Summary of Advancing Sleep Detection by Modelling Weak Label Sets: a Novel Weakly Supervised Learning Approach, By Matthias Boeker et al.
Advancing sleep detection by modelling weak label sets: A novel weakly supervised learning approach
by Matthias Boeker, Vajira Thambawita, Michael Riegler, Pål Halvorsen, Hugo L. Hammer
First submitted to arxiv on: 27 Feb 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 A novel approach for sleep detection using weakly supervised learning is introduced, which relies on a set of weak labels derived from conventional sleep detection algorithms. The proposed method models the number of weak sleep labels as an outcome of a binomial distribution, linking it to neural networks trained to detect sleep based on actigraphy. This framework maximizes the likelihood function, equivalent to minimizing soft cross-entropy loss. Additionally, the Brier score is explored as a loss function for weak labels. The study demonstrates the efficacy of this framework using the Multi-Ethnic Study of Atherosclerosis dataset, showing that an LSTM trained on soft cross-entropy outperforms other neural networks and loss functions in accuracy and model calibration. This research advances sleep detection techniques in scenarios with scarce ground truth data and contributes to weakly supervised learning by introducing innovative approaches for modelling sets of weak labels. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study introduces a new way to detect when people are sleeping or awake, even when we don’t have perfect information about when they’re actually asleep. It’s like trying to figure out what someone is thinking without directly asking them! The scientists used special algorithms that predict sleep patterns and then used those predictions to create a better system for detecting sleep. They tested this new approach on a big dataset of people’s activity levels and found it worked really well. This research could help us understand sleep patterns better, which is important for our physical and mental health. |
Keywords
* Artificial intelligence * Cross entropy * Likelihood * Loss function * Lstm * Supervised
