It is well known in the field of machine learning that committee models improve accuracy, provide generalization error estimates, and enable active learning strategies. In this work, we adapt these concepts to interatomic potentials based on artificial neural networks. Instead of a single model, multiple models that share the same atomic environment descriptors yield an average that outperforms its individual members as well as a measure of the generalization error in the form of the committee disagreement. We not only use this disagreement to identify the most relevant configurations to build up the model's training set in an active learning procedure, but also monitor and bias it during simulations to control the generalization error. This facilitates the adaptive development of committee neural network potentials and their training sets, while keeping the number of ab initio calculations to a minimum. To illustrate the benefits of this methodology, we apply it to the development of a committee model for water in the condensed phase. Starting from a single reference ab initio simulation, we use active learning to expand into new state points and to describe the quantum nature of the nuclei. The final model, trained on 814 reference calculations, yields excellent results under a range of conditions, from liquid water at ambient and elevated temperatures and pressures to different phases of ice, and the air-water interface - all including nuclear quantum effects. This approach to committee models will enable the systematic development of robust machine learning models for a broad range of systems.
The committee machine: Computational to statistical gaps in learning a two-layers neural network
Institut de Physique Théorique, CNRS & CEA & Université Paris-Saclay（法国国家科学研究中心、法国原子能和替代能源委员会及巴黎萨克雷大学理论物理研究所）；Laboratoire de Physique Statistique, CNRS & Sorbonnes Universités & École Normale Supérieure, PSL University（法国国家科学研究中心、索邦大学及巴黎高等师范学院统计物理实验室）；Laboratoire de Théorie des Communications, École Polytechnique Fédérale de Lausanne（洛桑联邦理工学院通信理论实验室）；International Center for Theoretical Physics（国际理论物理中心）
Kyushu Institute of Technology（九州工业大学）；The Institute of Statistical Mathematics（统计数学研究所）；RIKEN AIP（理化学研究所人工智能研究中心）
Active learning is a framework in which the learning machine can select the samples to be used for training. This technique is promising, particularly when the cost of data acquisition and labeling is high. In active learning, determining the timing at which learning should be stopped is a critical issue. In this study, we propose a criterion for automatically stopping active learning. The proposed stopping criterion is based on the difference in the expected generalization errors and hypothesis testing. We derive a novel upper bound for the difference in expected generalization errors before …
Multi-head committees enable direct uncertainty prediction for atomistic foundation models
Charles University, Faculty of Mathematics and Physics（查理大学数学物理学院）；Cavendish Laboratory, Department of Physics, University of Cambridge（剑桥大学物理系卡文迪许实验室）；Lennard-Jones Centre, University of Cambridge（剑桥大学伦纳德-琼斯中心）
Delft University of Technology（代尔夫特理工大学）；Radboud University Nijmegen（拉德堡德大学奈梅亨分校）；University of Copenhagen（哥本哈根大学）
Active learning algorithms propose which unlabeled objects should be queried for their labels to improve a predictive model the most. We study active learners that minimize generalization bounds and uncover relationships between these bounds that lead to an improved approach to active learning. In particular we show the relation between the bound of the state-of-the-art Maximum Mean Discrepancy (MMD) active learner, the bound of the Discrepancy, and a new and looser bound that we refer to as the Nuclear Discrepancy bound. We motivate this bound by a probabilistic argument: we show it considers…
Conventional active learning algorithms assume a single labeler that produces noiseless label at a given, fixed cost, and aim to achieve the best generalization performance for given classifier under a budget constraint. However, in many real settings, different labelers have different labeling costs and can yield different labeling accuracies. Moreover, a given labeler may exhibit different labeling accuracies for different instances. This setting can be referred to as active learning with diverse labelers with varying costs and accuracies, and it arises in many important real settings. It is…
Institute of Theoretical and Applied Informatics, Polish Academy of Sciences（波兰科学院理论与应用信息学研究所）；Yaşar University（亚沙尔大学）
During the COVID-19 pandemic, a massive number of attempts on the predictions of the number of cases and the other future trends of this pandemic have been made. However, they fail to predict, in a reliable way, the medium and long term evolution of fundamental features of COVID-19 outbreak within acceptable accuracy. This paper gives an explanation for the failure of machine learning models in this particular forecasting problem. The paper shows that simple linear regression models provide high prediction accuracy values reliably but only for a 2-weeks period and that relatively complex machi…
Google Research（谷歌研究院）；Courant Institute of Mathematical Sciences（库朗数学科学研究所）；New York University（纽约大学）
We present a new active learning algorithm that adaptively partitions the input space into a finite number of regions, and subsequently seeks a distinct predictor for each region, both phases actively requesting labels. We prove theoretical guarantees for both the generalization error and the label complexity of our algorithm, and analyze the number of regions defined by the algorithm under some mild assumptions. We also report the results of an extensive suite of experiments on several real-world datasets demonstrating substantial empirical benefits over existing single-region and non-adaptiv…
Kyushu Institute of Technology（九州工业大学）；The Institute of Statistical Mathematics（统计数理研究所）；RIKEN AIP（理化学研究所革新智能统合研究中心）
Active learning is a framework for supervised learning to improve the predictive performance by adaptively annotating a small number of samples. To realize efficient active learning, both an acquisition function that determines the next datum and a stopping criterion that determines when to stop learning should be considered. In this study, we propose a stopping criterion based on error stability, which guarantees that the change in generalization error upon adding a new sample is bounded by the annotation cost and can be applied to any Bayesian active learning. We demonstrate that the propose…
We address the problem of active learning under label shift: when the class proportions of source and target domains differ. We introduce a "medial distribution" to incorporate a tradeoff between importance weighting and class-balanced sampling and propose their combined usage in active learning. Our method is known as Mediated Active Learning under Label Shift (MALLS). It balances the bias from class-balanced sampling and the variance from importance weighting. We prove sample complexity and generalization guarantees for MALLS which show active learning reduces asymptotic sample complexity ev…
Active learning from demonstration allows a robot to query a human for specific types of input to achieve efficient learning. Existing work has explored a variety of active query strategies; however, to our knowledge, none of these strategies directly minimize the performance risk of the policy the robot is learning. Utilizing recent advances in performance bounds for inverse reinforcement learning, we propose a risk-aware active inverse reinforcement learning algorithm that focuses active queries on areas of the state space with the potential for large generalization error. We show that risk-…
University of Minnesota, Twin Cities（明尼苏达大学双城分校）；University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）
In spite of advances in understanding lazy training, recent work attributes the practical success of deep learning to the rich regime with complex inductive bias. In this paper, we study rich regime training empirically with benchmark datasets, and find that while most parameters are lazy, there is always a small number of active parameters which change quite a bit during training. We show that re-initializing (resetting to their initial random values) the active parameters leads to worse generalization. Further, we show that most of the active parameters are in the bottom layers, close to the…
The workhorse of machine learning is stochastic gradient descent. To access stochastic gradients, it is common to consider iteratively input/output pairs of a training dataset. Interestingly, it appears that one does not need full supervision to access stochastic gradients, which is the main motivation of this paper. After formalizing the "active labeling" problem, which focuses on active learning with partial supervision, we provide a streaming technique that provably minimizes the ratio of generalization error over the number of samples. We illustrate our technique in depth for robust regres…
University of Florida（佛罗里达大学）；University of North Carolina at Chapel Hill（北卡罗来纳大学教堂山分校）；Los Alamos National Laboratory（洛斯阿拉莫斯国家实验室）
The development of accurate and transferable machine learning (ML) potentials for predicting molecular energetics is a challenging task. The process of data generation to train such ML potentials is a task neither well understood nor researched in detail. In this work, we present a fully automated approach for the generation of datasets with the intent of training universal ML potentials. It is based on the concept of active learning (AL) via Query by Committee (QBC), which uses the disagreement between an ensemble of ML potentials to infer the reliability of the ensemble's prediction. QBC all…
Several pool-based active learning algorithms (AL) were employed to model potential energy surfaces (PESs) with a minimum number of electronic structure calculations. Theoretical and empirical results suggest that superior strategies can be obtained by sampling molecular structures corresponding to large uncertainties in their predictions while at the same time not deviating much from the true distribution of the data. To model PESs in an AL framework we propose to use a regression version of stochastic query by forest, a hybrid method that samples points corresponding to large uncertainties w…
