Research map / Probabilistic and scientific modeling / Dynamical systems
Independent Component Discovery in Temporal Count Data
Alexandre Chaussard, Anna Bonnet, Sylvain Le Corff
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 |
|---|---|---|---|
| fit ARPLN-ICA with a count-specific emission instead of linear ICA (UwedgeICA, Picard) on log-counts | mixing recovery · simulated mixing recovery (three scenarios) | average cosine similarity: 46% and 44% (baselines) vs 74.6% and 84.5% (ARPLN-ICA, moderate and high coherence) | Improved |
| use structured auto-regressive variational family instead of mean-field (ablation) | ARPLN-ICA · simulated data, fit to oracle distribution | sliced Wasserstein | Improved |
| use structured auto-regressive variational family instead of mean-field (ablation) | ARPLN-ICA · simulated data, mixing recovery | cosine similarity | No clear effect |
| increase number of training trajectories n (ablation) | ARPLN-ICA · simulated moderate coherence | cosine similarity | Improved |
| increase column coherence or reduce excitation (more zeros) in the simulated mixing (ablation) | ARPLN-ICA · simulated mixing recovery | cosine similarity | Hurt |
| add regime switching (C=2) over time-homogeneous ARPLN-ICA (ablation) | ARPLN-ICA · microbiome forecasting (gnotobiotic mice, LOO CV) | MAE log1p, Aitchison, Poisson deviance | Improved |
| ARPLN-ICA versus forecasting baselines (CLR-VAR(1), gLV-L2, P-rSLDS, CLR-ICA) | microbiome forecasting · gnotobiotic mice, LOO CV | Poisson deviance | Improved |
Closest papers by meaning
- Variational Inference for sparse network reconstruction from count data (ICML 2019 · similarity 0.86)
- Deep Poisson gamma dynamical systems (NeurIPS 2018 · similarity 0.85)
- Robust and Scalable Models of Microbiome Dynamics (ICML 2018 · similarity 0.85)
- Learning Interpretable Hierarchical Dynamical Systems Models from Time Series Data (ICLR 2025 · similarity 0.84)
- Discovering Mixtures of Structural Causal Models from Time Series Data (ICML 2024 · similarity 0.83)
- Generating Hypotheses of Dynamic Causal Graphs in Neuroscience: Leveraging Generative Factor Models of Observed Time Series (ICML 2025 · similarity 0.83)
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.
