改进蛋白质模型推理策略，无需重训练就能显著提升生成效果。
Unlocking Multimodal Protein Language Models at Inference Time
- 设计三阶段框架，系统测试不同推理采样方法
- 在四项任务中实现显著性能提升，最高突破现有上限
- 发现任务特异性采样偏好，挑战原有模型认知
多模态蛋白质语言模型（pLMs）学习蛋白序列与结构的联合分布，其生成性能也高度依赖推理时的采样策略。然而，先前研究更多关注模型训练，而忽视了推理策略的影响。本文建立一个三阶段实证研究框架，跨三种代表性pLMs和四项基础任务，评估原始采样、任务特定无分类器引导以及奖励引导束搜索等策略，分别控制采样分布、每步逻辑值和并行轨迹。通过探索-利用权衡的互补优化，我们（1）揭示默认推理协议的次优性，并识别出任务导向的采样偏好；（2）在各项任务中观察到显著的定量提升，无需更新模型参数即可持续提升多模态pLMs的上限性能；（3）得出关于基础模型的新结论，与以往共识相异。
原文摘要 · Abstract (English)
Multimodal protein language models (pLMs) learn joint protein sequence-structure distributions, and their generation performance should also depend critically on inference-time sampling strategies. Yet prior work has focused more on model training than on how inference-time strategies behave. In this paper, we establish a three-stage investigation framework to empirically study the inference design space of multimodal pLMs across three representative pLMs and four fundamental tasks. We evaluate vanilla sampling, task-specific classifier-free guidance, and reward-guided beam search on multimodal pLMs, corresponding to controls over sampling distributions, per-step logits, and parallel trajectories. Throughout the complementary advancements centered on exploration-exploitation trade-off, we (1) reveal the suboptimality of default inference protocols and identify task-oriented sampling preferences; (2) observe substantial quantitative gains across tasks, consistently boosting the upper bound performance of multimodal pLMs without updating model parameters; (3) derive conclusions about base models that differ from prior consensus.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。