Track trials, findings, and clinical evidence as it lands.
Search papers, labs, and topics across Lattice.
This paper addresses the challenge of generative models producing diverse outputs by externalizing diversity control into a test-time sampling procedure. The MINE-SAMPLE-RENDER (MSR) pipeline first mines an interpretable categorical schema and associated statistics from a corpus, then samples attribute configurations based on these statistics, and finally uses the generative model solely as a conditional renderer. MSR demonstrably improves alignment with corpus distributions and enhances the diversity of generated outputs across various domains, including text-to-image and 3D asset generation.
Generative models can achieve significantly higher output diversity and corpus alignment by decoupling sampling from rendering at test time.
Modern generative models produce high-quality individual outputs but struggle to sample diversely, even when explicitly instructed to do so. We introduce MINE-SAMPLE-RENDER (MSR), a test-time approach that externalizes the diversity into an explicit, corpus-derived sampling procedure. Our pipeline (i) mines an interpretable categorical schema from a corpus of real artifacts and supplies statistics for sampling; (ii) samples attribute configurations from the corpus-derived statistics; (iii) uses a generative model purely as a conditional renderer of a sampled configuration. Across weather report generation, user simulation, webpage design, text-to-image generation, and agent-generated 3D assets, we evaluate the diversity and output quality of MSR-guided generation and show that MSR improves alignment with corpus distributions and diversity of generated outputs.