Despite their prevalence in deep-learning communities, over-parameterized models convey high demands of computational costs for proper training. This work studies the fine-grained, modular-level learning dynamics of over-parameterized models to attain a more efficient and fruitful training strategy. Empirical evidence reveals that when scaling down into network modules, such as heads in self-attention models, we can observe varying learning patterns implicitly associated with each module's trainability. To describe such modular-level learning capabilities, we introduce a novel concept dubbed modular neural tangent kernel (mNTK), and we demonstrate that the quality of a module's learning is tightly associated with its mNTK's principal eigenvalue $\lambda_{\max}$. A large $\lambda_{\max}$ indicates that the module learns features with better convergence, while those miniature ones may impact generalization negatively. Inspired by the discovery, we propose a novel training strategy termed Modular Adaptive Training (MAT) to update those modules with their $\lambda_{\max}$ exceeding a dynamic threshold selectively, concentrating the model on learning common features and ignoring those inconsistent ones. Unlike most existing training schemes with a complete BP cycle across all network modules, MAT can significantly save computations by its partially-updating strategy and can further improve performance. Experiments show that MAT nearly halves the computational cost of model training and outperforms the accuracy of baselines.

The prevalence of artificial intelligence (AI) has envisioned an era of healthcare democratisation that promises every stakeholder a new and better way of life. However, the advancement of clinical AI research is significantly hurdled by the dearth of data democratisation in healthcare. To truly democratise data for AI studies, challenges are two-fold: 1. the sensitive information in clinical data should be anonymised appropriately, and 2. AI-oriented clinical knowledge should flow freely across organisations. This paper considers a recent deep-learning advent, dataset condensation (DC), as a stone that kills two birds in democratising healthcare data. The condensed data after DC, which can be viewed as statistical metadata, abstracts original clinical records and irreversibly conceals sensitive information at individual levels; nevertheless, it still preserves adequate knowledge for learning deep neural networks (DNNs). More favourably, the compressed volumes and the accelerated model learnings of condensed data portray a more efficient clinical knowledge sharing and flowing system, as necessitated by data democratisation. We underline DC's prospects for democratising clinical data, specifically electrical healthcare records (EHRs), for AI research through experimental results and analysis across three healthcare datasets of varying data types.

Metric learning aims to learn a distance metric such that semantically similar instances are pulled together while dissimilar instances are pushed away. Many existing methods consider maximizing or at least constraining a distance "margin" that separates similar and dissimilar pairs of instances to guarantee their performance on a subsequent k-nearest neighbor classifier. However, such a margin in the feature space does not necessarily lead to robustness certification or even anticipated generalization advantage, since a small perturbation of test instance in the instance space could still potentially alter the model prediction. To address this problem, we advocate penalizing small distance between training instances and their nearest adversarial examples, and we show that the resulting new approach to metric learning enjoys a larger certified neighborhood with theoretical performance guarantee. Moreover, drawing on an intuitive geometric insight, the proposed new loss term permits an analytically elegant closed-form solution and offers great flexibility in leveraging it jointly with existing metric learning methods. Extensive experiments demonstrate the superiority of the proposed method over the state-of-the-arts in terms of both discrimination accuracy and robustness to noise.

Active learning aims to reduce annotation cost by predicting which samples are useful for a human teacher to label. However it has become clear there is no best active learning algorithm. Inspired by various philosophies about what constitutes a good criteria, different algorithms perform well on different datasets. This has motivated research into ensembles of active learners that learn what constitutes a good criteria in a given scenario, typically via multi-armed bandit algorithms. Though algorithm ensembles can lead to better results, they overlook the fact that not only does algorithm efficacy vary across datasets, but also during a single active learning session. That is, the best criteria is non-stationary. This breaks existing algorithms' guarantees and hampers their performance in practice. In this paper, we propose dynamic ensemble active learning as a more general and promising research direction. We develop a dynamic ensemble active learner based on a non-stationary multi-armed bandit with expert advice algorithm. Our dynamic ensemble selects the right criteria at each step of active learning. It has theoretical guarantees, and shows encouraging results on $13$ popular datasets.

Active learning (AL) aims to enable training high performance classifiers with low annotation cost by predicting which subset of unlabelled instances would be most beneficial to label. The importance of AL has motivated extensive research, proposing a wide variety of manually designed AL algorithms with diverse theoretical and intuitive motivations. In contrast to this body of research, we propose to treat active learning algorithm design as a meta-learning problem and learn the best criterion from data. We model an active learning algorithm as a deep neural network that inputs the base learner state and the unlabelled point set and predicts the best point to annotate next. Training this active query policy network with reinforcement learning, produces the best non-myopic policy for a given dataset. The key challenge in achieving a general solution to AL then becomes that of learner generalisation, particularly across heterogeneous datasets. We propose a multi-task dataset-embedding approach that allows dataset-agnostic active learners to be trained. Our evaluation shows that AL algorithms trained in this way can directly generalise across diverse problems.

In this paper, a general Maximum K-Min approach for classification is proposed. With the physical meaning of optimizing the classification confidence of the K worst instances, Maximum K-Min Gain/Minimum K-Max Loss (MKM) criterion is introduced. To make the original optimization problem with combinational constraints computationally tractable, the optimization techniques are adopted and a general compact representation lemma for MKM Criterion is summarized. Based on the lemma, a Nonlinear Maximum K-Min (NMKM) classifier and a Semi-supervised Maximum K-Min (SMKM) classifier are presented for traditional classification task and semi-supervised classification task respectively. Based on the experiment results of publicly available datasets, our Maximum K-Min methods have achieved competitive performance when comparing against Hinge Loss classifiers.

In this paper, a general Maximum K-Min approach for classification is proposed, which focuses on maximizing the gain obtained by the K worst-classified instances while ignoring the remaining ones. To make the original optimization problem with combinational constraints computationally tractable,  the optimization techniques are adopted and a general compact representation lemma is summarized. Based on the lemma, a Nonlinear Maximum K -Min (NMKM) classifier is presented and the experiment results demonstrate the superior performance of the Maximum K -Min Approach.