Research map / Federated learning and privacy / Personalized and heterogeneous FL
FedDNA: DNA Sequence Reconstruction via Deep Evidential Learning and Personalized Federated Aggregation
Haiyan Lin, Qi Shen, Fei Zhu, Zixuan Qin, Liu Yang, Yuping Duan
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add post-hoc evidence fusion over reads in a cluster (ablation) | FedDNA baseline · DNA datasets avg | Success Rate (%): 68.97 -> 92.30 | Improved |
| add deep evidential learning (ablation) | FedDNA with PHF · DNA datasets avg | Success Rate (%): 92.30 -> 96.64 | Improved |
| add uncertainty-decomposed federated aggregation weights (ablation) | FedDNA with PHF+DEL · DNA datasets avg | Success Rate (%): 96.64 -> 97.91 | Improved |
| add uncertainty-decomposed federated aggregation weights (ablation) | FedDNA with PHF+DEL · PE AYB | Success Rate (%): 95.81 -> 96.89 | Improved |
| add uncertainty-decomposed federated aggregation weights (ablation) | FedDNA with PHF+DEL · P10_5 | Success Rate (%): 97.86 -> 97.96 | No clear effect |
| train with federated learning instead of centralized training (ablation) | FedDNA · Seq 1D | Success Rate (%): 95.01 -> 97.26 | Improved |
| federated training instead of centralized (ablation) | FedDNA · PE AYB | Success Rate (%): 94.53 -> 96.89 | Improved |
| FedDNA aggregation vs FedEvi | FedDNA · PE AYB | Edit Error Rate (%): 0.30 -> 0.23 | Hurt |
| FedDNA vs FedEvi aggregation | FedDNA · Seq 1D | Success Rate (%): 95.09 -> 97.26 | Improved |
| FedDNA vs classical FedAvg | FedDNA · DNA datasets avg | Success Rate (%): 92.30 -> 97.91 | Improved |
| FedDNA vs Iterative Reconstruction (non-neural) | FedDNA · Seq 1D | Success Rate (%): 97.45 vs 97.26 | Hurt |
Closest papers by meaning
- Model Decides How to Tokenize: Adaptive DNA Sequence Tokenization with MxDNA (NeurIPS 2024 · similarity 0.83)
- Levenshtein Distance Embedding with Poisson Regression for DNA Storage (AAAI 2024 · similarity 0.81)
- Revisiting Convolution Architecture in the Realm of DNA Foundation Models (ICLR 2025 · similarity 0.80)
- GeneZip: Region-Aware Compression for Long Context DNA Modeling (NeurIPS 2026 · similarity 0.80)
- JanusDNA: A Powerful Bi-directional Hybrid DNA Foundation Model (NeurIPS 2025 · similarity 0.80)
- What DNA Foundation Models Learn Beyond Sequence Composition (NeurIPS 2026 · similarity 0.79)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
