Summary of Abstracted Shapes As Tokens — a Generalizable and Interpretable Model For Time-series Classification, by Yunshi Wen et al.
Abstracted Shapes as Tokens – A Generalizable and Interpretable Model for Time-series Classification
by Yunshi Wen, Tengfei Ma, Tsui-Wei Weng, Lam M. Nguyen, Anak Agung Julius
First submitted to arxiv on: 1 Nov 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Machine Learning (stat.ML)
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 introduces VQShape, a pre-trained model for time-series representation learning and classification that provides interpretable and generalizable representations. By leveraging vector quantization, the authors demonstrate that different domains’ time-series data can be described using a unified set of low-dimensional codes. These codes can be visualized as abstracted shapes in the time domain, allowing for the development of interpretable classifiers that achieve comparable performance to specialist models. Moreover, VQShape’s pre-trained weights and codebook enable zero-shot learning on previously unseen datasets and domains not included in the training process. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper creates a new way to understand and work with time-series data from different sources. It develops a special model called VQShape that can take in various types of time-series data and produce a simple, understandable representation of each series. This allows for easier interpretation and the creation of accurate classifiers. The model is also good at generalizing to new data it hasn’t seen before, which is important for real-world applications. |
Keywords
» Artificial intelligence » Classification » Quantization » Representation learning » Time series » Zero shot
