Research map / Probabilistic and scientific modeling / Dynamical systems
Expectation Propagation with Stochastic Kinetic Model in Complex Interaction Systems
Le Fang, Fan Yang, Wen Dong, Tong Guan, Chunming Qiao
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 |
|---|---|---|---|
| EP with discrete fixed point (DFP) instead of particle filter | stochastic kinetic model · SynthTown | R2: 0.50 -> 0.85 | Improved |
| DFP instead of extended Kalman filter | stochastic kinetic model · Berlin | R2: 0.45 -> 0.66 | Improved |
| DG gradient ascent instead of RNN | stochastic kinetic model · SynthTown | MSE: 407 -> 104 | Improved |
| DFP instead of FNN | stochastic kinetic model · Berlin | R2: 0.31 -> 0.66 | Improved |
| DFP instead of PF on larger dataset | stochastic kinetic model · Berlin | MSE: 678 -> 20 | Improved |
| scale from SynthTown to Berlin (more locations) | DFP · Berlin vs SynthTown | R2: 0.85 -> 0.66 | Hurt |
| DFP runtime growth with larger dataset vs PF and EKF | DFP · Berlin | runtime: 47 sec -> 29 min | Improved |
| gradient ascent (DG) instead of fixed point (ablation) | DG · SynthTown | MSE: 181 -> 104 | Improved |
| gradient ascent (DG) instead of fixed point (ablation) | DG · Berlin | runtime: 29 min -> 56 min | Hurt |
| Gaussian latent state (GFP) instead of discrete (DFP) (ablation) | GFP · Berlin | runtime: 29 min -> 21 min | Improved |
| Neural network (RNN) instead of FNN | RNN · Berlin | R2: 0.31 -> 0.51 | Improved |
Closest papers by meaning
- Fearless Stochasticity in Expectation Propagation (NeurIPS 2024 · similarity 0.84)
- A Graph Dynamics Prior for Relational Inference (AAAI 2024 · similarity 0.84)
- Learning the Causal Structure of Networked Dynamical Systems under Latent Nodes and Structured Noise (AAAI 2024 · similarity 0.84)
- Latent Representation and Simulation of Markov Processes via Time-Lagged Information Bottleneck (ICLR 2024 · similarity 0.83)
- Benchmarking Structural Inference Methods for Interacting Dynamical Systems with Synthetic Data (NeurIPS 2024 · similarity 0.83)
- Parametric model reduction of mean-field and stochastic systems via higher-order action matching (NeurIPS 2024 · 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.
