High-quality arguments are an essential part of decision-making. Automatically predicting the quality of an argument is a complex task that recently got much attention in argument mining. However, the annotation effort for this task is exceptionally high. Therefore, we test uncertainty-based active learning (AL) methods on two popular argument-strength data sets to estimate whether sample-efficient learning can be enabled. Our extensive empirical evaluation shows that uncertainty-based acquisition functions can not surpass the accuracy reached with the random acquisition on these data sets.
Pool-based active learning (AL) aims to optimize the annotation process (i.e., labeling) as the acquisition of annotations is often time-consuming and therefore expensive. For this purpose, an AL strategy queries annotations intelligently from annotators to train a high-performance classification model at a low annotation cost. Traditional AL strategies operate in an idealized framework. They assume a single, omniscient annotator who never gets tired and charges uniformly regardless of query difficulty. However, in real-world applications, we often face human annotators, e.g., crowd or in-house workers, who make annotation mistakes and can be reluctant to respond if tired or faced with complex queries. Recently, a wide range of novel AL strategies has been proposed to address these issues. They differ in at least one of the following three central aspects from traditional AL: (1) They explicitly consider (multiple) human annotators whose performances can be affected by various factors, such as missing expertise. (2) They generalize the interaction with human annotators by considering different query and annotation types, such as asking an annotator for feedback on an inferred classification rule. (3) They take more complex cost schemes regarding annotations and misclassifications into account. This survey provides an overview of these AL strategies and refers to them as real-world AL. Therefore, we introduce a general real-world AL strategy as part of a learning cycle and use its elements, e.g., the query and annotator selection algorithm, to categorize about 60 real-world AL strategies. Finally, we outline possible directions for future research in the field of AL.
Machine learning researchers have long noticed the phenomenon that the model training process will be more effective and efficient when the training samples are densely sampled around the underlying decision boundary. While this observation has already been widely applied in a range of machine learning security techniques, it lacks theoretical analyses of the correctness of the observation. To address this challenge, we first add particular perturbation to original training examples using adversarial attack methods so that the generated examples could lie approximately on the decision boundary of the ML classifiers. We then investigate the connections between active learning and these particular training examples. Through analyzing various representative classifiers such as k-NN classifiers, kernel methods as well as deep neural networks, we establish a theoretical foundation for the observation. As a result, our theoretical proofs provide support to more efficient active learning methods with the help of adversarial examples, contrary to previous works where adversarial examples are often used as destructive solutions. Experimental results show that the established theoretical foundation will guide better active learning strategies based on adversarial examples.
Artificial neural networks (NNs) are one of the most frequently used machine learning approaches to construct interatomic potentials and enable efficient large-scale atomistic simulations with almost ab initio accuracy. However, the simultaneous training of NNs on energies and forces, which are a prerequisite for, e.g., molecular dynamics simulations, can be demanding. In this work, we present an improved NN architecture based on the previous GM-NN model [V. Zaverkin and J. K\"astner, J. Chem. Theory Comput. 16, 5410-5421 (2020)], which shows an improved prediction accuracy and considerably reduced training times. Moreover, we extend the applicability of Gaussian moment-based interatomic potentials to periodic systems and demonstrate the overall excellent transferability and robustness of the respective models. The fast training by the improved methodology is a pre-requisite for training-heavy workflows such as active learning or learning-on-the-fly.
We report a workflow and the output of a natural language processing (NLP)-based procedure to mine the extant metal-organic framework (MOF) literature describing structurally characterized MOFs and their solvent removal and thermal stabilities. We obtain over 2,000 solvent removal stability measures from text mining and 3,000 thermal decomposition temperatures from thermogravimetric analysis data. We assess the validity of our NLP methods and the accuracy of our extracted data by comparing to a hand-labeled subset. Machine learning (ML, i.e. artificial neural network) models trained on this data using graph- and pore-geometry-based representations enable prediction of stability on new MOFs with quantified uncertainty. Our web interface, MOFSimplify, provides users access to our curated data and enables them to harness that data for predictions on new MOFs. MOFSimplify also encourages community feedback on existing data and on ML model predictions for community-based active learning for improved MOF stability models.
In the United States, the parties to a lawsuit are required to search through their electronically stored information to find documents that are relevant to the specific case and produce them to their opposing party. Negotiations over the scope of these searches often reflect a fear that something will be missed (Fear of Missing Out: FOMO). A Recall level of 80%, for example, means that 20% of the relevant documents will be left unproduced. This paper makes the argument that eDiscovery is the process of identifying responsive information, not identifying documents. Documents are the carriers of the information; they are not the direct targets of the process. A given document may contain one or more topics or factoids and a factoid may appear in more than one document. The coupon collector's problem, Heaps law, and other analyses provide ways to model the problem of finding information from among documents. In eDiscovery, however, the parties do not know how many factoids there might be in a collection or their probabilities. This paper describes a simple model that estimates the confidence that a fact will be omitted from the produced set (the identified set), while being contained in the missed set. Two data sets are then analyzed, a small set involving microaggressions and larger set involving classification of web pages. Both show that it is possible to discover at least one example of each available topic within a relatively small number of documents, meaning the further effort will not return additional novel information. The smaller data set is also used to investigate whether the non-random order of searching for responsive documents commonly used in eDiscovery (called continuous active learning) affects the distribution of topics-it does not.
We introduce supervised contrastive active learning (SCAL) and propose efficient query strategies in active learning based on the feature similarity (featuresim) and principal component analysis based feature-reconstruction error (fre) to select informative data samples with diverse feature representations. We demonstrate our proposed method achieves state-of-the-art accuracy, model calibration and reduces sampling bias in an active learning setup for balanced and imbalanced datasets on image classification tasks. We also evaluate robustness of model to distributional shift derived from different query strategies in active learning setting. Using extensive experiments, we show that our proposed approach outperforms high performing compute-intensive methods by a big margin resulting in 9.9% lower mean corruption error, 7.2% lower expected calibration error under dataset shift and 8.9% higher AUROC for out-of-distribution detection.
The environments, in which autonomous cars act, are high-risky, dynamic, and full of uncertainty, demanding a continuous update of their sensory information and knowledge bases. The frequency of facing an unknown object is too high making hard the usage of Artificial Intelligence (AI) classical classification models that usually rely on the close-world assumption. This problem of classifying objects in this domain is better faced with and open-world AI approach. We propose an algorithm to identify not only all the known entities that may appear in front of the car, but also to detect and learn the classes of those unknown objects that may be rare to stand on an highway (e.g., a lost box from a truck). Our approach relies on the DOC algorithm from Lei Shu et. al. as well as on the Query-by-Committee algorithm.
This paper presents simple and efficient methods to mitigate sampling bias in active learning while achieving state-of-the-art accuracy and model robustness. We introduce supervised contrastive active learning by leveraging the contrastive loss for active learning under a supervised setting. We propose an unbiased query strategy that selects informative data samples of diverse feature representations with our methods: supervised contrastive active learning (SCAL) and deep feature modeling (DFM). We empirically demonstrate our proposed methods reduce sampling bias, achieve state-of-the-art accuracy and model calibration in an active learning setup with the query computation 26x faster than Bayesian active learning by disagreement and 11x faster than CoreSet. The proposed SCAL method outperforms by a big margin in robustness to dataset shift and out-of-distribution.
Flow network models can capture the underlying physics and operational constraints of many networked systems including the power grid and transportation and water networks. However, analyzing reliability of systems using computationally expensive flow-based models faces substantial challenges, especially for rare events. Existing actively trained meta-models, which present a new promising direction in reliability analysis, are not applicable to networks due to the inability of these methods to handle high-dimensional problems as well as discrete or mixed variable inputs. This study presents the first adaptive surrogate-based Network Reliability Analysis using Bayesian Additive Regression Trees (ANR-BART). This approach integrates BART and Monte Carlo simulation (MCS) via an active learning method that identifies the most valuable training samples based on the credible intervals derived by BART over the space of predictor variables as well as the proximity of the points to the estimated limit state. Benchmark power grids including IEEE 30, 57, 118, and 300-bus systems and their power flow models for cascading failure analysis are considered to investigate ANR-BART, MCS, subset simulation, and passively-trained optimal deep neural networks and BART. Results indicate that ANR-BART is robust and yields accurate estimates of network failure probability, while significantly reducing the computational cost of reliability analysis.
AstronomicAL is a human-in-the-loop interactive labelling and training dashboard that allows users to create reliable datasets and robust classifiers using active learning. This technique prioritises data that offer high information gain, leading to improved performance using substantially less data. The system allows users to visualise and integrate data from different sources and deal with incorrect or missing labels and imbalanced class sizes. AstronomicAL enables experts to visualise domain-specific plots and key information relating both to broader context and details of a point of interest drawn from a variety of data sources, ensuring reliable labels. In addition, AstronomicAL provides functionality to explore all aspects of the training process, including custom models and query strategies. This makes the software a tool for experimenting with both domain-specific classifications and more general-purpose machine learning strategies. We illustrate using the system with an astronomical dataset due to the field's immediate need; however, AstronomicAL has been designed for datasets from any discipline. Finally, by exporting a simple configuration file, entire layouts, models, and assigned labels can be shared with the community. This allows for complete transparency and ensures that the process of reproducing results is effortless
Labeling data can be an expensive task as it is usually performed manually by domain experts. This is cumbersome for deep learning, as it is dependent on large labeled datasets. Active learning (AL) is a paradigm that aims to reduce labeling effort by only using the data which the used model deems most informative. Little research has been done on AL in a text classification setting and next to none has involved the more recent, state-of-the-art Natural Language Processing (NLP) models. Here, we present an empirical study that compares different uncertainty-based algorithms with BERT$_{base}$ as the used classifier. We evaluate the algorithms on two NLP classification datasets: Stanford Sentiment Treebank and KvK-Frontpages. Additionally, we explore heuristics that aim to solve presupposed problems of uncertainty-based AL; namely, that it is unscalable and that it is prone to selecting outliers. Furthermore, we explore the influence of the query-pool size on the performance of AL. Whereas it was found that the proposed heuristics for AL did not improve performance of AL; our results show that using uncertainty-based AL with BERT$_{base}$ outperforms random sampling of data. This difference in performance can decrease as the query-pool size gets larger.
We propose Cartography Active Learning (CAL), a novel Active Learning (AL) algorithm that exploits the behavior of the model on individual instances during training as a proxy to find the most informative instances for labeling. CAL is inspired by data maps, which were recently proposed to derive insights into dataset quality (Swayamdipta et al., 2020). We compare our method on popular text classification tasks to commonly used AL strategies, which instead rely on post-training behavior. We demonstrate that CAL is competitive to other common AL methods, showing that training dynamics derived from small seed data can be successfully used for AL. We provide insights into our new AL method by analyzing batch-level statistics utilizing the data maps. Our results further show that CAL results in a more data-efficient learning strategy, achieving comparable or better results with considerably less training data.
Physics-informed Machine Learning has recently become attractive for learning physical parameters and features from simulation and observation data. However, most existing methods do not ensure that the physics, such as balance laws (e.g., mass, momentum, energy conservation), are constrained. Some recent works (e.g., physics-informed neural networks) softly enforce physics constraints by including partial differential equation (PDE)-based loss functions but need re-discretization of the PDEs using auto-differentiation. Training these neural nets on observational data showed that one could solve forward and inverse problems in one shot. They evaluate the state variables and the parameters in a PDE. This re-discretization of PDEs is not necessarily an attractive option for domain scientists that work with physics-based codes that have been developed for decades with sophisticated discretization techniques to solve complex process models and advanced equations of state. This paper proposes a physics constrained machine learning framework, AdjointNet, allowing domain scientists to embed their physics code in neural network training workflows. This embedding ensures that physics is constrained everywhere in the domain. Additionally, the mathematical properties such as consistency, stability, and convergence vital to the numerical solution of a PDE are still satisfied. We show that the proposed AdjointNet framework can be used for parameter estimation (and uncertainty quantification by extension) and experimental design using active learning. The applicability of our framework is demonstrated for four flow cases. Results show that AdjointNet-based inversion can estimate process model parameters with reasonable accuracy. These examples demonstrate the applicability of using existing software with no changes in source code to perform accurate and reliable inversion of model parameters.
The key elements of seismic probabilistic risk assessment studies are the fragility curves which express the probabilities of failure of structures conditional to a seismic intensity measure. A multitude of procedures is currently available to estimate these curves. For modeling-based approaches which may involve complex and expensive numerical models, the main challenge is to optimize the calls to the numerical codes to reduce the estimation costs. Adaptive techniques can be used for this purpose, but in doing so, taking into account the uncertainties of the estimates (via confidence intervals or ellipsoids related to the size of the samples used) is an arduous task because the samples are no longer independent and possibly not identically distributed. The main contribution of this work is to deal with this question in a mathematical and rigorous way. To this end, we propose and implement an active learning methodology based on adaptive importance sampling for parametric estimations of fragility curves. We prove some theoretical properties (consistency and asymptotic normality) for the estimator of interest. Moreover, we give a convergence criterion in order to use asymptotic confidence ellipsoids. Finally, the performances of the methodology are evaluated on analytical and industrial test cases of increasing complexity.
Common acquisition functions for active learning use either uncertainty or diversity sampling, aiming to select difficult and diverse data points from the pool of unlabeled data, respectively. In this work, leveraging the best of both worlds, we propose an acquisition function that opts for selecting \textit{contrastive examples}, i.e. data points that are similar in the model feature space and yet the model outputs maximally different predictive likelihoods. We compare our approach, CAL (Contrastive Active Learning), with a diverse set of acquisition functions in four natural language understanding tasks and seven datasets. Our experiments show that CAL performs consistently better or equal than the best performing baseline across all tasks, on both in-domain and out-of-domain data. We also conduct an extensive ablation study of our method and we further analyze all actively acquired datasets showing that CAL achieves a better trade-off between uncertainty and diversity compared to other strategies.
Annotated images are required for both supervised model training and evaluation in image classification. Manually annotating images is arduous and expensive, especially for multi-labeled images. A recent trend for conducting such laboursome annotation tasks is through crowdsourcing, where images are annotated by volunteers or paid workers online (e.g., workers of Amazon Mechanical Turk) from scratch. However, the quality of crowdsourcing image annotations cannot be guaranteed, and incompleteness and incorrectness are two major concerns for crowdsourcing annotations. To address such concerns, we have a rethinking of crowdsourcing annotations: Our simple hypothesis is that if the annotators only partially annotate multi-label images with salient labels they are confident in, there will be fewer annotation errors and annotators will spend less time on uncertain labels. As a pleasant surprise, with the same annotation budget, we show a multi-label image classifier supervised by images with salient annotations can outperform models supervised by fully annotated images. Our method contributions are 2-fold: An active learning way is proposed to acquire salient labels for multi-label images; and a novel Adaptive Temperature Associated Model (ATAM) specifically using partial annotations is proposed for multi-label image classification. We conduct experiments on practical crowdsourcing data, the Open Street Map (OSM) dataset and benchmark dataset COCO 2014. When compared with state-of-the-art classification methods trained on fully annotated images, the proposed ATAM can achieve higher accuracy. The proposed idea is promising for crowdsourcing data annotation. Our code will be publicly available.
Active learning has emerged as a standard paradigm in areas with scarcity of labeled training data, such as in the medical domain. Language models have emerged as the prevalent choice of several natural language tasks due to the performance boost offered by these models. However, in several domains, such as medicine, the scarcity of labeled training data is a common issue. Also, these models may not work well in cases where class imbalance is prevalent. Active learning may prove helpful in these cases to boost the performance with a limited label budget. To this end, we propose a novel method using sampling techniques based on submodular optimization and optimal transport for active learning in language models, dubbed ALLWAS. We construct a sampling strategy based on submodular optimization of the designed objective in the gradient domain. Furthermore, to enable learning from few samples, we propose a novel strategy for sampling from the Wasserstein barycenters. Our empirical evaluations on standard benchmark datasets for text classification show that our methods perform significantly better (>20% relative increase in some cases) than existing approaches for active learning on language models.
Wind farm design primarily depends on the variability of the wind turbine wake flows to the atmospheric wind conditions, and the interaction between wakes. Physics-based models that capture the wake flow-field with high-fidelity are computationally very expensive to perform layout optimization of wind farms, and, thus, data-driven reduced order models can represent an efficient alternative for simulating wind farms. In this work, we use real-world light detection and ranging (LiDAR) measurements of wind-turbine wakes to construct predictive surrogate models using machine learning. Specifically, we first demonstrate the use of deep autoencoders to find a low-dimensional \emph{latent} space that gives a computationally tractable approximation of the wake LiDAR measurements. Then, we learn the mapping between the parameter space and the (latent space) wake flow-fields using a deep neural network. Additionally, we also demonstrate the use of a probabilistic machine learning technique, namely, Gaussian process modeling, to learn the parameter-space-latent-space mapping in addition to the epistemic and aleatoric uncertainty in the data. Finally, to cope with training large datasets, we demonstrate the use of variational Gaussian process models that provide a tractable alternative to the conventional Gaussian process models for large datasets. Furthermore, we introduce the use of active learning to adaptively build and improve a conventional Gaussian process model predictive capability. Overall, we find that our approach provides accurate approximations of the wind-turbine wake flow field that can be queried at an orders-of-magnitude cheaper cost than those generated with high-fidelity physics-based simulations.
Quality control is a key activity performed by manufacturing enterprises to ensure products meet quality standards and avoid potential damage to the brand's reputation. The decreased cost of sensors and connectivity enabled an increasing digitalization of manufacturing. In addition, artificial intelligence enables higher degrees of automation, reducing overall costs and time required for defect inspection. In this research, we compare three active learning approaches and five machine learning algorithms applied to visual defect inspection with real-world data provided by Philips Consumer Lifestyle BV. Our results show that active learning reduces the data labeling effort without detriment to the models' performance.