A Ccelerating N Eural a Rchitecture S Earch Us
نویسندگان
چکیده
Methods for neural network hyperparameter optimization and architecture search are computationally expensive due to the need to train a large number of model configurations. In this paper, we show that simple regression models can predict the final performance of partially trained model configurations using features based on network architectures, hyperparameters, and time-series validation performance data. We empirically show that our performance prediction models are much more accurate than prominent Bayesian counterparts, are simpler to implement, and are faster to train. Our models can predict final performance in both visual classification and language modeling domains, are effective for predicting performance of drastically varying model architectures, and can even generalize between model classes. Using these prediction models, we also implement an early stopping method for hyperparameter optimization and architecture search, which obtains a speedup of a factor up to 6x in both hyperparameter optimization and architecture search. Finally, we empirically show that our early stopping method can be seamlessly incorporated into both reinforcement learning-based architecture selection algorithms and bandit based search methods. Through extensive experimentation, we empirically show our performance prediction models and early stopping algorithm are state-of-the-art in terms of prediction accuracy and speedup achieved while still identifying the optimal model configurations.
منابع مشابه
Modeling and Forecasting Effects of Crude Oil Price Changes on the US and UK GDP
       This paper proposes a new forecasting model for investigating relationship between the price of crude oil, as an important energy source and GDP of the US, as the largest oil consumer, and the UK, as the oil producer. GMDH neural network and MLFF neural network approaches, which are both non-linear models, are employed to forecast GDP responses to the oil price changes. The resul...
متن کاملA Genetic Fuzzy Neural Network for Pattern Recognition
In, th,is paper, a gen,etic fu,zzy n,eu,ro,l n,etwork f o r pattern, recognition, is proposed by applyin,g genwtic algorith,ms to th,e Kuian,-Cai f v z z y n,eural network. A genekic-guided self-organizinq learn,in,g a1gorith.m i s CO,pable of redncinuj th,e nmm'ber of f m z y neurons an,d in,creasinq recognxikion rates f o r th,e fixed n,um,ber of outpu,t neurons. Th.e simulataon,s h,ave indic...
متن کاملAmitraz Poisoning; A case study
A m i t r a z, a n i ns e c t i c i d e /a ca ri c i de of the f o r m a m i d i n e p e st i c i d e s group, is a ? 2 a d r e n e r g i c ag on i st a nd of t he a m i d i ne c h e m i ca l f a m il y generally us e d to c o n t r ol animal e c top a r a s i t e s. Poisoning due to am i t r a z i s r a r e and character...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2018