trigger word detection (sometimes also called keyword detection, or wake word detection)
Trigger word detection is the technology that allows devices like Amazon Alexa, Google Home, Apple Siri, and Baidu DuerOS to wake up upon hearing a certain word.
y_hat when that that clip has finished saying "activate".
Rather than updating the label of a single time step, we will update several(50) steps of the output to have target label 1.
That updating several consecutive time steps can make the training data more balanced.
5 Sequence Models - 3 Sequence models & Attention mechanism
Various sequence to sequence architectures
Basic Models
Encoder-decoder model
Sequence to sequence model
예don't
want arandomly
chosen translation, you maybewant
the most likely
translationPicking the most likely sentence
conditional language model
Beam Search
Refinements to Beam Search
a normalized log
likelihood objective.production-> log (sum)the number of words
in your translationError analysis in beam search
Bleu Score (optional)
Attention Model Intuition
기존 구조의 한계
Attention Model 아이디어
Attention Model
Speech recognition - Audio data
Speech recognition
spectrograms
음성데이터에 대해 생각할 수 있는 한가지 neural network model![image](https://user-images.githubusercontent.com/21363169/89144010-7fd22080-d587-11ea-83c5-b149e35f0758.png)
CTC (Connectionist temporal classification)
Trigger Word Detection
trigger word detection (sometimes also called keyword detection, or wake word detection)
y_hat when that that clip has finished saying "activate".
Rather than updating the label of a single time step, we will update several(50) steps of the output to have target label 1.