arXiv：cs.LG· Ziyuan Gao·· 3 小时前
视觉类增量学习方法能否迁移到蛋白质？冻结蛋白质编码器的诊断性基准
Do Vision CIL Methods Transfer to Protein? A Diagnostic Benchmark of Frozen Protein Encoders
AI 导读
一项诊断性基准在五种序列、结构和序列-结构蛋白质编码器与两个不同规模的酶功能数据集上，对九种代表性 CIL 方法进行了受控评测，编码器全程冻结。结果显示没有任何方法能在不同编码器间稳定迁移，且失败在替代阶段构造下依然存在。诊断分析将其归因于基于梯度的稳定性信号偏弱、嵌入几何的显著异质性，以及协方差假设的合适容量随数据集规模和编码器变化，说明表征本身应被视为持续学习设计的核心组件。
正文
Abstract:Protein function annotation operates over an expanding label space, making class-incremental learning (CIL) a natural framework for incorporating newly discovered functions without full retraining. Yet existing CIL methods have been developed largely in vision and have not been systematically evaluated over frozen protein representations. We conduct a controlled benchmark of nine representative CIL methods across five sequence-, structure-, and sequence--structure-based protein encoders and two enzyme function datasets at different scales, while keeping the encoders frozen and evaluation settings consistent across representations. We find that no method transfers reliably across encoders, and that these failures persist across alternative phase constructions. Diagnostic analyses reveal that this instability arises from weak gradient-based stability signals, substantial heterogeneity in embedding geometry, and covariance assumptions whose appropriate capacity changes with both dataset scale and encoder. These findings suggest that the representation itself should be treated as a central component of continual-learning design rather than a passive feature extractor. More generally, reliable protein CIL will require encoder-aware methods that adapt their statistical assumptions and capacity to the geometry of the representation on which they operate. Code is available at this https URL.
| Comments: | Accepted at NeurIPS 2026 Workshop on Continual Learning for Foundation Models and Agents (CL4FM Agents) |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Quantitative Methods (q-bio.QM) |
| Cite as: | arXiv:2610.18879 [cs.LG] |
| (or arXiv:2610.18879v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.18879 arXiv-issued DOI via DataCite |
Submission history
From: Ziyuan Gao [view email]
[v1]
Thu, 1 Oct 2026 02:12:27 UTC (4,971 KB)
来源：arXiv：cs.LG · arxiv.org