Summary of Conditionally-conjugate Gaussian Process Factor Analysis For Spike Count Data Via Data Augmentation, by Yididiya Y. Nadew et al.
Conditionally-Conjugate Gaussian Process Factor Analysis for Spike Count Data via Data Augmentation
by Yididiya Y. Nadew, Xuhui Fan, Christopher J. Quinn
First submitted to arxiv on: 19 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary Gaussian process factor analysis (GPFA) is a widely used latent variable modeling technique for identifying underlying neural trajectories in high-dimensional recordings. By treating spiking rates as Gaussian observations, GPFA enables tractable inference. Recently, GPFA has been extended to model spike count data. However, the non-conjugacy of the likelihood renders inference intractable, requiring black-box techniques or numerical approximations. To overcome this challenge, we propose conditionally-conjugate Gaussian process factor analysis (ccGPFA), which provides both analytically and computationally tractable inference for modeling neural activity from spike count data. We develop a novel data augmentation method to render the model conditionally conjugate, allowing simple closed-form updates using variational EM algorithms. Additionally, our model can be scaled using sparse Gaussian Processes and accelerated via natural gradients. Empirical experiments demonstrate the efficacy of ccGPFA. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine trying to understand what’s going on in a person’s brain by looking at lots of different measurements. This is hard because there are so many variables involved! Researchers have developed a way called Gaussian process factor analysis (GPFA) to simplify this process and identify the underlying patterns in these measurements. Recently, they’ve been able to use GPFA to analyze data from individual neurons, but this has its own challenges. To overcome these challenges, scientists have come up with a new approach called conditionally-conjugate GPFA. This method allows them to quickly and easily analyze the neural activity data and understand what’s going on in the brain. |
Keywords
» Artificial intelligence » Data augmentation » Inference » Likelihood
