Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Neuronal Gaussian Process Regression
Johannes Friedrich
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 |
|---|---|---|---|
| BioNN (sparse GP-style network with local rules) vs VFE | BioNN · Boston Housing | predictive log likelihood: -2.483 -> -2.605 | Hurt |
| BioNN vs VFE | BioNN · Energy Efficiency | predictive log likelihood: -0.712 -> -0.729 | No clear effect |
| BioNN vs VFE | BioNN · Kin8nm | predictive log likelihood: 0.972 -> 1.031 | Improved |
| BioNN vs VFE | BioNN · Naval Propulsion | predictive log likelihood: 8.800 -> 9.059 | Improved |
| Replace VFE weights with ANN weights trained by Adam (NLL loss, 40 passes) from GP initialization | sparse GP network · Naval Propulsion | predictive log likelihood: 8.800 -> 9.171 | Improved |
| ANN vs MC Dropout and PBP | ANN · UCI datasets | predictive log likelihood | Improved |
| BioNN vs FITC approximation to full GP | BioNN · Concrete Strength | KL(p||q): 69,425.31 -> 29.71 | Improved |
| BioNN vs VFE approximation to full GP | BioNN · Boston Housing | KL(p||q): 15.37 -> 29.05 | Hurt |
| BioNN vs VFE approximation to full GP | BioNN · Energy Efficiency | KL(p||q): 4.79 -> 5.24 | No clear effect |
| Increase number of inducing points/neurons to an unnecessarily large number (ablation) | BioNN · Snelson | NLPD | Hurt |
| Learn noise level from output y in network instead of VFE/FITC predictive variance | BioNN · Snelson, varied noise | noise variance estimate / NLPD | Improved |
| Optimize tuning curve centers (and widths) with REINFORCE | BioNN · Snelson | test RMSE / NLPD | Improved |
Closest papers by meaning
- Estimating Nonlinear Neural Response Functions using GP Priors and Kronecker Methods (NeurIPS 2016 · similarity 0.88)
- A sampling-based circuit for optimal decision making (NeurIPS 2021 · similarity 0.86)
- A simple normative network approximates local non-Hebbian learning in the cortex (NeurIPS 2020 · similarity 0.85)
- Controlled maximal variability along with reliable performance in recurrent neural networks (NeurIPS 2024 · similarity 0.85)
- Modeling Latent Neural Dynamics with Gaussian Process Switching Linear Dynamical Systems (NeurIPS 2024 · similarity 0.84)
- Modeling Dynamic Neural Activity by combining Naturalistic Video Stimuli and Stimulus-independent Latent Factors (NeurIPS 2025 · 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.
