7 papers
Calibrating Generative Models to Feature Distributions with MMD Finetuning
Nathaniel L. Diamant, Brian L. Trippe
Generative models can produce individually plausible samples while deviating substantially from a target set in the distribution of key features. For example, a model pretrained on…
Closing the Approximation Gap in Simulation-free Latent SDEs
Henry D. Smith, Brian L. Trippe, Scott W. Linderman
Recovering dynamical systems from noisy observations is a recurring challenge across scientific domains, including neuroscience and physics. Latent stochastic differential equation…
Diffusion Language Model Parallel Decoding via Product-of-Experts Bridge
Juntong Shi, Brian L. Trippe, Jure Leskovec +2
Diffusion language models (DLMs) offer substantial speed advantages through parallel decoding, but the lack of token dependencies limits generation quality compared to autoregressi…
Calibrating Generative Models to Distributional Constraints
Henry D. Smith, Nathaniel L. Diamant, Brian L. Trippe
Generative models frequently suffer miscalibration, wherein statistics of the sampling distribution, such as the fraction of generations in a given class, deviate from desired valu…
Predicting mutational effects on protein binding from folding energy
Arthur Deng, Karsten Householder, Fang Wu +3
Accurate estimation of mutational effects on protein-protein binding energies is an open problem with applications in structural biology and therapeutic design. Several deep learni…
MotifBench: A standardized protein design benchmark for motif-scaffolding problems
Zhuoqi Zheng, Bo Zhang, Kieran Didi +5
The motif-scaffolding problem is a central task in computational protein design: Given the coordinates of atoms in a geometry chosen to confer a desired biochemical function (a mot…