Automated Machine Learning - Methods, Systems, Challenges

2019

1,493 citations recorded by the source corpus.

Citations in the corpus

• FedNAS: Federated Deep Learning via Neural Architecture Search (2020, background): “L G ] 18 A p r 2020 significantly outperforms many manually designed architectures [6].”
• Benchmarking Deep Learning Models and Automated Model Design for COVID-19 Detection with Chest CT Scans (2020, methodology): “In recent years, Automated Machine Learning (AutoML) has created many SOTA results by automatically searching model architectures and hyper-parameters for specific tasks [22], [23], [29].”
• Automated Machine Learning: From Principles to Practices (2018, methodology): “To overcome this problem, the gradient-based method [22, 23, 87] is proposed for one-shot NAS.”
• The Challenge of Machine Learning in Space Weather: Nowcasting and Forecasting (2019, unknown): “Hence, promising techniques use genetic algorithm to make different networks compete, in search of the most performing one for a given task (Hutter et al., 2019).”
• Deep Networks from the Principle of Rate Reduction (2020, background): “…even take to the extreme by searching for effective network structures and training strategies through extensive random search techniques, such as Neural Architecture Search (Zoph & Le, 2017; Baker et al., 2017), AutoML (Hutter et al., 2019), and Learning to Learn (Andrychowicz et al., 2016).”
• Integrating Scientific Knowledge with Machine Learning for Engineering and Environmental Systems (2020, background): “Because of this, there is increasing interest in automated neural architecture search methods [20, 85, 126].”
• A Survey on Deep Learning Techniques for Stereo-Based Depth Estimation (2020, unknown): “[149] show how to use and extend existing AutoML techniques [150] to efficiently optimize large-scale U-Netlike encoder-decoder architectures for stereo-based depth estimation.”
• Grid search in hyperparameter optimization of machine learning models for prediction of HIV/AIDS test results (2021, background): “GS operates by assessing the Cartesian product of a finite set of values defined by the user [24].”
• Optimizer Benchmarking Needs to Account for Hyperparameter Tuning (2019, unknown): “This has led to calls for using automatic hyperparameter optimization methods (HPO) with a fixed budget for a fairer comparison of models (Sculley et al., 2018; Hutter et al., 2019b; Eggensperger et al., 2019).”
• Efficient Neural Architecture Search via Proximal Iterations (2019, methodology): “As NAS can be seen as a black-box optimization problem (Yao et al. 2018; Hutter, Kotthoff, and Vanschoren 2018), here, we bring the wisdom of constraint optimization to deal with the NAS problem.”
• Automated Machine Learning in Practice: State of the Art and Recent Results (2019, background): “models without (or as little as possible) human intervention [3].”
• Automatic Machine Learning by Pipeline Synthesis using Model-Based Reinforcement Learning and a Grammar (2019, methodology): “There are Automatic machine learning (AutoML) (Hutter et al., 2019) problems of increasing difficulty, starting with hyperparameter optimization of a specific algorithm, to the selection of algorithms and their hyperparameter optimization, and finally meta learning, which entails synthesizing an entire…”