machine learning

Learning-Theoretic Foundation for General Coded Computing: The Straggler Setting

GCC formulates coded computing through a natural end-to-end mean-squared error loss that directly measures the discrepancy between the desired computations and their recovered estimates, and enables theoretical performance guarantees for GCC under two complementary straggler regimes.

Off-Policy Evaluation for Semantic ID Recommenders: Does the Model's Own Code Hierarchy Help?

This work asks a simple question: can the model's own SID tree serve as the action abstraction for that OPE, and explains how resolution depth is the operative knob and a conditional bias bound links the coarsening bias to the quantizer's worst-case reconstruction residual and the target-logging divergence.

Conservative Hybrid Graph Networks for Process Systems with Learned Routing

The Conservative Hybrid Graph Network (CHGN), which learns routing, regime assignment, and removal rates as data-driven surrogates and inserts them into a fixed transport equation, so that the mass balance holds by construction for any predicted routing.

The Halt Vector: Internalizing a Causal Steering Intervention for Efficient Reasoning

The mechanism is a halt vector: a difference-of-means direction at layer 18 of this model whose steering strength controls how long it thinks, while a replicated value axis does nothing, and what works is reconstructing the whole steered activation with those dimensions pinned to their natural values.

Equivariant Sheaf Neural Networks: Learning Geometric Transport on Graphs

ESNN is introduced, an Equivariant Sheaf Neural Network that enriches this interaction by learning directed, matrix-valued transport between neighboring vector features while preserving exact Euclidean equivariance.

Curvature Cryptanalysis of Smooth Transformer Feed-Forward Networks

This work formalizes resulting Hessian collection as a partially symmetric decomposition to establish conditions for local identifiability and stability to exploit vector-output stencil reuse to reduce the structural query cost by a factor of 16.

Unsupervised Latent Space Alignment with Hyperspherical Geodesic Matching

HGA (Hyperspherical Gaussian Alignment), a method that directly optimizes a transformation between two latent spaces by maximizing a geometric measure of "fit" between them.

ERR+: Sequential Entropy Resolution for Efficient and Decisive LLM Reasoning

A formal analysis showing that joint optimization of the two objectives induces gradient conflict in early training, motivating the sequential design of ERR+, a two-phase RLVR framework grounded in this observation.

TransfHAR: Self-Supervised Wrist Representations for On-Demand Activity Recognition

TransfHAR is implemented as a real-time smartwatch application that lets users define and expand their own activity set for personalized recognition from only a few demonstrations, and indicates that broad self-supervised wrist pretraining provides an effective foundation for on-demand fine-grained activity recognition.

VBVR-Pro: A Scalable and Verifiable Suite for Native Visual Reasoning

This work introduces VBVR-Pro, a closed-loop testbed that makes native visual reasoning through generation trainable, verifiable, optimizable, and experimentally controllable, and identifies recurring failure modes of the prevalent VLM-as-a-judge paradigm.

PhysCoRe: Physics-Corrected Residual World Models for Material-Aware Deformable Dynamics

Experiments on real deformable-object manipulation sequences show that PhysCoRe outperforms state-of-the-art baselines in prediction accuracy, and that its predicted confidence forms a reliable distribution across the object's geometry, providing a natural signal for future confidence-guided exploration.

Learning from agile software development