lxcnju / fedkws-ui

FedKWS-UI

Source Code for Interspeech 2022 paper: Avoid Overfitting User Specific Information in Federated Keyword Spotting

Dataset

• Download Speech Commands dataset: Torchaudio SC
• Preprocess: preprocess.py
• Load Data: speech_commands_data.py and speech_commands_feddata.py

Running Tips

• Run FedAvg, FedMMD, FedProx, FedOpt, FedKWS-UI (Ours) with train_xxx.py.
• Algorithm code: fedavg.py, fedkwsui.py, etc.

Citation

• Xin-Chun Li, Jin-Lin Tang, Shaoming Song, Bingshuai Li, Yinchuan Li, Yunfeng Shao, Le Gan, De-Chuan Zhan. Avoid Overfitting User Specific Information in Federated Keyword Spotting. In: Proceedings of the 23rd INTERSPEECH Conference (INTERSPEECH'2022), online conference, Songdo ConvensiA, Incheon, Korea, 2022.
• [Bibtex]