ShortListing Model: A Streamlined SimplexDiffusion for Discrete Variable Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Song, Yuxuan, Zhang, Zhe, Pei, Yu, Gong, Jingjing, Yu, Qiying, Zhang, Zheng, Wang, Mingxuan, Zhou, Hao, Liu, Jingjing, Ma, Wei-Ying
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912551871184896
author Song, Yuxuan
Zhang, Zhe
Pei, Yu
Gong, Jingjing
Yu, Qiying
Zhang, Zheng
Wang, Mingxuan
Zhou, Hao
Liu, Jingjing
Ma, Wei-Ying
author_facet Song, Yuxuan
Zhang, Zhe
Pei, Yu
Gong, Jingjing
Yu, Qiying
Zhang, Zheng
Wang, Mingxuan
Zhou, Hao
Liu, Jingjing
Ma, Wei-Ying
contents 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
format Preprint
id arxiv_https___arxiv_org_abs_2508_17345
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ShortListing Model: A Streamlined SimplexDiffusion for Discrete Variable Generation
Song, Yuxuan
Zhang, Zhe
Pei, Yu
Gong, Jingjing
Yu, Qiying
Zhang, Zheng
Wang, Mingxuan
Zhou, Hao
Liu, Jingjing
Ma, Wei-Ying
Machine Learning
Genomics
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
title ShortListing Model: A Streamlined SimplexDiffusion for Discrete Variable Generation
topic Machine Learning
Genomics
url https://arxiv.org/abs/2508.17345