ShortListing Model: A Streamlined SimplexDiffusion for Discrete Variable Generation

Yuxuan Song, Zhe Zhang, Yu Pei, Jingjing Gong, Qiying Yu +5 more
8/24/2025
cs.LGq-bio.GN

Official code: GenSI-THUAIR/SLM (12 stars, MIT, last commit 2025-09-30).

Abstract

Generative modeling of discrete variables is challenging yet crucial for applications in natural language processing and biological sequence design. We introduce the Shortlisting Model (SLM), a novel simplex-based diffusion model inspired by progressive candidate pruning. SLM operates on simplex centroids, reducing generation complexity and enhancing scalability. Additionally, SLM incorporates a flexible implementation of classifier-free guidance, enhancing unconditional generation performance. Extensive experiments on DNA promoter and enhancer design, protein design, character-level and large-vocabulary language modeling demonstrate the competitive performance and strong potential of SLM. Our code can be found at https://github.com/GenSI-THUAIR/SLM

View on arXivView PDF

Code Implementations(1)

GenSI-THUAIR/SLMOfficial100%
MIT

Project Page of ShortListing Model

122Python, MakefileAug 1, 20251 years agoMIT

Similar Papers

DiscDiff: Latent Diffusion Model for DNA Sequence Generation

82% match1 repo342024

Diffusion Language Models Are Versatile Protein Learners

81% match1 repo3452024

Diffusion on language model encodings for protein sequence generation

80% match1 repo252024

Diffusion on Language Model Encodings for Protein Sequence Generation

80% match2024

TaxDiff: Taxonomic-Guided Diffusion Model for Protein Sequence Generation

80% match1 repo752024

Latent Diffusion Model for DNA Sequence Generation

79% match1 repo342023

Cite this paper

@article{song2025shortlisting,
  title  = {ShortListing Model: A Streamlined SimplexDiffusion for Discrete Variable Generation},
  author = {Yuxuan Song and Zhe Zhang and Yu Pei and Jingjing Gong and Qiying Yu and Zheng Zhang and Mingxuan Wang and Hao Zhou and Jingjing Liu and Wei-Ying Ma},
  year   = {2025},
  eprint = {2508.17345},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2508.17345v1},
  journal = {NEURIPS 2025 2025}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2508.17345.svg)](https://codeofpaper.com/paper/2508.17345)

Discussion