Active learning aims to reduce annotation cost by predicting which samples are useful for a human expert to label. Although this field is quite old, several important challenges to using active learning in real-world settings still remain unsolved. In particular, most selection strategies are hand-designed, and it has become clear that there is no best active learning strategy that consistently outperforms all others in all applications. This has motivated research into meta-learning algorithms for "learning how to actively learn". In this paper, we compare this kind of approach with the association of a Random Forest with the margin sampling strategy, reported in recent comparative studies as a very competitive heuristic. To this end, we present the results of a benchmark performed on 20 datasets that compares a strategy learned using a recent meta-learning algorithm with margin sampling. We also present some lessons learned and open future perspectives.
Active learning can reduce the number of samples needed to perform a hypothesis test and to estimate the parameters of a model. In this paper, we revisit the work of Chernoff that described an asymptotically optimal algorithm for performing a hypothesis test. We obtain a novel sample complexity bound for Chernoff's algorithm, with a non-asymptotic term that characterizes its performance at a fixed confidence level. We also develop an extension of Chernoff sampling that can be used to estimate the parameters of a wide variety of models and we obtain a non-asymptotic bound on the estimation error. We apply our extension of Chernoff sampling to actively learn neural network models and to estimate parameters in real-data linear and non-linear regression problems, where our approach performs favorably to state-of-the-art methods.
Active learning frameworks offer efficient data annotation without remarkable accuracy degradation. In other words, active learning starts training the model with a small size of labeled data while exploring the space of unlabeled data in order to select most informative samples to be labeled. Generally speaking, representing the uncertainty is crucial in any active learning framework, however, deep learning methods are not capable of either representing or manipulating model uncertainty. On the other hand, from the real world application perspective, uncertainty representation is getting more and more attention in the machine learning community. Deep Bayesian active learning frameworks and generally any Bayesian active learning settings, provide practical consideration in the model which allows training with small data while representing the model uncertainty for further efficient training. In this paper, we briefly survey recent advances in Bayesian active learning and in particular deep Bayesian active learning frameworks.
Deep Gaussian processes (DGPs) are increasingly popular as predictive models in machine learning (ML) for their non-stationary flexibility and ability to cope with abrupt regime changes in training data. Here we explore DGPs as surrogates for computer simulation experiments whose response surfaces exhibit similar characteristics. In particular, we transport a DGP's automatic warping of the input space and full uncertainty quantification (UQ), via a novel elliptical slice sampling (ESS) Bayesian posterior inferential scheme, through to active learning (AL) strategies that distribute runs non-uniformly in the input space -- something an ordinary (stationary) GP could not do. Building up the design sequentially in this way allows smaller training sets, limiting both expensive evaluation of the simulator code and mitigating cubic costs of DGP inference. When training data sizes are kept small through careful acquisition, and with parsimonious layout of latent layers, the framework can be both effective and computationally tractable. Our methods are illustrated on simulation data and two real computer experiments of varying input dimensionality. We provide an open source implementation in the "deepgp" package on CRAN.
Structure-based virtual screening is an important tool in early stage drug discovery that scores the interactions between a target protein and candidate ligands. As virtual libraries continue to grow (in excess of $10^8$ molecules), so too do the resources necessary to conduct exhaustive virtual screening campaigns on these libraries. However, Bayesian optimization techniques can aid in their exploration: a surrogate structure-property relationship model trained on the predicted affinities of a subset of the library can be applied to the remaining library members, allowing the least promising compounds to be excluded from evaluation. In this study, we assess various surrogate model architectures, acquisition functions, and acquisition batch sizes as applied to several protein-ligand docking datasets and observe significant reductions in computational costs, even when using a greedy acquisition strategy; for example, 87.9% of the top-50000 ligands can be found after testing only 2.4% of a 100M member library. Such model-guided searches mitigate the increasing computational costs of screening increasingly large virtual libraries and can accelerate high-throughput virtual screening campaigns with applications beyond docking.
Node classification on graphs is an important task in many practical domains. It usually requires labels for training, which can be difficult or expensive to obtain in practice. Given a budget for labelling, active learning aims to improve performance by carefully choosing which nodes to label. Previous graph active learning methods learn representations using labelled nodes and select some unlabelled nodes for label acquisition. However, they do not fully utilize the representation power present in unlabelled nodes. We argue that the representation power in unlabelled nodes can be useful for active learning and for further improving performance of active learning for node classification. In this paper, we propose a latent space clustering-based active learning framework for node classification (LSCALE), where we fully utilize the representation power in both labelled and unlabelled nodes. Specifically, to select nodes for labelling, our framework uses the K-Medoids clustering algorithm on a latent space based on a dynamic combination of both unsupervised features and supervised features. In addition, we design an incremental clustering module to avoid redundancy between nodes selected at different steps. Extensive experiments on five datasets show that our proposed framework LSCALE consistently and significantly outperforms the stateof-the-art approaches by a large margin.
Skew-Gaussian processes (SkewGPs) extend the multivariate Unified Skew-Normal distributions over finite dimensional vectors to distribution over functions. SkewGPs are more general and flexible than Gaussian processes, as SkewGPs may also represent asymmetric distributions. In a recent contribution we showed that SkewGP and probit likelihood are conjugate, which allows us to compute the exact posterior for non-parametric binary classification and preference learning. In this paper, we generalize previous results and we prove that SkewGP is conjugate with both the normal and affine probit likelihood, and more in general, with their product. This allows us to (i) handle classification, preference, numeric and ordinal regression, and mixed problems in a unified framework; (ii) derive closed-form expression for the corresponding posterior distributions. We show empirically that the proposed framework based on SkewGP provides better performance than Gaussian processes in active learning and Bayesian (constrained) optimization. These two tasks are fundamental for design of experiments and in Data Science.
Humans can leverage hierarchical structures to split a task into sub-tasks and solve problems efficiently. Both imitation and reinforcement learning or a combination of them with hierarchical structures have been proven to be an efficient way for robots to learn complex tasks with sparse rewards. However, in the previous work of hierarchical imitation and reinforcement learning, the tested environments are in relatively simple 2D games, and the action spaces are discrete. Furthermore, many imitation learning works focusing on improving the policies learned from the expert polices that are hard-coded or trained by reinforcement learning algorithms, rather than human experts. In the scenarios of human-robot interaction, humans can be required to provide demonstrations to teach the robot, so it is crucial to improve the learning efficiency to reduce expert efforts, and know human's perception about the learning/training process. In this project, we explored different imitation learning algorithms and designed active learning algorithms upon the hierarchical imitation and reinforcement learning framework we have developed. We performed an experiment where five participants were asked to guide a randomly initialized agent to a random goal in a maze. Our experimental results showed that using DAgger and reward-based active learning method can achieve better performance while saving more human efforts physically and mentally during the training process.
A novel design optimization approach (ActivO) that employs an ensemble of machine learning algorithms is presented. The proposed approach is a surrogate-based scheme, where the predictions of a weak leaner and a strong learner are utilized within an active learning loop. The weak learner is used to identify promising regions within the design space to explore, while the strong learner is used to determine the exact location of the optimum within promising regions. For each design iteration, exploration is done by randomly selecting evaluation points within regions where the weak learner-predicted fitness is high. The global optimum obtained by using the strong learner as a surrogate is also evaluated to enable rapid convergence once the most promising region has been identified. First, the performance of ActivO was compared against five other optimizers on a cosine mixture function with 25 local optima and one global optimum. In the second problem, the objective was to minimize indicated specific fuel consumption of a compression-ignition internal combustion (IC) engine while adhering to desired constraints associated with in-cylinder pressure and emissions. Here, the efficacy of the proposed approach is compared to that of a genetic algorithm, which is widely used within the internal combustion engine community for engine optimization, showing that ActivO reduces the number of function evaluations needed to reach the global optimum, and thereby time-to-design by 80%. Furthermore, the optimization of engine design parameters leads to savings of around 1.9% in energy consumption, while maintaining operability and acceptable pollutant emissions.
Majorly classical Active Learning (AL) approach usually uses statistical theory such as entropy and margin to measure instance utility, however it fails to capture the data distribution information contained in the unlabeled data. This can eventually cause the classifier to select outlier instances to label. Meanwhile, the loss associated with mislabeling an instance in a typical classification task is much higher than the loss associated with the opposite error. To address these challenges, we propose a Cost-Based Bugdet Active Learning (CBAL) which considers the classification uncertainty as well as instance diversity in a population constrained by a budget. A principled approach based on the min-max is considered to minimize both the labeling and decision cost of the selected instances, this ensures a near-optimal results with significantly less computational effort. Extensive experimental results show that the proposed approach outperforms several state-of -the-art active learning approaches.
Active learning generally involves querying the most representative samples for human labeling, which has been widely studied in many fields such as image classification and object detection. However, its potential has not been explored in the more complex instance segmentation task that usually has relatively higher annotation cost. In this paper, we propose a novel and principled semi-supervised active learning framework for instance segmentation. Specifically, we present an uncertainty sampling strategy named Triplet Scoring Predictions (TSP) to explicitly incorporate samples ranking clues from classes, bounding boxes and masks. Moreover, we devise a progressive pseudo labeling regime using the above TSP in semi-supervised manner, it can leverage both the labeled and unlabeled data to minimize labeling effort while maximize performance of instance segmentation. Results on medical images datasets demonstrate that the proposed method results in the embodiment of knowledge from available data in a meaningful way. The extensive quantitatively and qualitatively experiments show that, our method can yield the best-performing model with notable less annotation costs, compared with state-of-the-arts.
Kernel-based machine learning regression algorithms (MLRAs) are potentially powerful methods for being implemented into operational biophysical variable retrieval schemes. However, they face difficulties in coping with large training datasets. With the increasing amount of optical remote sensing data made available for analysis and the possibility of using a large amount of simulated data from radiative transfer models (RTMs) to train kernel MLRAs, efficient data reduction techniques will need to be implemented. Active learning (AL) methods enable to select the most informative samples in a dataset. This letter introduces six AL methods for achieving optimized biophysical variable estimation with a manageable training dataset, and their implementation into a Matlab-based MLRA toolbox for semi-automatic use. The AL methods were analyzed on their efficiency of improving the estimation accuracy of leaf area index and chlorophyll content based on PROSAIL simulations. Each of the implemented methods outperformed random sampling, improving retrieval accuracy with lower sampling rates. Practically, AL methods open opportunities to feed advanced MLRAs with RTM-generated training data for development of operational retrieval models.
In supervised learning, acquiring labeled training data for a predictive model can be very costly, but acquiring a large amount of unlabeled data is often quite easy. Active learning is a method of obtaining predictive models with high precision at a limited cost through the adaptive selection of samples for labeling. This paper explains the basic problem settings of active learning and recent research trends. In particular, research on learning acquisition functions to select samples from the data for labeling, theoretical work on active learning algorithms, and stopping criteria for sequential data acquisition are highlighted. Application examples for material development and measurement are introduced.
An effective way to oppose global warming and mitigate climate change is to electrify our energy sectors and supply their electric power from renewable wind and solar. Spatio-temporal predictions of electric load become increasingly important for planning this transition, while deep learning prediction models provide increasingly accurate predictions for it. The data used for training deep learning models, however, is usually collected at random using a passive learning approach. This naturally results in a large demand for data and associated costs for sensors like smart meters, posing a large barrier for electric utilities in decarbonizing their grids. Here, we test active learning where we leverage additional computation for collecting a more informative subset of data. We show how electric utilities can apply active learning to better distribute smart meters and collect their data for more accurate predictions of load with about half the data compared to when applying passive learning.
Thanks to the tractability of their likelihood, several deep generative models show promise for seemingly straightforward but important applications like anomaly detection, uncertainty estimation, and active learning. However, the likelihood values empirically attributed to anomalies conflict with the expectations these proposed applications suggest. In this paper, we take a closer look at the behavior of distribution densities through the lens of reparametrization and show that these quantities carry less meaningful information than previously thought, beyond estimation issues or the curse of dimensionality. We conclude that the use of these likelihoods for anomaly detection relies on strong and implicit hypotheses, and highlight the necessity of explicitly formulating these assumptions for reliable anomaly detection.
Knowledge distillation (KD) is a general neural network training approach that uses a teacher model to guide the student model. Existing works mainly study KD from the network output side (e.g., trying to design a better KD loss function), while few have attempted to understand it from the input side. Especially, its interplay with data augmentation (DA) has not been well understood. In this paper, we ask: Why do some DA schemes (e.g., CutMix) inherently perform much better than others in KD? What makes a "good" DA in KD? Our investigation from a statistical perspective suggests that a good DA scheme should reduce the covariance of the teacher-student cross-entropy. A practical metric, the stddev of teacher's mean probability (T. stddev), is further presented and well justified empirically. Besides the theoretical understanding, we also introduce a new entropy-based data-mixing DA scheme, CutMixPick, to further enhance CutMix. Extensive empirical studies support our claims and demonstrate how we can harvest considerable performance gains simply by using a better DA scheme in knowledge distillation.
In the recent past, psychological stress has been increasingly observed in humans, and early detection is crucial to prevent health risks. Stress detection using on-device deep learning algorithms has been on the rise owing to advancements in pervasive computing. However, an important challenge that needs to be addressed is handling unlabeled data in real-time via suitable ground truthing techniques (like Active Learning), which should help establish affective states (labels) while also selecting only the most informative data points to query from an oracle. In this paper, we propose a framework with capabilities to represent model uncertainties through approximations in Bayesian Neural Networks using Monte-Carlo (MC) Dropout. This is combined with suitable acquisition functions for active learning. Empirical results on a popular stress and affect detection dataset experimented on a Raspberry Pi 2 indicate that our proposed framework achieves a considerable efficiency boost during inference, with a substantially low number of acquired pool points during active learning across various acquisition functions. Variation Ratios achieves an accuracy of 90.38% which is comparable to the maximum test accuracy achieved while training on about 40% lesser data.
Recently, leveraging pre-trained Transformer based language models in down stream, task specific models has advanced state of the art results in natural language understanding tasks. However, only a little research has explored the suitability of this approach in low resource settings with less than 1,000 training data points. In this work, we explore fine-tuning methods of BERT -- a pre-trained Transformer based language model -- by utilizing pool-based active learning to speed up training while keeping the cost of labeling new data constant. Our experimental results on the GLUE data set show an advantage in model performance by maximizing the approximate knowledge gain of the model when querying from the pool of unlabeled data. Finally, we demonstrate and analyze the benefits of freezing layers of the language model during fine-tuning to reduce the number of trainable parameters, making it more suitable for low-resource settings.
In this paper, we explore how to efficiently combine crowdsourcing and machine intelligence for the problem of document screening, where we need to screen documents with a set of machine-learning filters. Specifically, we focus on building a set of machine learning classifiers that evaluate documents, and then screen them efficiently. It is a challenging task since the budget is limited and there are countless number of ways to spend the given budget on the problem. We propose a multi-label active learning screening specific sampling technique -- objective-aware sampling -- for querying unlabelled documents for annotating. Our algorithm takes a decision on which machine filter need more training data and how to choose unlabeled items to annotate in order to minimize the risk of overall classification errors rather than minimizing a single filter error. We demonstrate that objective-aware sampling significantly outperforms the state of the art active learning sampling strategies.
Unsupervised Domain Adaptation (UDA) aims to bridge the gap between a source domain, where labelled data are available, and a target domain only represented with unlabelled data. If domain invariant representations have dramatically improved the adaptability of models, to guarantee their good transferability remains a challenging problem. This paper addresses this problem by using active learning to annotate a small budget of target data. Although this setup, called Active Domain Adaptation (ADA), deviates from UDA's standard setup, a wide range of practical applications are faced with this situation. To this purpose, we introduce \textit{Stochastic Adversarial Gradient Embedding} (SAGE), a framework that makes a triple contribution to ADA. First, we select for annotation target samples that are likely to improve the representations' transferability by measuring the variation, before and after annotation, of the transferability loss gradient. Second, we increase sampling diversity by promoting different gradient directions. Third, we introduce a novel training procedure for actively incorporating target samples when learning invariant representations. SAGE is based on solid theoretical ground and validated on various UDA benchmarks against several baselines. Our empirical investigation demonstrates that SAGE takes the best of uncertainty \textit{vs} diversity samplings and improves representations transferability substantially.