2023. 9. 5. 13:59ㆍGoogle ML Bootcamp/1. Neural Networks and Deep Learning
Why is deep learning suddenly working so well?
- over the last 10 years maybe is that for a lot of problems we went from having a relatively small amout of data.
but now, we have a fairly large amout of data and all of this was thanks to the digitization of a society.
for hit this very high level of performance ten wyou need two things.
1. size of Neural Network
2. enough data(=labeled data)
if not enough amout of data, algorithm level is not defined well. so it is really different from personal engineering skills
activate function을 sigmoid에서 ReLU로 바꿈으로써 절편의 기울기가 0이되는 현상을 없앨 수 있었다 = computation이 빨라졌다
- 개발자들은 많은 아이디어를 실험할 수 있게 됨으로써 더 많은 알고리즘이 개발될 수 있는 초석이 될 수 있다.
'Google ML Bootcamp > 1. Neural Networks and Deep Learning' 카테고리의 다른 글
| 6. Logistic Regression Cost Function (0) | 2023.09.08 |
|---|---|
| 5. Logistic Regression (0) | 2023.09.08 |
| 4. Binary Classification (0) | 2023.09.08 |
| 2. Supervised Learning with Neural Networks (0) | 2023.09.05 |
| 1. What is a Neural Network? (0) | 2023.09.05 |