Research map / Probabilistic and scientific modeling / Variational and generative sampling
Efficient Non-conjugate Gaussian Process Factor Models for Spike Count Data using Polynomial Approximations
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 second-order orthogonal-polynomial approximations to marginalize non-conjugate count likelihoods | Count-GPFA · Simulated count-GPFA data | rate reconstruction accuracy | Helped |
| Use second-order orthogonal-polynomial approximations to marginalize non-conjugate count likelihoods | Count-GPFA · Simulated count-GPFA data | convergence time | Helped |
| Initialize BBVI hyperparameters with PAL estimates instead of random initialization (ablation) | BBVI for count-GPFA · Simulated count-GPFA data | ELBO optimization stability | Helped |
| Initialize BBVI hyperparameters with PAL estimates instead of random initialization (ablation) | BBVI for count-GPFA · Simulated count-GPFA data | convergence time | Helped |
| Use PAL inference for Poisson-GPFA on high-variance neural activity with abrupt rate changes | Poisson-GPFA · Mouse neural data | cross-validated log-likelihood | Hurt |
| Combine PAL initialization with BBVI rather than using PAL alone | Poisson-GPFA · Mouse neural data | cross-validated log-likelihood | Helped |
| Use PAL initialization with BBVI rather than BBVI alone (ablation) | Count-GPFA · Mouse neural data | cross-validated log-likelihood | Helped |
| Use PAL estimation rather than BBVI for binomial or negative-binomial GPFA | Count-GPFA · Mouse neural data | cross-validated log-likelihood | No clear effect |
| Use PAL, BBVI, or PAL-initialized BBVI for count-GPFA | Count-GPFA · Primate neural data | cross-validated log-likelihood | No clear effect |
| Use binomial-GPFA rather than the other count models | Count-GPFA · Mouse and primate neural data | cross-validated log-likelihood | Helped |
Closest papers by meaning
- Conditionally-Conjugate Gaussian Process Factor Analysis for Spike Count Data via Data Augmentation (ICML 2024 · similarity 0.94)
- Multi-modal Gaussian Process Variational Autoencoders for Neural and Behavioral Data (ICLR 2024 · similarity 0.87)
- Scalable inference of functional neural connectivity at submillisecond timescales (NeurIPS 2025 · similarity 0.86)
- A Differentiable Partially Observable Generalized Linear Model with Forward-Backward Message Passing (ICML 2024 · similarity 0.84)
- Modeling Latent Neural Dynamics with Gaussian Process Switching Linear Dynamical Systems (NeurIPS 2024 · similarity 0.83)
- Turbocharging Gaussian Process Inference with Approximate Sketch-and-Project (NeurIPS 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.