Summary of Knowledge-data Fusion Based Source-free Semi-supervised Domain Adaptation For Seizure Subtype Classification, by Ruimin Peng et al.
Knowledge-Data Fusion Based Source-Free Semi-Supervised Domain Adaptation for Seizure Subtype Classification
by Ruimin Peng, Jiayu An, Dongrui Wu
First submitted to arxiv on: 29 Nov 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
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 Knowledge-Data Fusion based SF-SSDA approach, KDF-MutualSHOT, leverages both raw EEG data and expert knowledge to enhance the accuracy of seizure subtype classification. The model consists of a feature-driven Decision Tree-based component and a data-driven Transformer-based component, which are mutually learned using Jensen-Shannon Divergence. To adapt this model to a new target dataset, an SF-SSDA algorithm, MutualSHOT, is developed, featuring a consistency-based pseudo-label selection strategy. Experimental results on the TUSZ and CHSZ datasets demonstrate that KDF-MutualSHOT outperforms other supervised and source-free domain adaptation approaches in cross-subject seizure subtype classification. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary EEG-based seizure subtype classification can improve clinical diagnosis efficiency. A new approach, KDF-MutualSHOT, combines expert knowledge with raw EEG data to classify seizures more accurately. The model uses two different methods to analyze the data: a decision tree and a transformer. This helps the model learn from both the structure of the brain signals (EEG) and the patterns in the data (Transformer). To use this model on new data, the MutualSHOT algorithm adjusts it based on how well it agrees with expert opinions. Tests showed that KDF-MutualSHOT did better than other approaches at classifying seizure types. |
Keywords
* Artificial intelligence * Classification * Decision tree * Domain adaptation * Supervised * Transformer
