Track new results across pure and applied mathematics.
Search papers, labs, and topics across Lattice.
PRODiGI replaces dataset-specific density estimation by amortizing the inference of explicit, standalone generative programs into a single feedforward pass. This architecture combines categorical template prediction with non-autoregressive parameter decoding, yielding inspectable symbolic models that natively support exact sampling, density computation, and score evaluation. Across tabular benchmarks, the model outperforms existing pretrained density estimators in density and score MAE with multi-fold speedups, while subsequent program-space fine-tuning slashes generation MMD by 84%.
Instead of spending compute fitting bespoke generative models for every dataset, a single feedforward pass can now compile empirical samples directly into inspectable, executable generative programs.
Estimating probability densities from a finite set of samples typically requires dataset-specific model fitting. We introduce PRODiGI, a pretrained data-to-program model that infers an explicit, executable generative program in a single forward pass. Pretrained on synthetic datasets paired with their ground-truth programs, PRODiGI accommodates diverse generative families and data dimensionalities through template prediction and non-autoregressive program parameter decoding. Its inferred programs support direct sampling, density and score evaluation, and inspection independently of the pretrained model. We further introduce program-space fine-tuning, which refines differentiable program parameters by matching generated and empirical samples while keeping model parameters intact. Experiments show that PRODiGI achieves lower average density and score MAE than existing pretrained models, while offering multi-fold speedups over its closest competitors. Program-space fine-tuning further reduces generation MMD by 84%. By turning empirical data into explicit, reusable programs, PRODiGI introduces a new direction for fast, interpretable tabular generative modeling.