Neural architecture search (NAS) has achieved remarkable results in deep neural network design. Differentiable architecture search converts the search over discrete architectures into a hyperparameter optimization problem which can be solved by gradient descent. However, questions have been raised regarding the effectiveness and generalizability of gradient methods for solving non-convex architecture hyperparameter optimization problems. In this paper, we propose L$^{2}$NAS, which learns to intelligently optimize and update architecture hyperparameters via an actor neural network based on the distribution of high-performing architectures in the search history. We introduce a quantile-driven training procedure which efficiently trains L$^{2}$NAS in an actor-critic framework via continuous-action reinforcement learning. Experiments show that L$^{2}$NAS achieves state-of-the-art results on NAS-Bench-201 benchmark as well as DARTS search space and Once-for-All MobileNetV3 search space. We also show that search policies generated by L$^{2}$NAS are generalizable and transferable across different training datasets with minimal fine-tuning.
We aim to identify how different components in the KD pipeline affect the resulting performance and how much the optimal KD pipeline varies across different datasets/tasks, such as the data augmentation policy, the loss function, and the intermediate representation for transferring the knowledge between teacher and student. To tease apart their effects, we propose Distiller, a meta KD framework that systematically combines a broad range of techniques across different stages of the KD pipeline, which enables us to quantify each component's contribution. Within Distiller, we unify commonly used objectives for distillation of intermediate representations under a universal mutual information (MI) objective and propose a class of MI-$\alpha$ objective functions with better bias/variance trade-off for estimating the MI between the teacher and the student. On a diverse set of NLP datasets, the best Distiller configurations are identified via large-scale hyperparameter optimization. Our experiments reveal the following: 1) the approach used to distill the intermediate representations is the most important factor in KD performance, 2) among different objectives for intermediate distillation, MI-$\alpha$ performs the best, and 3) data augmentation provides a large boost for small training datasets or small student networks. Moreover, we find that different datasets/tasks prefer different KD algorithms, and thus propose a simple AutoDistiller algorithm that can recommend a good KD pipeline for a new dataset.
Algorithm parameters, in particular hyperparameters of machine learning algorithms, can substantially impact their performance. To support users in determining well-performing hyperparameter configurations for their algorithms, datasets and applications at hand, SMAC3 offers a robust and flexible framework for Bayesian Optimization, which can improve performance within a few evaluations. It offers several facades and pre-sets for typical use cases, such as optimizing hyperparameters, solving low dimensional continuous (artificial) global optimization problems and configuring algorithms to perform well across multiple problem instances. The SMAC3 package is available under a permissive BSD-license at https://github.com/automl/SMAC3.
Transformer-based language models such as BERT have outperformed previous models on a large number of English benchmarks, but their evaluation is often limited to English or a small number of well-resourced languages. In this work, we evaluate monolingual, multilingual, and randomly initialized language models from the BERT family on a variety of Uralic languages including Estonian, Finnish, Hungarian, Erzya, Moksha, Karelian, Livvi, Komi Permyak, Komi Zyrian, Northern S\'ami, and Skolt S\'ami. When monolingual models are available (currently only et, fi, hu), these perform better on their native language, but in general they transfer worse than multilingual models or models of genetically unrelated languages that share the same character set. Remarkably, straightforward transfer of high-resource models, even without special efforts toward hyperparameter optimization, yields what appear to be state of the art POS and NER tools for the minority Uralic languages where there is sufficient data for finetuning.
To achieve peak predictive performance, hyperparameter optimization (HPO) is a crucial component of machine learning and its applications. Over the last years, the number of efficient algorithms and tools for HPO grew substantially. At the same time, the community is still lacking realistic, diverse, computationally cheap, and standardized benchmarks. This is especially the case for multi-fidelity HPO methods. To close this gap, we propose HPOBench, which includes 7 existing and 5 new benchmark families, with a total of more than 100 multi-fidelity benchmark problems. HPOBench allows to run this extendable set of multi-fidelity HPO benchmarks in a reproducible way by isolating and packaging the individual benchmarks in containers. It also provides surrogate and tabular benchmarks for computationally affordable yet statistically sound evaluations. To demonstrate HPOBench's broad compatibility with various optimization tools, as well as its usefulness, we conduct an exemplary large-scale study evaluating 13 optimizers from 6 optimization tools. We provide HPOBench here: https://github.com/automl/HPOBench.
When developing and analyzing new hyperparameter optimization methods, it is vital to empirically evaluate and compare them on well-curated benchmark suites. In this work, we propose a new set of challenging and relevant benchmark problems motivated by desirable properties and requirements for such benchmarks. Our new surrogate-based benchmark collection consists of 14 scenarios that in total constitute over 700 multi-fidelity hyperparameter optimization problems, which all enable multi-objective hyperparameter optimization. Furthermore, we empirically compare surrogate-based benchmarks to the more widely-used tabular benchmarks, and demonstrate that the latter may produce unfaithful results regarding the performance ranking of HPO methods. We examine and compare our benchmark collection with respect to defined requirements and propose a single-objective as well as a multi-objective benchmark suite on which we compare 7 single-objective and 7 multi-objective optimizers in a benchmark experiment. Our software is available at [https://github.com/slds-lmu/yahpo_gym].
It is reported that the number of online payment users in China has reached 854 million; with the emergence of community e-commerce platforms, the trend of integration of e-commerce and social applications is increasingly intense. Community e-commerce is not a mature and sound comprehensive e-commerce with fewer categories and low brand value. To effectively retain community users and fully explore customer value has become an important challenge for community e-commerce operators. Given the above problems, this paper uses the data-driven method to study the prediction of community e-commerce customers' repurchase behaviour. The main research contents include 1. Given the complex problem of feature engineering, the classic model RFM in the field of customer relationship management is improved, and an improved model is proposed to describe the characteristics of customer buying behaviour, which includes five indicators. 2. In view of the imbalance of machine learning training samples in SMOTE-ENN, a training sample balance using SMOTE-ENN is proposed. The experimental results show that the machine learning model can be trained more effectively on balanced samples. 3. Aiming at the complexity of the parameter adjustment process, an automatic hyperparameter optimization method based on the TPE method was proposed. Compared with other methods, the model's prediction performance is improved, and the training time is reduced by more than 450%. 4. Aiming at the weak prediction ability of a single model, the soft voting based RF-LightgBM model was proposed. The experimental results show that the RF-LighTGBM model proposed in this paper can effectively predict customer repurchase behaviour, and the F1 value is 0.859, which is better than the single model and previous research results.
Novel technologies in automated machine learning ease the complexity of algorithm selection and hyperparameter optimization. Hyperparameters are important for machine learning models as they significantly influence the performance of machine learning models. Many optimization techniques have achieved notable success in hyperparameter tuning and surpassed the performance of human experts. However, depending on such techniques as blackbox algorithms can leave machine learning practitioners without insight into the relative importance of different hyperparameters. In this paper, we consider building the relationship between the performance of the machine learning models and their hyperparameters to discover the trend and gain insights, with empirical results based on six classifiers and 200 datasets. Our results enable users to decide whether it is worth conducting a possibly time-consuming tuning strategy, to focus on the most important hyperparameters, and to choose adequate hyperparameter spaces for tuning. The results of our experiments show that gradient boosting and Adaboost outperform other classifiers across 200 problems. However, they need tuning to boost their performance. Overall, the results obtained from this study provide a quantitative basis to focus efforts toward guided automated hyperparameter optimization and contribute toward the development of better-automated machine learning frameworks.
The training of neural networks using different deep learning frameworks may lead to drastically differing accuracy levels despite the use of the same neural network architecture and identical training hyperparameters such as learning rate and choice of optimization algorithms. Currently, our ability to build standardized deep learning models is limited by the availability of a suite of neural network and corresponding training hyperparameter benchmarks that expose differences between existing deep learning frameworks. In this paper, we present a living dataset of models and hyperparameters, called CrossedWires, that exposes semantic differences between two popular deep learning frameworks: PyTorch and Tensorflow. The CrossedWires dataset currently consists of models trained on CIFAR10 images using three different computer vision architectures: VGG16, ResNet50 and DenseNet121 across a large hyperparameter space. Using hyperparameter optimization, each of the three models was trained on 400 sets of hyperparameters suggested by the HyperSpace search algorithm. The CrossedWires dataset includes PyTorch and Tensforflow models with test accuracies as different as 0.681 on syntactically equivalent models and identical hyperparameter choices. The 340 GB dataset and benchmarks presented here include the performance statistics, training curves, and model weights for all 1200 hyperparameter choices, resulting in 2400 total models. The CrossedWires dataset provides an opportunity to study semantic differences between syntactically equivalent models across popular deep learning frameworks. Further, the insights obtained from this study can enable the development of algorithms and tools that improve reliability and reproducibility of deep learning frameworks. The dataset is freely available at https://github.com/maxzvyagin/crossedwires through a Python API and direct download link.
From the past few years, due to advancements in technologies, the sedentary living style in urban areas is at its peak. This results in individuals getting a victim of obesity at an early age. There are various health impacts of obesity like Diabetes, Heart disease, Blood pressure problems, and many more. Machine learning from the past few years is showing its implications in all expertise like forecasting, healthcare, medical imaging, sentiment analysis, etc. In this work, we aim to provide a framework that uses machine learning algorithms namely, Random Forest, Decision Tree, XGBoost, Extra Trees, and KNN to train models that would help predict obesity levels (Classification), Bodyweight, and fat percentage levels (Regression) using various parameters. We also applied and compared various hyperparameter optimization (HPO) algorithms such as Genetic algorithm, Random Search, Grid Search, Optuna to further improve the accuracy of the models. The website framework contains various other features like making customizable Diet plans, workout plans, and a dashboard to track the progress. The framework is built using the Python Flask. Furthermore, a weighing scale using the Internet of Things (IoT) is also integrated into the framework to track calories and macronutrients from food intake.
Predicting clinical outcomes from medical images using quantitative features (``radiomics'') requires many method design choices, Currently, in new clinical applications, finding the optimal radiomics method out of the wide range of methods relies on a manual, heuristic trial-and-error process. We introduce a novel automated framework that optimizes radiomics workflow construction per application by standardizing the radiomics workflow in modular components, including a large collection of algorithms for each component, and formulating a combined algorithm selection and hyperparameter optimization problem. To solve it, we employ automated machine learning through two strategies (random search and Bayesian optimization) and three ensembling approaches. Results show that a medium-sized random search and straight-forward ensembling perform similar to more advanced methods while being more efficient. Validated across twelve clinical applications, our approach outperforms both a radiomics baseline and human experts. Concluding, our framework improves and streamlines radiomics research by fully automatically optimizing radiomics workflow construction. To facilitate reproducibility, we publicly release six datasets, software of the method, and code to reproduce this study.
Differentiable architecture search (DAS) is a widely researched tool for the discovery of novel architectures, due to its promising results for image classification. The main benefit of DAS is the effectiveness achieved through the weight-sharing one-shot paradigm, which allows efficient architecture search. In this work, we investigate DAS in a systematic case study of inverse problems, which allows us to analyze these potential benefits in a controlled manner. We demonstrate that the success of DAS can be extended from image classification to signal reconstruction, in principle. However, our experiments also expose three fundamental difficulties in the evaluation of DAS-based methods in inverse problems: First, the results show a large variance in all test cases. Second, the final performance is strongly dependent on the hyperparameters of the optimizer. And third, the performance of the weight-sharing architecture used during training does not reflect the final performance of the found architecture well. While the results on image reconstruction confirm the potential of the DAS paradigm, they challenge the common understanding of DAS as a one-shot method.
Tuning the hyperparameters in the differentially private stochastic gradient descent (DPSGD) is a fundamental challenge. Unlike the typical SGD, private datasets cannot be used many times for hyperparameter search in DPSGD; e.g., via a grid search. Therefore, there is an essential need for algorithms that, within a given search space, can find near-optimal hyperparameters for the best achievable privacy-utility tradeoffs efficiently. We formulate this problem into a general optimization framework for establishing a desirable privacy-utility tradeoff, and systematically study three cost-effective algorithms for being used in the proposed framework: evolutionary, Bayesian, and reinforcement learning. Our experiments, for hyperparameter tuning in DPSGD conducted on MNIST and CIFAR-10 datasets, show that these three algorithms significantly outperform the widely used grid search baseline. As this paper offers a first-of-a-kind framework for hyperparameter tuning in DPSGD, we discuss existing challenges and open directions for future studies. As we believe our work has implications to be utilized in the pipeline of private deep learning, we open-source our code at https://github.com/AmanPriyanshu/DP-HyperparamTuning.
Graphs can be used to describe complex systems. Recently, whole graph embedding (graph representation learning) can compress a graph into a compact lower-dimension vector while preserving intrinsic properties, earning much attention. However, most graph embedding methods have problems such as tedious parameter tuning or poor explanation. This paper presents a simple and hyperparameter-free whole graph embedding method based on the DHC (Degree, H-index, and Coreness) theorem and Shannon Entropy (E), abbreviated as DHC-E. The DHC-E can provide a trade-off between simplicity and quality for supervised classification learning tasks involving molecular, social, and brain networks. Moreover, it performs well in lower-dimensional graph visualization. Overall, the DHC-E is simple, hyperparameter-free, and explainable for whole graph embedding with promising potential for exploring graph classification and lower-dimensional graph visualization.
Bilevel optimization has become a powerful framework in various machine learning applications including meta-learning, hyperparameter optimization, and network architecture search. There are generally two classes of bilevel optimization formulations for machine learning: 1) problem-based bilevel optimization, whose inner-level problem is formulated as finding a minimizer of a given loss function; and 2) algorithm-based bilevel optimization, whose inner-level solution is an output of a fixed algorithm. For the first class, two popular types of gradient-based algorithms have been proposed for hypergradient estimation via approximate implicit differentiation (AID) and iterative differentiation (ITD). Algorithms for the second class include the popular model-agnostic meta-learning (MAML) and almost no inner loop (ANIL). However, the convergence rate and fundamental limitations of bilevel optimization algorithms have not been well explored. This thesis provides a comprehensive convergence rate analysis for bilevel algorithms in the aforementioned two classes. We further propose principled algorithm designs for bilevel optimization with higher efficiency and scalability. For the problem-based formulation, we provide a convergence rate analysis for AID- and ITD-based bilevel algorithms. We then develop acceleration bilevel algorithms, for which we provide shaper convergence analysis with relaxed assumptions. We also provide the first lower bounds for bilevel optimization, and establish the optimality by providing matching upper bounds under certain conditions. We finally propose new stochastic bilevel optimization algorithms with lower complexity and higher efficiency in practice. For the algorithm-based formulation, we develop a theoretical convergence for general multi-step MAML and ANIL, and characterize the impact of parameter selections and loss geometries on the their complexities.
Bilevel optimization has been recently used in many machine learning problems such as hyperparameter optimization, policy optimization, and meta learning. Although many bilevel optimization methods have been proposed, they still suffer from the high computational complexities and do not consider the more general bilevel problems with nonsmooth regularization. In the paper, thus, we propose a class of enhanced bilevel optimization methods with using Bregman distance to solve bilevel optimization problems, where the outer subproblem is nonconvex and possibly nonsmooth, and the inner subproblem is strongly convex. Specifically, we propose a bilevel optimization method based on Bregman distance (BiO-BreD) to solve deterministic bilevel problems, which achieves a lower computational complexity than the best known results. Meanwhile, we also propose a stochastic bilevel optimization method (SBiO-BreD) to solve stochastic bilevel problems based on stochastic approximated gradients and Bregman distance. Moreover, we further propose an accelerated version of SBiO-BreD method (ASBiO-BreD) using the variance-reduced technique, which can achieve a lower computational complexity than the best known computational complexities with respect to condition number $\kappa$ and target accuracy $\epsilon$ for finding an $\epsilon$-stationary point. We conduct data hyper-cleaning task and hyper-representation learning task to demonstrate that our new algorithms outperform related bilevel optimization approaches.
Machine learning algorithms such as random forests or xgboost are gaining more importance and are increasingly incorporated into production processes in order to enable comprehensive digitization and, if possible, automation of processes. Hyperparameters of these algorithms used have to be set appropriately, which can be referred to as hyperparameter tuning or optimization. Based on the concept of tunability, this article presents an overview of theoretical and practical results for popular machine learning algorithms. This overview is accompanied by an experimental analysis of 30 hyperparameters from six relevant machine learning algorithms. In particular, it provides (i) a survey of important hyperparameters, (ii) two parameter tuning studies, and (iii) one extensive global parameter tuning study, as well as (iv) a new way, based on consensus ranking, to analyze results from multiple algorithms. The R package mlr is used as a uniform interface to the machine learning models. The R package SPOT is used to perform the actual tuning (optimization). All additional code is provided together with this paper.
Most machine learning algorithms are configured by one or several hyperparameters that must be carefully chosen and often considerably impact performance. To avoid a time consuming and unreproducible manual trial-and-error process to find well-performing hyperparameter configurations, various automatic hyperparameter optimization (HPO) methods, e.g., based on resampling error estimation for supervised machine learning, can be employed. After introducing HPO from a general perspective, this paper reviews important HPO methods such as grid or random search, evolutionary algorithms, Bayesian optimization, Hyperband and racing. It gives practical recommendations regarding important choices to be made when conducting HPO, including the HPO algorithms themselves, performance evaluation, how to combine HPO with ML pipelines, runtime improvements, and parallelization. This work is accompanied by an appendix that contains information on specific software packages in R and Python, as well as information and recommended hyperparameter search spaces for specific learning algorithms. We also provide notebooks that demonstrate concepts from this work as supplementary files.
Owing to the remarkable capability of extracting effective graph embeddings, graph convolutional network (GCN) and its variants have been successfully applied to a broad range of tasks, such as node classification, link prediction, and graph classification. Traditional GCN models suffer from the issues of overfitting and oversmoothing, while some recent techniques like DropEdge could alleviate these issues and thus enable the development of deep GCN. However, training GCN models is non-trivial, as it is sensitive to the choice of hyperparameters such as dropout rate and learning weight decay, especially for deep GCN models. In this paper, we aim to automate the training of GCN models through hyperparameter optimization. To be specific, we propose a self-tuning GCN approach with an alternate training algorithm, and further extend our approach by incorporating the population based training scheme. Experimental results on three benchmark datasets demonstrate the effectiveness of our approaches on optimizing multi-layer GCN, compared with several representative baselines.
Gaussian process hyperparameter optimization requires linear solves with, and log-determinants of, large kernel matrices. Iterative numerical techniques are becoming popular to scale to larger datasets, relying on the conjugate gradient method (CG) for the linear solves and stochastic trace estimation for the log-determinant. This work introduces new algorithmic and theoretical insights for preconditioning these computations. While preconditioning is well understood in the context of CG, we demonstrate that it can also accelerate convergence and reduce variance of the estimates for the log-determinant and its derivative. We prove general probabilistic error bounds for the preconditioned computation of the log-determinant, log-marginal likelihood and its derivatives. Additionally, we derive specific rates for a range of kernel-preconditioner combinations, showing that up to exponential convergence can be achieved. Our theoretical results enable provably efficient optimization of kernel hyperparameters, which we validate empirically on large-scale benchmark problems. There our approach accelerates training by up to an order of magnitude.