This paper establishes a large deviation principle for the generalization error of interpolating classifiers in the overparametrized regime.

The paper formalizes the problem of representation identifiability in supervised learning, showing that a representation property is identifiable if and only if it is constant across all possible fact…

The paper proposes Test-Time Training for Supervised Causal Learning (TTT-SCL), a novel framework that dynamically generates training data aligned with specific test instances to significantly improve…

This paper extends gradient boosting to functions of vector inputs using a simple algorithm with histogram-based decision trees.

A new boosting algorithm that strong learns concept classes closed under O(log 1/γ)-XOR using O(log 1/ε) calls to a γ-advantage weak learner and additional samples, by connecting boosting with list-de…

This paper analyzes the poor performance of Meta-learning for Training-data Selection (MTS) and proposes that increasing the batch size and incorporating informative features can significantly improve…

This paper characterizes proper binary classification from positive-only samples, revealing a rich landscape that differs from standard PAC learning.

This paper develops statistical learning theory for gradient boosting in Peaks-over-Threshold modeling using Generalized Pareto distributions, deriving error bounds and reducing gradient correlation.

A novel self-supervised hard negative sampling technique is proposed for two-tower recommendation systems using a large language model to generate challenging and informative negatives.

The twoblock clustering tree is introduced as a new type of interpretable regression tree for multivariate responses, using local multivariate linear models as leaves and twoblock dimension reduction…

This paper introduces survey sampling techniques to estimate or minimize empirical pairwise loss functions, showing that targeting informative pairs significantly reduces computational cost while main…

The paper establishes tight upper and lower bounds on the statistical cost of approximate machine unlearning for smooth strongly convex losses, showing that the optimal unlearning rate depends critica…

The paper introduces PRAXIS, a novel algorithm that efficiently approximates the computation of 'Rashomon sets' for decision trees, significantly reducing memory and runtime complexity.

The paper proposes pretraining a Perceiver-style in-context learner on synthetic data to solve Multiple Instance Learning (MIL) tasks efficiently in the low-label regime.

The paper introduces a novel, non-deep neural network architecture that achieves the performance of LLMs by finding the global optimum of the loss function in a single, closed-form iteration, eliminat…