Nikolay Malkin

Discrete Compositional Generation via General Soft Operators and Robust Reinforcement Learning

Data-to-Energy Stochastic Dynamics

Multi-Marginal Flow Matching with Adversarially Learnt Interpolants

Adaptive Destruction Processes for Diffusion Samplers

A Comedy of Estimators: On KL Regularization in RL Training of LLMs

Recursive Self-Aggregation Unlocks Deep Thinking in Large Language Models

Forgetting is Everywhere

On Designing Diffusion Autoencoders for Efficient Generation and Representation Learning

Bayesian Symbolic Regression with Entropic Reinforcement Learning

Learning Diverse Attacks on Large Language Models for Robust Red-Teaming and Safety Tuning

Action abstractions for amortized sampling

Outsourced Diffusion Sampling: Efficient Posterior Inference in Latent Spaces of Generative Models

Adaptive teachers for amortized samplers

PQMass: Probabilistic Assessment of the Quality of Generative Models using Probability Mass Estimation

Can a Bayesian oracle prevent harm from an agent?

From discrete-time policies to continuous-time diffusion samplers: Asymptotic equivalences and faster training