REVIEW 5 cited by
Implicit Generation and Generalization in Energy-Based Models
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Energy based models (EBMs) are appealing due to their generality and simplicity in likelihood modeling, but have been traditionally difficult to train. We present techniques to scale MCMC based EBM training on continuous neural networks, and we show its success on the high-dimensional data domains of ImageNet32x32, ImageNet128x128, CIFAR-10, and robotic hand trajectories, achieving better samples than other likelihood models and nearing the performance of contemporary GAN approaches, while covering all modes of the data. We highlight some unique capabilities of implicit generation such as compositionality and corrupt image reconstruction and inpainting. Finally, we show that EBMs are useful models across a wide variety of tasks, achieving state-of-the-art out-of-distribution classification, adversarially robust classification, state-of-the-art continual online class learning, and coherent long term predicted trajectory rollouts.
Forward citations
Cited by 5 Pith papers
-
Wasserstein normalized autoencoder for anomaly detection
A Wasserstein-distance-trained normalized autoencoder detects semivisible jets in simulated LHC events with AUCs around 0.69–0.77, outperforming standard and normalized autoencoders on a ttbar background.
-
Compositional Scene Understanding through Inverse Generative Modeling
Composing per-concept diffusion models and inverting them with denoising loss enables multi-object scene understanding that generalizes beyond the training distribution.
-
Efficient Stochastic Optimisation via Sequential Monte Carlo
Sequential Monte Carlo samplers can approximate intractable gradients inside a first-order optimizer, yielding a general SOSMC framework that speeds up reward tuning of energy-based models in the reported settings.
-
AERO: A Redirection-Based Optimization Framework Inspired by Judo for Robust Probabilistic Forecasting
AERO's experimental optimizer is momentum SGD with Gaussian gradient noise, and its claimed state-of-the-art gains lack any baseline comparison.
- CodeDiffuser: Attention-Enhanced Diffusion Policy via VLM-Generated Code for Instruction Ambiguity
Discussion (0). Continue with ORCID to comment.