Bidirectional Representations Augmented Autoregressive Biological Sequence Generation

preprint · NeurIPS 2025 · 2025

preprint · NeurIPS 2025 · 2025. Xiang Zhang (Shanghai AI Lab) et al. Autoregressive (AR) models, common in sequence generation, are limited in many biological tasks such as de…
Date 2025-10-09
Type preprint
Venue NeurIPS 2025
Publisher arXiv
Contribution algorithm
DOI 10.48550/arXiv.2510.08169
Citations (OpenAlex) 0

Abstract

Autoregressive (AR) models, common in sequence generation, are limited in many biological tasks such as de novo peptide sequencing and protein modeling by their unidirectional nature, failing to capture crucial global bidirectional token dependencies. Non-Autoregressive (NAR) models offer holistic, bidirectional representations but face challenges with generative coherence and scalability. To transcend this, we propose a hybrid framework enhancing AR generation by dynamically integrating rich contextual information from non-autoregressive mechanisms. Our approach couples a shared input encoder with two decoders: a non-autoregressive one learning latent bidirectional biological features, and an AR decoder synthesizing the biological sequence by leveraging these bidirectional features. A novel cross-decoder attention module enables the AR decoder to iteratively query and integrate these bidirectional features, enriching its predictions. This synergy is cultivated via a tailored training strategy with importance annealing for balanced objectives and cross-decoder gradient blocking for stable, focused learning. Evaluations on a demanding nine-species benchmark of de novo peptide sequencing show that our model substantially surpasses AR and NAR baselines. It uniquely harmonizes AR stability with NAR contextual awareness, delivering robust, superior performance on diverse downstream data. This research advances biological sequence modeling techniques and contributes a novel architectural paradigm for augmenting AR models with enhanced bidirectional understanding for complex sequence generation. Code is available at https://github.com/BEAM-Labs/denovo.

Authors

  1. Xiang Zhang (Shanghai AI Lab) · Fudan University, Shanghai Artificial Intelligence Laboratory, University of British Columbia
  2. Jiaqi Wei · Shanghai Artificial Intelligence Laboratory, Zhejiang University
  3. Zijie Qiu · Fudan University, Shanghai Artificial Intelligence Laboratory
  4. Sheng Xu · Fudan University, Shanghai Artificial Intelligence Laboratory
  5. Zhi Jin · Shanghai Artificial Intelligence Laboratory, Soochow University
  6. Zhiqiang Gao · Shanghai Artificial Intelligence Laboratory
  7. Nanqing Dong · Shanghai Artificial Intelligence Laboratory
  8. Siqi Sun · Fudan University, Shanghai Artificial Intelligence Laboratory

Methods and tools

Cites (18)

Seen in the charts

Back to the full map

Back to top