Research map / Probabilistic and scientific modeling / Molecules and proteins
scCBGM: Single-Cell Editing via Concept Bottlenecks
Alma Andersson, Aya Abdelsalam Ismail, Edward De Brouwer, Doron Haviv, Tommaso Biancalani, Kyunghyun Cho, Gabriele Scalia, Aicha BenTaieb, Hector Corrada Bravo
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 |
|---|---|---|---|
| use scCBGM (concept bottleneck with skip connections and cross-covariance loss) instead of standard CBGM | scCBGM · Kang PBMC IFN-b, B cells | rMMD: 0.902 -> 0.112 | Improved |
| condition flow matching on scCBGM embeddings instead of observed concepts only (Vanilla-FM) | Flow matching · Kang, Cui, Nault datasets (14 cell groups) | rMMD: improved or matched in 12/14 cell groups | Improved |
| use encode-edit counterfactual procedure instead of decode-only sampling (ablation) | scCBGM-FM · Kang, Cui, Nault datasets | rMMD: Kang B cells 0.106 -> 0.093 | Improved |
| use concept bottleneck head instead of concept embedding head (ablation) | scCBGM · synthetic data, counterfactual MSE | MSE | Improved |
| add cross-covariance loss in place of cosine orthogonality loss (ablation) | scCBGM (CBM family) · synthetic data, counterfactual MSE | MSE | Improved |
| add decoder concept skip connections (ablation) | scCBGM · synthetic data, counterfactual MSE | MSE | Mixed |
| use scCBGM under noisy concept annotations | scCBGM · synthetic data, four noise types | MSE | Improved |
| use scCBGM instead of baselines on Monocytes (CD14) in Kang data | scCBGM · Kang PBMC IFN-b, Monocytes (CD14) | rMMD: Vanilla-FM (edit) 0.394 vs scCBGM-FM (edit) 1.171 | Hurt |
| edit pathway-activity concepts of control stellate cells jointly with simulated TCDD treatment | scCBGM · Nault dataset stellate cells case study | rMMD / expression change | Improved |
Closest papers by meaning
- Multi-Modal and Multi-Attribute Generation of Single Cells with CFGen (ICLR 2025 · similarity 0.86)
- scPilot: Large Language Model Reasoning Toward Automated Single-Cell Analysis and Discovery (NeurIPS 2025 · similarity 0.86)
- A Simple and Comprehensive Benchmark for Single-Cell Transcriptomics (AAAI 2025 · similarity 0.84)
- scSSL-Bench: Benchmarking Self-Supervised Learning for Single-Cell Data (ICML 2025 · similarity 0.84)
- Cradle-VAE: Enhancing Single-Cell Gene Perturbation Modeling with Counterfactual Reasoning-based Artifact Disentanglement (AAAI 2025 · similarity 0.84)
- Semi-supervised Knowledge Transfer Across Multi-omic Single-cell Data (NeurIPS 2024 · 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.
