Research map / Continual learning / Class-incremental
Gene Incremental Learning for Single-Cell Transcriptomics
Jiaxin Qi, Yan Cui, Jianqiang Huang, Gaogang Xie
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 |
|---|---|---|---|
| train only on new-stage genes without any knowledge-preserving method (baseline) over stages | gene incremental learning baseline · 2-stage GIL regression | regression loss forgetting: 0.279 average drop in regression | Hurt |
| add gene replay (1,000 samples) | baseline GIL · 2-stage GIL regression | regression loss forgetting: 0.279 -> 0.037 | Improved |
| add gene distillation (lambda 5.0) | baseline GIL · 2-stage GIL regression | regression loss forgetting: 0.279 -> 0.212 | Improved |
| add gene distillation (lambda 5.0) | baseline GIL · 2-stage GIL downstream classification | accuracy change (%): -1.816 -> -2.473 | Hurt |
| add gene replay (1,000 samples) | baseline GIL · 2-stage GIL downstream classification | accuracy change (%): -1.816 -> -1.075 | Improved |
| add gene replay or distillation | baseline GIL · 3-stage GIL regression (ICol-Myel-Panc) | regression loss forgetting: 0.181 -> 0.028 (replay); 0.181 -> 0.177 (distill) | Mixed |
| increase number of replayed samples from 50 to 104 (ablation) | Gene Replay · Norman-Lupus regression | forgetting delta: 0.121 -> 0.018 | Improved |
| increase distillation coefficient lambda from 0.5 to 10.0 (ablation) | Gene Distillation · Norman-Lupus regression | forgetting delta: 0.248 -> 0.154 | Improved |
| increase distillation coefficient lambda from 0.5 to 10.0 (ablation) | Gene Distillation · Norman-Lupus new-gene (Lupus) regression loss | regression loss on stage-2 genes: 0.134 -> 0.143 | Hurt |
Closest papers by meaning
- CellPLM: Pre-training of Cell Language Model Beyond Single Cells (ICLR 2024 · similarity 0.87)
- A Simple and Comprehensive Benchmark for Single-Cell Transcriptomics (AAAI 2025 · similarity 0.87)
- ViTally Consistent: Scaling Biological Representation Learning for Cell Microscopy (ICML 2025 · similarity 0.85)
- CellVerse: Do Large Language Models Really Understand Cell Biology? (NeurIPS 2025 · similarity 0.85)
- sciLaMA: A Single-Cell Representation Learning Framework to Leverage Prior Knowledge from Large Language Models (ICML 2025 · similarity 0.84)
- How Post-Training Shapes Biological Reasoning Models (NeurIPS 2026 · similarity 0.84)
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.
