José Miguel Hernández-Lobato

Associate Professor@University of Cambridge·英国·OpenReview

研究方向

Bayesia Machine Learning · Bayesian Optimization · Bayesian Neural Networks · Gaussian Process · Variational Inference · Expectation Propagation · Deep Generative Models

José Miguel Hernández-Lobato

RNE: plug-and-play diffusion inference-time control and energy-based training

There Was Never a Bottleneck in Concept Bottleneck Models

Vector Quantization using Gaussian Variational Autoencoder

Mitigating Forgetting in Low Rank Adaptation

GAPA: Post-Hoc Uncertainty Quantification for Pre-Trained Models via Activation-Space Gaussian Processes

Exploring the Trade-off between Quality and Diversity of Language Models during Reinforcement Learning

CREPE: Controlling diffusion with REPlica Exchange

Expanding the Chaos: Neural Operator for Stochastic (Partial) Differential Equations

Towards Training One-Step Diffusion Models Without Distillation

Weighted Conditional Flow Matching

A Diffusive Classification Loss for Learning Energy-based Generative Models

FEAT: Free energy Estimators with Adaptive Transport

Uncertainty Modeling in Graph Neural Networks via Stochastic Differential Equations

Scalable Gaussian Processes with Latent Kronecker Structure

Domain-Adapted Diffusion Model for PROTAC Linker Design Through the Lens of Density Ratio in Chemical Space

Aligning Multimodal Representations through an Information Bottleneck

BNEM: A Boltzmann Sampler Based on Bootstrapped Noised Energy Matching

Getting Free Bits Back from Rotational Symmetries in LLMs

Batched Bayesian optimization with correlated candidate uncertainties

Improving Antibody Design with Force-Guided Sampling in Diffusion Models

Progressive Tempering Sampler with Diffusion