2017. 8. 25. Weekly research update
Aug 25, 2017 · 1 min read
- EEG-based Intention Recognition from Spatio-Temporal Representations via Cascade and Parallel Convolutional Recurrent Neural Networks [https://arxiv.org/abs/1708.06578] : input EEG -> output simple behavior(eye close, hand grasp). Consider the EEG has spatio-temporal data. Used CNN — FC — RNN(LSTM) and showed 99% accuracy.
- Distortion Distribution of Neural Spike Train Sequence Matching with Optogenetics [https://arxiv.org/abs/1708.06641]: Optogenetics implementation to train the neural spike pattern.
- Intonational speech prosody encoding in the human auditory cortex [http://science.sciencemag.org/content/357/6353/797?utm_campaign=toc_sci-mag_2017-08-24&et_rid=34829324&et_cid=1509313] : Intonation play a important role on speech recognition. Could it be implemented to ML architecture?
