Neural Architecture Search with Reinforcement Learning
Citations
20 citations
Cites background from "Neural Architecture Search with Rei..."
...Also, some networks generated by automatic network architecture search can have many feed-forward connections [14], which can substantially (∼10X) increase the size of the activation data....
[...]
..., network architecture search [3, 14] and hyperparameter search [10, 13], reductions in Ops or parameters may not always improve the network efficiency....
[...]
20 citations
Cites background or methods from "Neural Architecture Search with Rei..."
...SIGKDD Explorations Volume 22, Issue 2 43 Figure 9: The RNN controller [102] predicts a sequence of strings for whole network....
[...]
...Automated Deep Learning (AutoDL) AutoKeras [42] NASBot [44] NAS [102] NASNet [103] ENAS [70] LargeEvoNet [74] AgingEvoNet [73] HierEvoNet [59] MorphEvoNet [93] DARTS [60] Proxyless [10] NAONet [61] NASP [95] -...
[...]
...The method [102] following REINFORCE rule expresses the reward gradient ∇θcJ(θc) as follows,...
[...]
...AutoDL has been a quick-evolving subfield in AutoML, since Zoph and Le [102] disclosed the feasibility of neural architecture search using reinforcement learning....
[...]
...One way to train the agent is policy gradient [102]....
[...]
20 citations
20 citations
Cites background from "Neural Architecture Search with Rei..."
...Recent results shows that automatically designed deep learning models can achieve state-of-the-art performance on academic benchmarks (Zoph and Le 2016; Real et al. 2019), as well as offer practical usage (Tan and Le 2019)....
[...]
20 citations
Cites methods from "Neural Architecture Search with Rei..."
...Architectures automatically found by algorithms have achieved highly competitive performance in high-level vision tasks such as image classification [46], object detection [8, 36] and semantic segmentation [17, 29]....
[...]
References
123,388 citations
111,197 citations
55,235 citations
"Neural Architecture Search with Rei..." refers methods in this paper
...Along with this success is a paradigm shift from feature designing to architecture designing, i.e., from SIFT (Lowe, 1999), and HOG (Dalal & Triggs, 2005), to AlexNet (Krizhevsky et al., 2012), VGGNet (Simonyan & Zisserman, 2014), GoogleNet (Szegedy et al., 2015), and ResNet (He et al., 2016a)....
[...]
42,067 citations
31,952 citations
"Neural Architecture Search with Rei..." refers methods in this paper
...Along with this success is a paradigm shift from feature designing to architecture designing, i.e., from SIFT (Lowe, 1999), and HOG (Dalal & Triggs, 2005), to AlexNet (Krizhevsky et al., 2012), VGGNet (Simonyan & Zisserman, 2014), GoogleNet (Szegedy et al., 2015), and ResNet (He et al., 2016a)....
[...]