Machine Learning

This paper presents a novel approach for embedding supervised models in hyperbolic space, linking loss functions to hyperbolic distances and introducing monotonic decision trees for unambiguous visual…

Hyper-opinion Evidential Deep Learning (HEDL) enhances out-of-distribution detection by integrating sharp and vague evidence for superior uncertainty estimation and classification accuracy.

HYDRA-FL: A novel hybrid knowledge distillation method makes federated learning robust against poisoning attacks while maintaining accuracy!

Hybrid RL algorithms achieve sharper error/regret bounds than existing offline/online RL methods in linear MDPs, improving sample efficiency without stringent assumptions on behavior policy quality.

Multi-head transformers utilize distinct attention patterns across layers—multiple heads are essential for initial data preprocessing, while a single head suffices for subsequent optimization steps, o…

CODA: A novel method for solving contextual goal-oriented problems with offline datasets, using unlabeled trajectories and context-goal pairs to create a fully labeled dataset, outperforming other bas…

Deep network pruning’s fundamental limits are characterized, revealing how weight magnitude and network sharpness determine the maximum achievable sparsity.

Boost ensemble accuracy by predicting performance with fewer classifiers using a novel polarization law and refined error bounds.

Low reward variance drastically improves contextual bandit regret, defying minimax assumptions and highlighting the crucial role of eluder dimension.

TAG-CF boosts collaborative filtering accuracy by up to 39.2% on cold users, using only a single message-passing step at test time, avoiding costly training-time computations.

CATY-IRL: A novel, provably efficient algorithm solves Inverse Reinforcement Learning’s scalability issues for large state spaces, improving upon state-of-the-art methods.

HORSE: A novel attention-based neural network significantly improves large-scale neural subset selection by up to 20%, addressing limitations in existing methods.

Higher-rank irreducible Cartesian tensors boost accuracy and efficiency in equivariant message-passing neural networks for atomistic simulations.

Adversarial training achieves minimax-optimal prediction error in high-dimensional linear regression under l∞-perturbation, improving upon existing methods.

Hierarchical Programmatic Option framework (HIPO) uses human-readable programs as options in reinforcement learning to solve long, repetitive tasks with improved interpretability and generalization.

Hierarchical Hybrid Sliced Wasserstein (H2SW) solves the challenge of comparing complex, heterogeneous joint distributions by introducing novel slicing operators, leading to a scalable and statistical…

MTGC tackles multi-timescale model drift in hierarchical federated learning.

HHD-GP leverages Helmholtz-Hodge decomposition within Gaussian Processes to learn physically meaningful components of dynamical systems, enhancing prediction accuracy and interpretability.

HGDL: Heterogeneous Graph Label Distribution Learning, a new framework that leverages graph topology and content to enhance label distribution prediction.

HiCS-FL: A novel federated learning client sampling method that leverages data heterogeneity for faster, more efficient global model training in non-IID settings.