistent computational bottleneck in the induction of decision trees, particularly as dataset dimensions scale. Building upon the recently proposed MSD-Splitting technique – which bins continuous data using the empirical mean and standard deviation to dramatically improve the efficiency and accuracy of the C4.5 algorithm – we introduce Adaptive MSD-Splitting (AMSD). While standard MSD-Splitting is highly effective for approximately symmetric distributions, its rigid adherence to fixed one-standard-deviation cutoffs can lead to catastrophic information loss in highly skewed data, a common artifact in real-world biomedical and financial datasets. AMSD addresses this by dynamically adjusting the standard deviation multiplier based on feature skewness, narrowing intervals in dense regions to preserve discriminative resolution. Furthermore, we integrate AMSD into ensemble methods, specifically presenting the Random Forest-AMSD (RF-AMSD) framework. Empirical evaluations on the Census Income, Heart Disease, Breast Cancer, and Forest Covertype datasets demonstrate that AMSD yields a 2-4% accuracy improvement over standard MSD-Splitting, while maintaining near-identical O(N) time complexity reductions compared to the O(N log N) exhaustive search. Our Random Forest extension achieves state-of-the-art accuracy at a fraction of standard computational costs, confirming the viability of adaptive statistical binning in large-scale ensemble learning architectures.
Read the full abstract

Machine Learning

Ultrametric OGP - parametric RDT symmetric binary perceptron connection

Mihailo Stojnic

In [97,99,100], an fl-RDT framework is introduced to characterize statistical computational gaps (SCGs). Studying symmetric binary perceptrons (SBPs), [100] obtained an algorithmic threshold estimate $\alpha_a\approx \alpha_c^{(7)}\approx 1.6093$ at the 7th lifting level (for $\kappa=1$ margin), closely approaching $1.58$ local entropy (LE) prediction [18]. In this paper, we further connect parametric RDT to overlap gap properties (OGPs), another key geometric feature of the solution space. Specifically, for any positive integer $s$, we consider $s$-level ultrametric OGPs ($ult_s$-OGPs) and rigorously upper-bound the associated constraint densities $\alpha_{ult_s}$. To achieve this, we develop an analytical union-bounding program consisting of combinatorial and probabilistic components. By casting the combinatorial part as a convex problem and the probabilistic part as a nested integration, we conduct numerical evaluations and obtain that the tightest bounds at the first two levels, $\bar{\alpha}_{ult_1} \approx 1.6578$ and $\bar{\alpha}_{ult_2} \approx 1.6219$, closely approach the 3rd and 4th lifting level parametric RDT estimates, $\alpha_c^{(3)} \approx 1.6576$ and $\alpha_c^{(4)} \approx 1.6218$. We also observe excellent agreement across other key parameters, including overlap values and the relative sizes of ultrametric clusters. Based on these observations, we propose several conjectures linking $ult$-OGP and parametric RDT. Specifically, we conjecture that algorithmic threshold $\alpha_a=\lim_{s\rightarrow\infty} \alpha_{ult_s} = \lim_{s\rightarrow\infty} \bar{\alpha}{ult_s} = \lim_{r\rightarrow\infty} \alpha_{c}^{(r)}$, and $\alpha_{ult_s} \leq \alpha_{c}^{(s+2)}$ (with possible equality for some (maybe even all) $s$). Finally, we discuss the potential existence of a full isomorphism connecting all key parameters of $ult$-OGP and parametric RDT.
Read the full abstract

Machine Learning

On two ways to use determinantal point processes for Monte Carlo integration

Guillaume Gautier, Rémi Bardenet and Michal Valko

The standard Monte Carlo estimator $\widehat{I}_N^{\mathrm{MC}}$ of $\int fd\omega$ relies on independent samples from $\omega$ and has variance of order $1/N$. Replacing the samples with a determinantal point process (DPP), a repulsive distribution, makes the estimator consistent, with variance rates that depend on how the DPP is adapted to $f$ and $\omega$. We examine two existing DPP-based estimators: one by Bardenet & Hardy (2020) with a rate of $\mathcal{O}(N^{-(1+1/d)})$ for smooth $f$, but relying on a fixed DPP. The other, by Ermakov & Zolotukhin (1960), is unbiased with rate of order $1/N$, like Monte Carlo, but its DPP is tailored to $f$. We revisit these estimators, generalize them to continuous settings, and provide sampling algorithms.
Read the full abstract

Machine Learning

Planning in entropy-regularized Markov decision processes and games

Jean-Bastien Grill, Omar Darwiche Domingues, Pierre Ménard et al. (+2 more)

We propose SmoothCruiser, a new planning algorithm for estimating the value function in entropy-regularized Markov decision processes and two-player games, given a generative model of the environment. SmoothCruiser makes use of the smoothness of the Bellman operator promoted by the regularization to achieve problem-independent sample complexity of order O~(1/epsilon^4) for a desired accuracy epsilon, whereas for non-regularized settings there are no known algorithms with guaranteed polynomial sample complexity in the worst case.
Read the full abstract

All authors Jean-Bastien Grill, Omar Darwiche Domingues, Pierre Ménard, Rémi Munos, Michal Valko

Machine Learning

PREF-XAI: Preference-Based Personalized Rule Explanations of Black-Box Machine Learning Models

Salvatore Greco, Jacek Karolczak, Roman Słowiński et al. (+1 more)

Explainable artificial intelligence (XAI) has predominantly focused on generating model-centric explanations that approximate the behavior of black-box models. However, such explanations often overlook a fundamental aspect of interpretability: different users require different explanations depending on their goals, preferences, and cognitive constraints. Although recent work has explored user-centric and personalized explanations, most existing approaches rely on heuristic adaptations or implicit user modeling, lacking a principled framework for representing and learning individual preferences. In this paper, we consider Preference-Based Explainable Artificial Intelligence (PREF-XAI), a novel perspective that reframes explanation as a preference-driven decision problem. Within PREF-XAI, explanations are not treated as fixed outputs, but as alternatives to be evaluated and selected according to user-specific criteria. In the PREF-XAI perspective, here we propose a methodology that combines rule-based explanations with formal preference learning. User preferences are elicited through a ranking of a small set of candidate explanations and modeled via an additive utility function inferred using robust ordinal regression. Experimental results on real-world datasets show that PREF-XAI can accurately reconstruct user preferences from limited feedback, identify highly relevant explanations, and discover novel explanatory rules not initially considered by the user. Beyond the proposed methodology, this work establishes a connection between XAI and preference learning, opening new directions for interactive and adaptive explanation systems.
Read the full abstract

All authors Salvatore Greco, Jacek Karolczak, Roman Słowiński, Jerzy Stefanowski

Machine Learning · Cross-listed from cs.RO

Learning Hybrid-Control Policies for High-Precision In-Contact Manipulation Under Uncertainty

Hunter L. Brown, Geoffrey Hollinger and Stefan Lee

Reinforcement learning-based control policies have been frequently demonstrated to be more effective than analytical techniques for many manipulation tasks. Commonly, these methods learn neural control policies that predict end-effector pose changes directly from observed state information. For tasks like inserting delicate connectors which induce force constraints, pose-based policies have limited explicit control over force and rely on carefully tuned low-level controllers to avoid executing damaging actions. In this work, we present hybrid position-force control policies that learn to dynamically select when to use force or position control in each control dimension. To improve learning efficiency of these policies, we introduce Mode-Aware Training for Contact Handling (MATCH) which adjusts policy action probabilities to explicitly mirror the mode selection behavior in hybrid control. We validate MATCH's learned policy effectiveness using fragile peg-in-hole tasks under extreme localization uncertainty. We find MATCH substantially outperforms pose-control policies – solving these tasks with up to 10% higher success rates and 5x fewer peg breaks than pose-only policies under common types of state estimation error. MATCH also demonstrates data efficiency equal to pose-control policies, despite learning in a larger and more complex action space. In over 1600 sim-to-real experiments, we find MATCH succeeds twice as often as pose policies in high noise settings (33% vs.~68%) and applies ~30% less force on average compared to variable impedance policies on a Franka FR3 in laboratory conditions.
Read the full abstract

Machine Learning

Budgeted Online Influence Maximization

Pierre Perrault, Jennifer Healey, Zheng Wen et al. (+1 more)

We introduce a new budgeted framework for online influence maximization, considering the total cost of an advertising campaign instead of the common cardinality constraint on a chosen influencer set. Our approach better models the real-world setting where the cost of influencers varies and advertisers want to find the best value for their overall social advertising budget. We propose an algorithm assuming an independent cascade diffusion model and edge level semi-bandit feedback, and provide both theoretical and experimental results. Our analysis is also valid for the cardinality constraint setting and improves the state of the art regret bound in this case.
Read the full abstract

All authors Pierre Perrault, Jennifer Healey, Zheng Wen, Michal Valko

Machine Learning

HardNet++: Nonlinear Constraint Enforcement in Neural Networks

Andrea Goertzen, Kaveh Alim, Youngjae Min et al. (+1 more)

Enforcing constraint satisfaction in neural network outputs is critical for safety, reliability, and physical fidelity in many control and decision-making applications. While soft-constrained methods penalize constraint violations during training, they do not guarantee constraint adherence during inference. Other approaches guarantee constraint satisfaction via a projection layer, but often rely on the existence of a tractable projection onto the feasible set, limiting their utility in more general problem settings. Many real-world problems of interest are nonlinear and lack the special structure admitting a tractable projection, motivating the development of methods that can enforce general nonlinear constraints. To this end, we introduce HardNet++, a constraint-satisfaction method that enforces linear and nonlinear equality and inequality constraints. Our approach iteratively adjusts the network output via damped local linearizations of the constraints. Each iteration is differentiable, admitting an end-to-end training framework, where the constraint satisfaction layer is active during training. We show that under certain regularity conditions, this procedure enforces nonlinear constraint satisfaction to arbitrary tolerance. Finally, we demonstrate tight constraint adherence without loss of optimality in a learning-for-optimization context, where we apply this method to a nonlinear model predictive control problem.
Read the full abstract

All authors Andrea Goertzen, Kaveh Alim, Youngjae Min, Navid Azizan

Machine Learning · Cross-listed from cs.CL

Chat2Workflow: A Benchmark for Generating Executable Visual Workflows with Natural Language

Yi Zhong, Buqiang Xu, Yijun Wang et al. (+4 more)

At present, executable visual workflows have emerged as a mainstream paradigm in real-world industrial deployments, offering strong reliability and controllability. However, in current practice, such workflows are almost entirely constructed through manual engineering: developers must carefully design workflows, write prompts for each step, and repeatedly revise the logic as requirements evolve – making development costly, time-consuming, and error-prone. To study whether large language models can automate this multi-round interaction process, we introduce Chat2Workflow, a benchmark for generating executable visual workflows directly from natural language, and propose a robust agentic baseline to improve performance. The benchmark is built from a large collection of real-world business workflows, with each instance designed so that the generated workflow can be transformed and directly deployed to practical workflow platforms such as Dify and Coze. Experimental results show that while state-of-the-art language models can often capture high-level intent, they struggle to generate correct, stable, and executable workflows, especially given complex and evolving requirements. Although our agentic baseline yields up to 6.05% resolve rate gains, the remaining real-world gap positions Chat2Workflow as a foundation for advancing industrial-grade automation. Code is available at https://github.com/zjunlp/Chat2Workflow.
Read the full abstract

All authors Yi Zhong, Buqiang Xu, Yijun Wang, Zifei Shan, Shuofei Qiao, Guozhou Zheng, Ningyu Zhang

Machine Learning · Cross-listed from cs.IR

From Top-1 to Top-K: A Reproducibility Study and Benchmarking of Counterfactual Explanations for Recommender Systems

Quang-Huy Nguyen, Thanh-Hai Nguyen, Khac-Manh Thai et al. (+6 more)

Counterfactual explanations (CEs) provide an intuitive way to understand recommender systems by identifying minimal modifications to user-item interactions that alter recommendation outcomes. Existing CE methods for recommender systems, however, have been evaluated under heterogeneous protocols, using different datasets, recommenders, metrics, and even explanation formats, which hampers reproducibility and fair comparison. Our paper systematically reproduces, re-implement, and re-evaluate eleven state-of-the-art CE methods for recommender systems, covering both native explainers (e.g., LIME-RS, SHAP, PRINCE, ACCENT, LXR, GREASE) and specific graph-based explainers originally proposed for GNNs. Here, a unified benchmarking framework is proposed to assess explainers along three dimensions: explanation format (implicit vs. explicit), evaluation level (item-level vs. list-level), and perturbation scope (user interaction vectors vs. user-item interaction graphs). Our evaluation protocol includes effectiveness, sparsity, and computational complexity metrics, and extends existing item-level assessments to top-K list-level explanations. Through extensive experiments on three real-world datasets and six representative recommender models, we analyze how well previously reported strengths of CE methods generalize across diverse setups. We observe that the trade-off between effectiveness and sparsity depends strongly on the specific method and evaluation setting, particularly under the explicit format; in addition, explainer performance remains largely consistent across item level and list level evaluations, and several graph-based explainers exhibit notable scalability limitations on large recommender graphs. Our results refine and challenge earlier conclusions about the robustness and practicality of CE generation methods in recommender systems: https://github.com/L2R-UET/CFExpRec.
Read the full abstract

All authors Quang-Huy Nguyen, Thanh-Hai Nguyen, Khac-Manh Thai, Duc-Hoang Pham, Huy-Son Nguyen, Cam-Van Thi Nguyen, Masoud Mansoury, Duc-Trong Le, Hoang-Quynh Le

Machine Learning

Disentangling Damage from Operational Variability: A Label-Free Self-Supervised Representation Learning Framework for Output-Only Structural Damage Identification

Xudong Jian, Charikleia Stoura, Simon Scandella et al. (+1 more)

Damage identification is a core task in structural health monitoring. In practice, however, its reliability is often compromised by confounding non-damage effects, such as variations in excitation and environmental conditions, which can induce changes comparable to or larger than those caused by structural damage. To address this challenge, this study proposes a self-supervised label-free disentangled representation learning framework for robust vibration-based structural damage identification. The proposed framework employs an autoencoder with two latent representations to learn directly from raw vibration acceleration signals. A self-supervised invariance regularization, implemented via Variance-Invariance-Covariance Regularization (VICReg), is imposed on one latent representation using baseline data where structural damage is assumed constant but operational and environmental conditions vary. In addition, a frequency-domain constraint is introduced to enforce agreement between the power spectral density reconstructed from the latent representation and that computed from the corresponding input time series. Together, these mechanisms promote disentanglement, enabling the learned representation to be sensitive to damage-related characteristics while remaining invariant to nuisance variability. The framework is trained in a fully end-to-end and label-free manner, requiring no prior information on damage, excitation, or environmental conditions, making it well-suited for real-world applications. Its effectiveness is validated on two distinct real-world vibration datasets, including a bridge and a gearbox. The results demonstrate robustness to operational variability, strong generalization capability, and good performance in both damage detection and quantification.
Read the full abstract

All authors Xudong Jian, Charikleia Stoura, Simon Scandella, Eleni Chatzi

Machine Learning

SAGE: Training-Free Semantic Evidence Composition for Edge-Cloud Inference under Hard Uplink Budgets

Inhyeok Choi and Hyuncheol Park

Edge-cloud hybrid inference offloads difficult inputs to a powerful remote model, but the uplink channel imposes hard per-request constraints on the number of bits that can be transmitted. We show that selecting transmitted content based solely on attention-based importance, the standard approach in collaborative inference, is inherently limited under hard budgets. Two findings support this claim. First, replacing high-importance units with low-importance but complementary ones improves server accuracy. This shows that what matters is not individual importance but how well the transmitted set covers diverse aspects of the input. Second, spatially uniform selection without any content information achieves competitive accuracy at moderate budgets. This confirms that spatial coverage alone carries independent value. Based on this analysis, we propose SAGE (Semantic Attention-Guided Evidence), a principled, training-free method that combines importance filtering with embedding-diversity sampling. SAGE achieves 93% of the server ceiling in offloaded accuracy while transmitting fewer than half of the available evidence units on ImageNet-1K, substantially outperforming importance-only composition.
Read the full abstract

Machine Learning · Cross-listed from cs.CL

RoLegalGEC: Legal Domain Grammatical Error Detection and Correction Dataset for Romanian

Mircea Timpuriu, Mihaela-Claudia Cercel and Dumitru-Clementin Cercel

The importance of clear and correct text in legal documents cannot be understated, and, consequently, a grammatical error correction tool meant to assist a professional in the law must have the ability to understand the possible errors in the context of a legal environment, correcting them accordingly, and implicitly needs to be trained in the same environment, using realistic legal data. However, the manually annotated data required by such a process is in short supply for languages such as Romanian, much less for a niche domain. The most common approach is the synthetic generation of parallel data; however, it requires a structured understanding of the Romanian grammar. In this paper, we introduce, to our knowledge, the first Romanian-language parallel dataset for the detection and correction of grammatical errors in the legal domain, RoLegalGEC, which aggregates 350,000 examples of errors in legal passages, along with error annotations. Moreover, we evaluate several neural network models that transform the dataset into a valuable tool for both detecting and correcting grammatical errors, including knowledge-distillation Transformers, sequence tagging architectures for detection, and a variety of pre-trained text-to-text Transformer models for correction. We consider that the set of models, together with the novel RoLegalGEC dataset, will enrich the resource base for further research on Romanian.
Read the full abstract

Machine Learning

An Efficient Black-Box Reduction from Online Learning to Multicalibration, and a New Route to $\Phi$-Regret Minimization

Gabriele Farina and Juan Carlos Perdomo

We give a Gordon-Greenwald-Marks (GGM) style black-box reduction from online learning to online multicalibration. Concretely, we show that to achieve high-dimensional multicalibration with respect to a class of functions $\mathcal{H}$, it suffices to combine any no-regret learner over $\mathcal {H} $ with an expected variational inequality (EVI) solver. We also prove a converse statement showing that efficient multicalibration implies efficient EVI solving, highlighting how EVIs in multicalibration mirror the role of fixed points in the GGM result for $\Phi$-regret. This first set of results addresses the high-dimensional analogue of the open question in Garg, Jung, Reingold, and Roth (SODA '24), showing that oracle-efficient online multicalibration with $\sqrt{T}$-type guarantees is possible in full generality. Furthermore, our GGM-style reduction unifies the analyses of existing algorithms, transfers guarantees from online learning to multicalibration for challenging environments with delayed observations or censored outcomes, and yields the first efficient black-box reduction between online learning and multiclass omniprediction. Our second main result is a fine-grained reduction from high-dimensional online multicalibration to (contextual) $\Phi$-regret minimization. Together with our first result, this establishes a new route from external regret to $\Phi$-regret that bypasses sophisticated fixed-point or semi-separation machinery, dramatically simplifies a result of Daskalakis, Farina, Fishelson, Pipis, and Schneider (STOC '25) while improving rates, and yields new algorithms that are robust to richer deviation classes, such as those belonging to any reproducing kernel Hilbert space.
Read the full abstract