Hyperparameters are configuration variables controlling the behavior of machine learning algorithms. They are ubiquitous in machine learning and artificial intelligence and the choice of their values determines the effectiveness of systems based on these technologies. Manual hyperparameter search is often time-consuming and becomes infeasible when the number of hyperparameters is large. Automating the search is an important step towards advancing, streamlining, and systematizing machine learning, freeing researchers and practitioners alike from the burden of finding a good set of hyperparameters by trial and error. In this survey, we present a unified treatment of hyperparameter optimization, providing the reader with examples, insights into the state-of-the-art, and numerous links to further reading. We cover the main families of techniques to automate hyperparameter search, often referred to as hyperparameter optimization or tuning, including random and quasi-random search, bandit-, model-, population-, and gradient-based approaches. We further discuss extensions, including online, constrained, and multi-objective formulations, touch upon connections with other fields, such as meta-learning and neural architecture search, and conclude with open questions and future research directions.

Pre-trained language models (PLM), for example BERT or RoBERTa, mark the state-of-the-art for natural language understanding task when fine-tuned on labeled data. However, their large size poses challenges in deploying them for inference in real-world applications, due to significant GPU memory requirements and high inference latency. This paper explores neural architecture search (NAS) for structural pruning to find sub-parts of the fine-tuned network that optimally trade-off efficiency, for example in terms of model size or latency, and generalization performance. We also show how we can utilize more recently developed two-stage weight-sharing NAS approaches in this setting to accelerate the search process. Unlike traditional pruning methods with fixed thresholds, we propose to adopt a multi-objective approach that identifies the Pareto optimal set of sub-networks, allowing for a more flexible and automated compression process.

Large language models (LLMs) mark the state-of-the-art for natural language understanding. However, their large size poses challenges in deploying them for inference in real-world applications, due to significant GPU memory requirements and high inference latency. This paper explores weight-sharing based neural architecture search (NAS) as a form of structural pruning to find sub-parts of the fine-tuned network that optimally trade-off efficiency, for example in terms of model size or latency, and generalization performance. Unlike traditional pruning methods with fixed thresholds, we propose to adopt a multi-objective approach that identifies the Pareto optimal set of sub-networks, allowing for a more flexible and automated compression process. Our NAS approach achieves up to 50% compression with less than 5% performance drop for a fine-tuned BERT model on 7 out of 8 text classification tasks.

Large Language Models (LLM) achieved considerable results on natural language understanding tasks. However, their sheer size causes a large memory consumption or high latency at inference time, which renders deployment on hardware-constrained applications challenging. Neural architecture search (NAS) demonstrated to be a promising framework to automatically design efficient neural network architectures. In this work, we discuss the relationship between NAS and structural pruning and apply multi-objective NAS to compress LLMs while optimizing their performance when fine-tuned on a downstream task. We provide an in-depth analysis of existing NAS strategies and compare different search space definitions. On sentence classification tasks, we can prune the popular BERT model by up to 50% without deteriorating performance.

Bayesian optimization (BO) is a widely popular approach for the hyperparameter optimization (HPO) in machine learning. At its core, BO iteratively evaluates promising configurations until a user-defined budget, such as wall-clock time or number of iterations, is exhausted. While the final performance after tuning heavily depends on the provided budget, it is hard to pre-specify an optimal value in advance. In this work, we propose an effective and intuitive termination criterion for BO that automatically stops the procedure if it is sufficiently close to the global optimum. Our key insight is that the discrepancy between the true objective (predictive performance on test data) and the computable target (validation performance) suggests stopping once the suboptimality in optimizing the target is dominated by the statistical estimation error. Across an extensive range of real-world HPO problems and baselines, we show that our termination criterion achieves a better trade-off between the test performance and optimization time. Additionally, we find that overfitting may occur in the context of HPO, which is arguably an overlooked problem in the literature, and show how our termination criterion helps to mitigate this phenomenon on both small and large datasets.

We present Syne Tune, a library for large-scale distributed hyperparameter optimization (HPO). 1 Syne Tune’s modular architecture allows users to easily switch between different execution backends to facilitate experimentation and makes it easy to contribute new optimization algorithms. To foster reproducible benchmarking, Syne Tune provides an efficient simulator backend and a benchmarking suite, which are essential for large-scale evaluations of distributed asynchronous HPO algorithms on tabulated and surrogate benchmarks. We showcase these functionalities with a range of state-of-the-art gradient-free optimizers, including multi-fidelity and transfer learning approaches on popular benchmarks from the literature. Additionally, we demonstrate the benefits of Syne Tune for constrained and multiobjective HPO applications through two use cases: the former considers hyperparameters that induce fair solutions and the latter automatically selects machine types along with the conventional hyperparameters.

Tuning machine learning models with Bayesian optimization (BO) is a successful strategy to find good hyperparameters. BO defines an iterative procedure where a cross-validated metric is evaluated on promising hyperparameters. In practice, however, an improvement of the validation metric may not translate in better predictive performance on a test set, especially when tuning models trained on small datasets. In other words, unlike conventional wisdom dictates, BO can overfit. In this paper, we carry out the first systematic investigation of overfitting in BO and demonstrate that this issue is serious, yet often overlooked in practice. We propose a novel criterion to early stop BO, which aims to maintain the solution quality while saving the unnecessary iterations that can lead to overfitting. Experiments on real-world hyperparameter optimization problems show that our approach effectively meets these goals and is more adaptive comparing to baselines.

Semantic how the meanings of words change over time — has preoccupied scholars since well before modern linguistics emerged in the late 19th and early 20th century, ushering in a new methodological turn in the study of language change. Compared to changes in sound and grammar, semantic change is the least understood. Ever since, the study of semantic change has progressed steadily, accumulating a vast store of knowledge for over a century, encompassing many languages and language families. Historical linguists also early on realized the potential of computers as research tools, with papers at the very first international conferences in computational linguistics in the 1960s. Such computational studies still tended to be small-scale, method-oriented, and qualitative. However, recent years have witnessed a sea-change in this regard. Big-data empirical quantitative investigations are now coming to the forefront, enabled by enormous advances in storage capability and processing power. Diachronic corpora have grown beyond imagination, defying exploration by traditional manual qualitative methods, and language technology has become increasingly data-driven and semantics-oriented. These developments present a golden opportunity for the empirical study of semantic change over both long and short time spans. A major challenge presently is to integrate the hard-earned knowledge and expertise of traditional historical linguistics with cutting-edge methodology explored primarily in computational linguistics. The idea for the present volume came out of a concrete response to this challenge. The 1st International Workshop on Computational Approaches to Historical Language Change (LChange'19), at ACL 2019, brought together scholars from both fields. This volume offers a survey of this exciting new direction in the study of semantic change, a discussion of the many remaining challenges that we face in pursuing it, and considerably updated and extended versions of a selection of the contributions to the LChange'19 workshop, addressing both more theoretical problems — e.g., discovery of laws of semantic change — and practical applications, such as information retrieval in longitudinal text archives.

Tuning complex machine learning systems is challenging. Machine learning typically requires to set hyperparameters, be it regularization, architecture, or optimization parameters, whose tuning is critical to achieve good predictive performance. To democratize access to machine learning systems, it is essential to automate the tuning. This paper presents Amazon SageMaker Automatic Model Tuning (AMT), a fully managed system for gradient-free optimization at scale. AMT finds the best version of a trained machine learning model by repeatedly evaluating it with different hyperparameter configurations. It leverages either random search or Bayesian optimization to choose the hyperparameter values resulting in the best model, as measured by the metric chosen by the user. AMT can be used with built-in algorithms, custom algorithms, and Amazon SageMaker pre-built containers for machine learning frameworks. We discuss the core functionality, system architecture, our design principles, and lessons learned. We also describe more advanced features of AMT, such as automated early stopping and warm-starting, showing in experiments their benefits to users.

Change and its precondition, variation, are inherent in languages. Over time, new words enter the lexicon, others become obsolete, and existing words acquire new senses. Associating a word’s correct meaning in its historical context is a central challenge in diachronic research. Historical corpora of classical languages, such as Ancient Greek and Latin, typically come with rich metadata, and existing models are limited by their inability to exploit contextual information beyond the document timestamp. While embedding-based methods feature among the current state of the art systems, they are lacking in the interpretative power. In contrast, Bayesian models provide explicit and interpretable representations of semantic change phenomena. In this chapter we build on GASC, a recent computational approach to semantic change based on a dynamic Bayesian mixture model. In this model, the evolution of word senses over time is based not only on distributional information of lexical nature, but also on text genres. We provide a systematic comparison of dynamic Bayesian mixture models for semantic change with state-of-the-art embedding-based models. On top of providing a full description of meaning change over time, we show that Bayesian mixture models are highly competitive approaches to detect binary semantic change in both Ancient Greek and Latin.

Variational inference is a powerful framework, used to approximate intractable posteriors through variational distributions. The de facto standard is to rely on Gaussian variational families, which come with numerous advantages: they are easy to sample from, simple to parametrize, and many expectations are known in closed-form or readily computed by quadrature. In this paper, we view the Gaussian variational approximation problem through the lens of gradient flows. We introduce a flexible and efficient algorithm based on a linear flow leading to a particle-based approximation. We prove that, with a sufficient number of particles, our algorithm converges linearly to the exact solution for Gaussian targets, and a low-rank approximation otherwise. In addition to the theoretical analysis, we show, on a set of synthetic and real-world high-dimensional problems, that our algorithm outperforms existing methods with Gaussian targets while performing on a par with non-Gaussian targets.

XGBoost, a scalable tree boosting algorithm, has proven effective for many prediction tasks of practical interest, especially using tabular datasets. Hyperparameter tuning can further improve the predictive performance, but unlike neural networks, full-batch training of many models on large datasets can be time consuming. Owing to the discovery that (i) there is a strong linear relation between dataset size&training time, (ii) XGBoost models satisfy the ranking hypothesis, and (iii) lower-fidelity models can discover promising hyperparameter configurations, we show that uniform subsampling makes for a simple yet fast baseline to speed up the tuning of large XGBoost models using multi-fidelity hyperparameter optimization with data subsets as the fidelity dimension. We demonstrate the effectiveness of this baseline on large-scale tabular datasets ranging from $15-70\mathrm{GB}$ in size.

In addition to the best model architecture and hyperparameters, a full AutoML solution requires selecting appropriate hardware automatically. This can be framed as a multi-objective optimization problem: there is not a single best hardware configuration but a set of optimal ones achieving different trade-offs between cost and runtime. In practice, some choices may be overly costly or take days to train. To lift this burden, we adopt a multi-objective approach that selects and adapts the hardware configuration automatically alongside neural architectures and their hyperparameters. Our method builds on Hyperband and extends it in two ways. First, we replace the stopping rule used in Hyperband by a non-dominated sorting rule to preemptively stop unpromising configurations. Second, we leverage hyperparameter evaluations from related tasks via transfer learning by building a probabilistic estimate of the Pareto front that finds promising configurations more efficiently than random search. We show in extensive NAS and HPO experiments that both ingredients bring significant speed-ups and cost savings, with little to no impact on accuracy. In three benchmarks where hardware is selected in addition to hyperparameters, we obtain runtime and cost reductions of at least 5.8x and 8.8x, respectively. Furthermore, when applying our multi-objective method to the tuning of hyperparameters only, we obtain a 10\% improvement in runtime while maintaining the same accuracy on two popular NAS benchmarks.

Given the increasing importance of machine learning (ML) in our lives, several algorithmic fairness techniques have been proposed to mitigate biases in the outcomes of the ML models. However, most of these techniques are specialized to cater to a single family of ML models and a specific definition of fairness, limiting their adaptibility in practice. We introduce a general constrained Bayesian optimization (BO) framework to optimize the performance of any ML model while enforcing one or multiple fairness constraints. BO is a model-agnostic optimization method that has been successfully applied to automatically tune the hyperparameters of ML models. We apply BO with fairness constraints to a range of popular models, including random forests, gradient boosting, and neural networks, showing that we can obtain accurate and fair solutions by acting solely on the hyperparameters. We also show empirically that our approach is competitive with specialized techniques that enforce model-specific fairness constraints, and outperforms preprocessing methods that learn fair representations of the input data. Moreover, our method can be used in synergy with such specialized fairness techniques to tune their hyperparameters. Finally, we study the relationship between fairness and the hyperparameters selected by BO. We observe a correlation between regularization and unbiased models, explaining why acting on the hyperparameters leads to ML models that generalize well and are fair.

Bayesian optimization (BO) is a popular method for optimizing expensive-to-evaluate black-box functions. BO budgets are typically given in iterations, which implicitly assumes each evaluation has the same cost. In fact, in many BO applications, evaluation costs vary significantly in different regions of the search space. In hyperparameter optimization, the time spent on neural network training increases with layer size; in clinical trials, the monetary cost of drug compounds vary; and in optimal control, control actions have differing complexities. Cost-constrained BO measures convergence with alternative cost metrics such as time, money, or energy, for which the sample efficiency of standard BO methods is ill-suited. For cost-constrained BO, cost efficiency is far more important than sample efficiency. In this paper, we formulate cost-constrained BO as a constrained Markov decision process (CMDP), and develop an efficient rollout approximation to the optimal CMDP policy that takes both the cost and future iterations into account. We validate our method on a collection of hyperparameter optimization problems as well as a sensor set selection application.

Bayesian optimization (BO) is a popular methodology to tune the hyperparameters of expensive black-box functions. Traditionally, BO focuses on a single task at a time and is not designed to leverage information from related functions, such as tuning performance objectives of the same algorithm across multiple datasets. In this work, we introduce a novel approach to achieve transfer learning across different datasets as well as different objectives. The main idea is to regress the mapping from hyperparameter to objective quantiles with a semi-parametric Gaussian Copula distribution, which provides robustness against different scales or outliers that can occur in different tasks. We introduce two methods to leverage this mapping: a Thompson sampling strategy as well as a Gaussian Copula process using such quantile estimate as a prior. We show that these strategies can combine the estimation of multiple objectives such as latency and accuracy, steering the hyperparameters optimization toward faster predictions for the same level of accuracy. Extensive experiments demonstrate significant improvements over state-of-the-art methods for both hyperparameter optimization and neural architecture search.

In many real-world applications, the performance of machine learning models is evaluated not along a single objective, but across multiple, potentially competing ones. For instance, for a model deciding whether to grant or deny loans, it is critical to make sure decisions are fair and not only accurate. As it is often infeasible to find a single model performing best across all objectives, practitioners are forced to find a trade-off between the individual objectives. While several multi-objective optimization (MO) techniques have been proposed in the machine learning literature (and beyond), little effort has been put towards using MO for hyperparameter optimization (HPO) problems; a task that has gained immense relevance and adoption in recent years. In this paper, we evaluate the suitability of existing MO algorithms for HPO and propose a novel multi-fidelity method for this problem. We evaluate our approach on public datasets with a special emphasis on fairness-motivated applications, and report substantially lower wall-clock times when approximating Pareto frontiers compared to the state-of-the-art.

Bayesian optimization (BO) is a class of global optimization algorithms, suitable for minimizing an expensive objective function in as few function evaluations as possible. While BO budgets are typically given in iterations, this implicitly measures convergence in terms of iteration count and assumes each evaluation has identical cost. In practice, evaluation costs may vary in different regions of the search space. For example, the cost of neural network training increases quadratically with layer size, which is a typical hyperparameter. Cost-aware BO measures convergence with alternative cost metrics such as time, energy, or money, for which vanilla BO methods are unsuited. We introduce Cost Apportioned BO (CArBO), which attempts to minimize an objective function in as little cost as possible. CArBO combines a cost-effective initial design with a cost-cooled optimization phase which depreciates a learned cost model as iterations proceed. On a set of 20 black-box function optimization problems we show that, given the same cost budget, CArBO finds significantly better hyperparameter configurations than competing methods.

Given the increasing importance of machine learning in our lives and the need for algorithmic fairness, several methods have been proposed to measure and mitigate biases in machine learning models. Commonly, these techniques are specialized approaches applied to a single type of model and a specific definition of fairness, limiting their effectiveness in practice. In this paper, we present a general constrained Bayesian optimization (BO) framework to optimize the performance of any black-box machine learning model while enforcing fairness constraints. BO is a class of global optimization algorithms that has been successfully applied to automatically tune the hyperparameters of machine learning models. We apply BO with fairness constraints to a range of popular models, including random forests, gradient boosting and neural networks, showing that we can obtain accurate and fair solutions by acting solely on the hyperparameters. We also show empirically that our approach is competitive with specialized techniques that explicitly enforce fairness constraints during training, and outperforms preprocessing methods that learn unbiased representations of the input data.