papers

Publications (25)

cs.LG2020

Using Large Ensembles of Control Variates for Variational Inference

Tomas Geffner, Justin Domke

Variational inference is increasingly being addressed with stochastic optimization. In this setting, the gradient's variance plays a crucial role in the optimization procedure, sin…

cs.LG2026

Latent Target Score Matching, with an application to Simulation-Based Inference

Joohwan Ko, Tomas Geffner

Denoising score matching (DSM) for training diffusion models may suffer from high variance at low noise levels. Target Score Matching (TSM) mitigates this when clean data scores ar…

cs.CV2024

Stochastic Flow Matching for Resolving Small-Scale Physics

Stathi Fotiadis, Noah Brenowitz, Tomas Geffner +4

Conditioning diffusion and flow models have proven effective for super-resolving small-scale details in natural images.However, in physical sciences such as weather, super-resolvin…

cs.LG2020

Approximation Based Variance Reduction for Reparameterization Gradients

Tomas Geffner, Justin Domke

Flexible variational distributions improve variational inference but are harder to optimize. In this work we present a control variate that is applicable for any reparameterizable…

cs.AI2018

Compact Policies for Fully-Observable Non-Deterministic Planning as SAT

Tomas Geffner, Hector Geffner

Fully observable non-deterministic (FOND) planning is becoming increasingly important as an approach for computing proper policies in probabilistic planning, extended temporal plan…

cs.LG2025

Efficient Molecular Conformer Generation with SO(3)-Averaged Flow Matching and Reflow

Zhonglin Cao, Mario Geiger, Allan dos Santos Costa +6

Fast and accurate generation of molecular conformers is desired for downstream computational chemistry and drug discovery tasks. Currently, training and sampling state-of-the-art d…

q-bio.BM2025

ProtComposer: Compositional Protein Structure Generation with 3D Ellipsoids

Hannes Stark, Bowen Jing, Tomas Geffner +4

We develop ProtComposer to generate protein structures conditioned on spatial protein layouts that are specified via a set of 3D ellipsoids capturing substructure shapes and semant…

cs.CL2025

Energy-Based Diffusion Language Models for Text Generation

Minkai Xu, Tomas Geffner, Karsten Kreis +5

Despite remarkable progress in autoregressive language models, alternative generative paradigms beyond left-to-right generation are still being actively explored. Discrete diffusio…

cs.LG2024

Joint control variate for faster black-box variational inference

Xi Wang, Tomas Geffner, Justin Domke

Black-box variational inference performance is sometimes hindered by the use of gradient estimators with high variance. This variance comes from two sources of randomness: Data sub…

q-bio.BM2024

Aligning Target-Aware Molecule Diffusion Models with Exact Energy Optimization

Siyi Gu, Minkai Xu, Alexander Powers +6

Generating ligand molecules for specific protein targets, known as structure-based drug design, is a fundamental problem in therapeutics development and biological discovery. Recen…

cs.LG2025

Learning Straight Flows by Learning Curved Interpolants

Shiv Shankar, Tomas Geffner

Flow matching models typically use linear interpolants to define the forward/noise addition process. This, together with the independent coupling between noise and target distribut…

cs.LG2026

La-Proteina: Atomistic Protein Generation via Partially Latent Flow Matching

Tomas Geffner, Kieran Didi, Zhonglin Cao +6

Recently, many generative models for de novo protein structure design have emerged. Yet, only few tackle the difficult task of directly generating fully atomistic structures jointl…

cs.LG2026

Scaling Atomistic Protein Binder Design with Generative Pretraining and Test-Time Compute

Kieran Didi, Zuobai Zhang, Guoqing Zhou +11

Protein interaction modeling is central to protein design, which has been transformed by machine learning with applications in drug discovery and beyond. In this landscape, structu…

q-bio.BM2025

Consistent Synthetic Sequences Unlock Structural Diversity in Fully Atomistic De Novo Protein Design

Danny Reidenbach, Zhonglin Cao, Zuobai Zhang +8

High-quality training datasets are crucial for the development of effective protein design models, but existing synthetic datasets often include unfavorable sequence-structure pair…

stat.ML2022

Deep End-to-end Causal Inference

Tomas Geffner, Javier Antoran, Adam Foster +9

Causal inference is essential for data-driven decision making across domains such as business engagement, medical treatment and policy making. However, research on causal discovery…

cs.LG2023

Langevin Diffusion Variational Inference

Tomas Geffner, Justin Domke

Many methods that build powerful variational distributions based on unadjusted Langevin transitions exist. Most of these were developed using a wide range of different approaches a…

cs.LG2025

Proteina: Scaling Flow-based Protein Structure Generative Models

Tomas Geffner, Kieran Didi, Zuobai Zhang +8

Recently, diffusion- and flow-based generative models of protein structures have emerged as a powerful tool for de novo protein design. Here, we develop Proteina, a new large-scale…

cs.LG2026

DiLaDiff: Distilled Latent-Augmented Diffusion for Language Modeling

Jean-Marie Lemercier, Tomas Geffner, Karsten Kreis +3

Diffusion language models intrinsically fail to capture correlations between decoded tokens, which leads to a harsh trade-off between sampling quality and throughput. To solve this…

cs.LG2023

Compositional Score Modeling for Simulation-based Inference

Tomas Geffner, George Papamakarios, Andriy Mnih

Neural Posterior Estimation methods for simulation-based inference can be ill-suited for dealing with posterior distributions obtained by conditioning on multiple observations, as…

stat.ML2021

On the Difficulty of Unbiased Alpha Divergence Minimization

Tomas Geffner, Justin Domke

Several approximate inference algorithms have been proposed to minimize an alpha-divergence between an approximating distribution and a target distribution. Many of these algorithm…

cs.LG2021

Empirical Evaluation of Biased Methods for Alpha Divergence Minimization

Tomas Geffner, Justin Domke

In this paper we empirically evaluate biased methods for alpha-divergence minimization. In particular, we focus on how the bias affects the final solutions found, and how this depe…

cs.LG2025

Truncated Consistency Models

Sangyun Lee, Yilun Xu, Tomas Geffner +4

Consistency models have recently been introduced to accelerate sampling from diffusion models by directly predicting the solution (i.e., data) of the probability flow ODE (PF ODE)…

cs.LG2019

A Rule for Gradient Estimator Selection, with an Application to Variational Inference

Tomas Geffner, Justin Domke

Stochastic gradient descent (SGD) is the workhorse of modern machine learning. Sometimes, there are many different potential gradient estimators that can be used. When so, choosing…

cs.LG2021

MCMC Variational Inference via Uncorrected Hamiltonian Annealing

Tomas Geffner, Justin Domke

Given an unnormalized target distribution we want to obtain approximate samples from it and a tight lower bound on its (log) normalization constant log Z. Annealed Importance Sampl…

cs.LG2022

Variational Inference with Locally Enhanced Bounds for Hierarchical Models

Tomas Geffner, Justin Domke

Hierarchical models represent a challenging setting for inference algorithms. MCMC methods struggle to scale to large models with many local variables and observations, and variati…