Summary of Flexible Heteroscedastic Count Regression with Deep Double Poisson Networks, by Spencer Young et al.
Flexible Heteroscedastic Count Regression with Deep Double Poisson Networks
by Spencer Young, Porter Jenkins, Lonchao Da, Jeff Dotson, Hua Wei
First submitted to arxiv on: 13 Jun 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 The proposed Deep Double Poisson Network (DDPN) aims to address the limitations of existing discrete models for real-world applications that require accurate, input-conditional uncertainty representations. Recent advances in heteroscedastic continuous regression have shown promise for calibrated uncertainty quantification on complex tasks like image regression, but these methods tend to produce pathologies when applied to discrete regression tasks such as crowd counting or ratings prediction. To address this, DDPN is designed to produce discrete predictive distributions of arbitrary flexibility, outperforming existing discrete models and meeting or exceeding the accuracy and flexibility of networks trained with Gaussian negative log likelihood. The proposed technique also enables tuning of mean fit and probabilistic calibration during training. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper proposes a new model called Deep Double Poisson Network (DDPN) that can be used for real-world applications like crowd counting, ratings prediction, or inventory estimation. This model is better than existing models because it can produce more accurate and flexible predictions. The authors also show how to train the model to get the best results. |
Keywords
* Artificial intelligence * Log likelihood * Regression
