Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
The paper introduces Spatiotemporal Noise-Contrastive Estimation (stNCE) for training energy-based models (EBMs) using stochastic interpolants, addressing limitations of existing spatial and temporal difference-based approaches. stNCE learns the energy function by contrasting joint spatiotemporal differences in the data distribution corrupted by noise at varying levels indexed by a time variable. Experiments on image and molecule datasets show that stNCE achieves competitive density estimation performance compared to state-of-the-art methods.
Combining spatial and temporal differences when training energy-based models unlocks better density estimation than using either alone.
Learning an energy-based model from data samples is a central problem in machine learning. Many recent and popular methods, such as denoising score matching for training energy-based diffusion models, use stochastic interpolants to corrupt data samples at different noise levels indexed by a time variable. This defines a joint density over both the data space and time, and most methods learn its energy through either spatial or temporal differences. We identify distinct failure modes for both of these approaches. To solve them, we propose Spatiotemporal Noise-Contrastive Estimation (stNCE), a framework for learning the energy through joint spatiotemporal differences. stNCE unifies many existing methods and leads to new training objectives. Experiments on images and molecules demonstrate performance competitive with state-of-the-art density estimation methods.
CMU ML