Summary of A Differentiable Partially Observable Generalized Linear Model with Forward-backward Message Passing, by Chengrui Li et al.
A Differentiable Partially Observable Generalized Linear Model with Forward-Backward Message Passing
by Chengrui Li, Weihan Li, Yule Wang, Anqi Wu
First submitted to arxiv on: 2 Feb 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Neurons and Cognition (q-bio.NC)
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 The partially observable generalized linear model (POGLM) is a powerful tool for understanding neural connectivity under the assumption of existing hidden neurons. A limitation of previous works using variational inference to learn POGLM lies in the difficulty of learning this latent variable model due to two main issues: the sampled Poisson hidden spike count hinders the use of the pathwise gradient estimator, and the existing design of the variational model is neither expressive nor time-efficient. To address these limitations, we propose a new differentiable POGLM that enables the pathwise gradient estimator, which outperforms the score function gradient estimator used in previous works. We also introduce the forward-backward message-passing sampling scheme for the variational model. Experimental results demonstrate that our differentiable POGLMs with our forward-backward message passing produce better performance on one synthetic and two real-world datasets, while also yielding more interpretable parameters. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary POGLM is a tool used in neuroscience to understand how neurons connect with each other. Researchers have tried to use this method before but it’s been tricky because they can’t see all the connections between neurons. They need a way to learn about these hidden connections from just looking at what happens when some of the neurons are active. The problem is that the math used in these attempts doesn’t work very well and takes too long. To solve this, scientists came up with new ways to do the math and test them on different datasets. They found that their new methods worked better than before and gave more useful results. |
Keywords
* Artificial intelligence * Inference
