Search Results for author: Yiwen Wang

Found 6 papers, 1 papers with code

Machine unlearning via GAN

no code implementations22 Nov 2021 Kongyang Chen, Yao Huang, Yiwen Wang

Machine learning models, especially deep models, may unintentionally remember information about their training data.

Inference Attack Membership Inference Attack

Subject Enveloped Deep Sample Fuzzy Ensemble Learning Algorithm of Parkinson's Speech Data

no code implementations17 Nov 2021 Yiwen Wang, Fan Li, Xiaoheng Zhang, Pin Wang, Yongming Li

Therefore, it is necessary to reconstruct the existing large segments within one subject into few segments even one segment within one subject, which can facilitate the extraction of relevant speech features to characterize diagnostic markers for the whole subject.

Ensemble Learning Speech Recognition

Lightweight machine unlearning in neural network

no code implementations10 Nov 2021 Kongyang Chen, Yiwen Wang, Yao Huang

Our method only needs to make a small perturbation of the weight of the target model and make it iterate in the direction of the model trained with the remaining data subset until the contribution of the unlearning data to the model is completely eliminated.

Incremental Learning

Controlled Evaluation of Grammatical Knowledge in Mandarin Chinese Language Models

1 code implementation EMNLP 2021 Yiwen Wang, Jennifer Hu, Roger Levy, Peng Qian

We find suggestive evidence that structural supervision helps with representing syntactic state across intervening content and improves performance in low-data settings, suggesting that the benefits of hierarchical inductive biases in acquiring dependency relationships may extend beyond English.

Tracking Fast Neural Adaptation by Globally Adaptive Point Process Estimation for Brain-Machine Interface

no code implementations27 Jul 2021 Shuhang Chen, Xiang Zhang, Xiang Shen, Yifan Huang, Yiwen Wang

In order to identify the active neurons in brain control and track their tuning property changes, we propose a globally adaptive point process method (GaPP) to estimate the neural modulation state from spike trains, decompose the states into the hyper preferred direction and reconstruct the kinematics in a dual-model framework.

Cannot find the paper you are looking for? You can Submit a new open access paper.