























Authors:Rong Fu, Muge Qi, Yang Li, Yabin Jin, Jiekai Wu, Jiaxuan Lu, Chunlei Meng, Youjin Wang, Zeli Su, Juntao Gao, Li Bao, Qi Zhao, Wei Luo, Simon Fong
Abstract:Repertoire-level analysis of T cell receptors offers a biologically grounded signal for disease detection and immune monitoring, yet practical deployment is impeded by label sparsity, cohort heterogeneity, and the computational burden of adapting large encoders to new tasks. We introduce a framework that synthesizes compact task-specific parameterizations from a learned dictionary of prototypes conditioned on lightweight task descriptors derived from repertoire probes and pooled embedding statistics. This synthesis produces small adapter modules applied to a frozen pretrained backbone, enabling immediate adaptation to novel tasks with only a handful of support examples and without full model fine-tuning. The architecture preserves interpretability through motif-aware probes and a calibrated motif discovery pipeline that links predictive decisions to sequence-level signals. Together, these components yield a practical, sample-efficient, and interpretable pathway for translating repertoire-informed models into diverse clinical and research settings where labeled data are scarce and computational resources are constrained.
| Comments: | 19 pages, 8 figures, 8 tables |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2602.01051 [cs.LG] |
| (or arXiv:2602.01051v4 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2602.01051 arXiv-issued DOI via DataCite |
From: Rong Fu [view email]
[v1]
Sun, 1 Feb 2026 06:30:31 UTC (7,139 KB)
[v2]
Sat, 14 Feb 2026 03:30:44 UTC (7,139 KB)
[v3]
Tue, 3 Mar 2026 10:09:10 UTC (7,140 KB)
[v4]
Wed, 22 Apr 2026 02:07:48 UTC (1,013 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理,仅供阅读参考。 原文来自 — 版权归原作者所有。