AutoML-Zero: Evolving Machine Learning Algorithms From Scratch
Esteban Real, Chen Liang, David R. So, Quoc V. Le
2020-03-06
:page_with_curl: Abstract(본문)
Machine learning research has advanced in multiple aspects, including model structures and learning methods. The effort to automate such research, known as AutoML, has also made significant progress. However, this progress has largely focused on the architecture of neural networks, where it has relied on sophisticated expert-designed layers as building blocks---or similarly restrictive search spaces. Our goal is to show that AutoML can go further: it is possible today to automatically discover complete machine learning algorithms just using basic mathematical operations as building blocks. We demonstrate this by introducing a novel framework that significantly reduces human bias through a generic search space. Despite the vastness of this space, evolutionary search can still discover two-layer neural networks trained by backpropagation. These simple neural networks can then be surpassed by evolving directly on tasks of interest, e.g. CIFAR-10 variants, where modern techniques emerge in the top algorithms, such as bilinear interactions, normalized gradients, and weight averaging. Moreover, evolution adapts algorithms to different task types: e.g., dropout-like techniques appear when little data is available. We believe these preliminary successes in discovering machine learning algorithms from scratch indicate a promising new direction for the field.
:mag_right: 어떤 논문인지 소개해주세요.
Google brain과 Google Research에서 발표한 새 AutoML 방법에 관한 논문입니다.
그 간의 AutoML은 neural networks의 구조나 제한적인 search space에 머물러 있었습니다. 그러나 본고는 기본적인 수학적 연산으로 완전하고 최적인 기계 학습 알고리즘을 자동으로 발견하도록 함으로써 AutoML이 더 멀리 갈 수 있다는 것을 보여줍니다.
광대한 search space를 탐구하는 방식엔 진화적 알고리즘을 적용했는데, 이 부분은 추후 다른 메타휴리스틱 알고리즘과 결합하여 더 효율적인 탐색이 가능한 새 제안을 생각해 볼 수 있겠습니다.
:clipboard: 논문의 정보를 알려주세요.
:page_with_curl: Abstract(본문)
:mag_right: 어떤 논문인지 소개해주세요.
:key: 핵심 키워드를 적어주세요.
AutoML, ,Genetic algorithm, Search space
:paperclip: URL
https://arxiv.org/abs/2003.03384