Summary of Learning Under Label Noise Through Few-shot Human-in-the-loop Refinement, by Aaqib Saeed et al.
Learning under Label Noise through Few-Shot Human-in-the-Loop Refinement
by Aaqib Saeed, Dimitris Spathis, Jungwoo Oh, Edward Choi, Ali Etemad
First submitted to arxiv on: 25 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Signal Processing (eess.SP)
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 This paper tackles a significant challenge in wearable technology: obtaining quality labels for various health metrics. Unlike video data, wearable data lacks obvious cues about physical manifestations, making label noise an issue. The proposed Few-Shot Human-in-the-Loop Refinement (FHLR) method addresses this problem by initially learning a seed model using weak labels, then fine-tuning it with expert corrections. FHLR achieves better generalizability and robustness by merging the seed and fine-tuned models via weighted parameter averaging. The approach is evaluated on four challenging tasks and datasets, comparing favorably to eight competitive baselines designed for noisy label learning. FHLR outperforms prior works by a large margin, with up to 19% accuracy improvement under symmetric and asymmetric noise. Notably, it remains robust to increased label noise. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Wearable devices track our physical activity, heart rate, sleep, and stress levels. A big problem is getting the right labels for this data. Unlike videos, wearable data doesn’t have clear signs about what’s happening, so we need extra information to make it accurate. The paper proposes a new way to solve this issue called Few-Shot Human-in-the-Loop Refinement (FHLR). It starts by learning a basic model using some correct labels, then refines it with expert help. FHLR makes the model better and more reliable by combining the original and refined models. The approach is tested on four tricky tasks and datasets, showing that it works really well. |
Keywords
* Artificial intelligence * Few shot * Fine tuning
