Bayesian optimization (BayesOpt) is a robust software extensively used for world optimization duties, corresponding to hyperparameter tuning, protein engineering, artificial chemistry, robotic studying, and even baking cookies. BayesOpt is a superb technique for these issues as a result of all of them contain optimizing black-box capabilities which might be costly to guage. A black-box operate’s underlying mapping from inputs (configurations of the factor we wish to optimize) to outputs (a measure of efficiency) is unknown. Nonetheless, we are able to try to grasp its inner workings by evaluating the operate for various combos of inputs. As a result of every analysis will be computationally costly, we have to discover the perfect inputs in as few evaluations as potential. BayesOpt works by repeatedly developing a surrogate mannequin of the black-box operate and strategically evaluating the operate on the most promising or informative enter location, given the data noticed thus far.
Gaussian processes are in style surrogate fashions for BayesOpt as a result of they’re simple to make use of, will be up to date with new knowledge, and supply a confidence stage about every of their predictions. The Gaussian course of mannequin constructs a likelihood distribution over potential capabilities. This distribution is specified by a imply operate (what these potential capabilities appear like on common) and a kernel operate (how a lot these capabilities can range throughout inputs). The efficiency of BayesOpt relies on whether or not the arrogance intervals predicted by the surrogate mannequin include the black-box operate. Historically, consultants use area data to quantitatively outline the imply and kernel parameters (e.g., the vary or smoothness of the black-box operate) to precise their expectations about what the black-box operate ought to appear like. Nonetheless, for a lot of real-world purposes like hyperparameter tuning, it is rather obscure the landscapes of the tuning goals. Even for consultants with related expertise, it may be difficult to slim down acceptable mannequin parameters.
In “Pre-trained Gaussian processes for Bayesian optimization”, we think about the problem of hyperparameter optimization for deep neural networks utilizing BayesOpt. We suggest Hyper BayesOpt (HyperBO), a extremely customizable interface with an algorithm that removes the necessity for quantifying mannequin parameters for Gaussian processes in BayesOpt. For brand new optimization issues, consultants can merely choose earlier duties which might be related to the present process they’re attempting to unravel. HyperBO pre-trains a Gaussian course of mannequin on knowledge from these chosen duties, and robotically defines the mannequin parameters earlier than working BayesOpt. HyperBO enjoys theoretical ensures on the alignment between the pre-trained mannequin and the bottom reality, in addition to the standard of its options for black-box optimization. We share robust outcomes of HyperBO each on our new tuning benchmarks for close to–state-of-the-art deep studying fashions and traditional multi-task black-box optimization benchmarks (HPO-B). We additionally reveal that HyperBO is powerful to the choice of related duties and has low necessities on the quantity of knowledge and duties for pre-training.
 |
| Within the conventional BayesOpt interface, consultants must fastidiously choose the imply and kernel parameters for a Gaussian course of mannequin. HyperBO replaces this guide specification with a choice of associated duties, making Bayesian optimization simpler to make use of. The chosen duties are used for pre-training, the place we optimize a Gaussian course of such that it will probably progressively generate capabilities which might be just like the capabilities equivalent to these chosen duties. The similarity manifests in particular person operate values and variations of operate values throughout the inputs. |
Loss capabilities for pre-training
We pre-train a Gaussian course of mannequin by minimizing the Kullback–Leibler divergence (a generally used divergence) between the bottom reality mannequin and the pre-trained mannequin. Because the floor reality mannequin is unknown, we can not instantly compute this loss operate. To unravel for this, we introduce two data-driven approximations: (1) Empirical Kullback–Leibler divergence (EKL), which is the divergence between an empirical estimate of the bottom reality mannequin and the pre-trained mannequin; (2) Detrimental log chance (NLL), which is the the sum of unfavorable log likelihoods of the pre-trained mannequin for all coaching capabilities. The computational value of EKL or NLL scales linearly with the variety of coaching capabilities. Furthermore, stochastic gradient–primarily based strategies like Adam will be employed to optimize the loss capabilities, which additional lowers the price of computation. In well-controlled environments, optimizing EKL and NLL result in the identical outcome, however their optimization landscapes will be very totally different. For instance, within the easiest case the place the operate solely has one potential enter, its Gaussian course of mannequin turns into a Gaussian distribution, described by the imply (m) and variance (s). Therefore the loss operate solely has these two parameters, m and s, and we are able to visualize EKL and NLL as follows:
 |
| We simulate the loss landscapes of EKL (left) and NLL (proper) for a easy mannequin with parameters m and s. The colours signify a heatmap of the EKL or NLL values, the place crimson corresponds to larger values and blue denotes decrease values. These two loss landscapes are very totally different, however they each purpose to match the pre-trained mannequin with the bottom reality mannequin. |
Pre-training improves Bayesian optimization
Within the BayesOpt algorithm, selections on the place to guage the black-box operate are made iteratively. The choice standards are primarily based on the arrogance ranges offered by the Gaussian course of, that are up to date in every iteration by conditioning on earlier knowledge factors acquired by BayesOpt. Intuitively, the up to date confidence ranges needs to be good: not overly assured or too uncertain, since in both of those two circumstances, BayesOpt can not make the choices that may match what an skilled would do.
In HyperBO, we change the hand-specified mannequin in conventional BayesOpt with the pre-trained Gaussian course of. Below delicate circumstances and with sufficient coaching capabilities, we are able to mathematically confirm good theoretical properties of HyperBO: (1) Alignment: the pre-trained Gaussian course of ensures to be near the bottom reality mannequin when each are conditioned on noticed knowledge factors; (2) Optimality: HyperBO ensures to discover a near-optimal resolution to the black-box optimization downside for any capabilities distributed in response to the unknown floor reality Gaussian course of.
 |
| We visualize the Gaussian course of (areas shaded in purple are 95% and 99% confidence intervals) conditional on observations (black dots) from an unknown take a look at operate (orange line). In comparison with the normal BayesOpt with out pre-training, the expected confidence ranges in HyperBO captures the unknown take a look at operate a lot better, which is a crucial prerequisite for Bayesian optimization. |
Empirically, to outline the construction of pre-trained Gaussian processes, we select to make use of very expressive imply capabilities modeled by neural networks, and apply well-defined kernel capabilities on inputs encoded to the next dimensional house with neural networks.
To guage HyperBO on difficult and life like black-box optimization issues, we created the PD1 benchmark, which incorporates a dataset for multi-task hyperparameter optimization for deep neural networks. PD1 was developed by coaching tens of 1000’s of configurations of close to–state-of-the-art deep studying fashions on in style picture and textual content datasets, in addition to a protein sequence dataset. PD1 incorporates roughly 50,000 hyperparameter evaluations from 24 totally different duties (e.g., tuning Vast ResNet on CIFAR100) with roughly 12,000 machine days of computation.
We reveal that when pre-training for just a few hours on a single CPU, HyperBO can considerably outperform BayesOpt with fastidiously hand-tuned fashions on unseen difficult duties, together with tuning ResNet50 on ImageNet. Even with solely ~100 knowledge factors per coaching operate, HyperBO can carry out competitively towards baselines.
 |
| Tuning validation error charges of ResNet50 on ImageNet and Vast ResNet (WRN) on the Avenue View Home Numbers (SVHN) dataset and CIFAR100. By pre-training on solely ~20 duties and ~100 knowledge factors per process, HyperBO can considerably outperform conventional BayesOpt (with a fastidiously hand-tuned Gaussian course of) on beforehand unseen duties. |
Conclusion and future work
HyperBO is a framework that pre-trains a Gaussian course of and subsequently performs Bayesian optimization with a pre-trained mannequin. With HyperBO, we now not must hand-specify the precise quantitative parameters in a Gaussian course of. As a substitute, we solely must establish associated duties and their corresponding knowledge for pre-training. This makes BayesOpt each extra accessible and more practical. An essential future path is to allow HyperBO to generalize over heterogeneous search areas, for which we’re growing new algorithms by pre-training a hierarchical probabilistic mannequin.
Acknowledgements
The next members of the Google Analysis Mind Workforce carried out this analysis: Zi Wang, George E. Dahl, Kevin Swersky, Chansoo Lee, Zachary Nado, Justin Gilmer, Jasper Snoek, and Zoubin Ghahramani. We might prefer to thank Zelda Mariet and Matthias Feurer for assist and session on switch studying baselines. We might additionally prefer to thank Rif A. Saurous for constructive suggestions, and Rodolphe Jenatton and David Belanger for suggestions on earlier variations of the manuscript. As well as, we thank Sharat Chikkerur, Ben Adlam, Balaji Lakshminarayanan, Fei Sha and Eytan Bakshy for feedback, and Setareh Ariafar and Alexander Terenin for conversations on animation. Lastly, we thank Tom Small for designing the animation for this put up.
