Mine, Sample, Render: Test-Time Diversity Alignment Of Generative Models
MSR improves generative-model diversity by sampling corpus-derived attributes and using the model only as a renderer.
The paper introduces MINE-SAMPLE-RENDER (MSR), a test-time method that improves diversity in generative models by separating sampling from rendering. MSR mines an interpretable categorical schema and statistics from a corpus of real artifacts, samples attribute configurations from those statistics, and uses a generative model only as a conditional renderer. Across weather-report generation, user simulation, webpage design, text-to-image generation, and agent-generated 3D assets, the authors report better alignment with corpus distributions and higher output diversity without changing the underlying model.
- MSR mines a categorical schema and sampling statistics from a real corpus.
- A generative model then renders only the sampled attribute configuration.
- Evaluated on weather reports, user simulation, web design, text-to-image, and 3D assets.
- Authors report better corpus-distribution alignment and more diverse outputs.
From the original article
1 min read at huggingface.co
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.
By Kostas Triaridis, Morris E. Sharp, Nebojsa Jojic · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.