Neural architecture search (NAS) finds high performing networks for a given task. Yet the results of NAS are fairly prosaic; they did not e.g. create a shift from convolutional structures to transformers. This is not least because the search spaces in NAS often aren’t diverse enough to include such transformations *a priori*. Instead, for NAS to provide greater potential for fundamental design shifts, we need a novel expressive search space design which is built from more fundamental operations. To this end, we introduce `einspace`, a search space based on a parameterised probabilistic context-free grammar. Our space is versatile, supporting architectures of various sizes and complexities, while also containing diverse network operations which allow it to model convolutions, attention components and more. It contains many existing competitive architectures, and provides flexibility for discovering new ones. Using this search space, we perform experiments to find novel architectures as well as improvements on existing ones on the diverse Unseen NAS datasets. We show that competitive architectures can be obtained by searching from scratch, and we consistently find large improvements when initialising the search with strong baselines. We believe that this work is an important advancement towards a transformative NAS paradigm where search space expressivity and strategic search initialisation play key roles.

In some medical imaging tasks and other settings where only small parts of the image are informative for the classification task, traditional CNNs can sometimes struggle to generalise. Manually annotated Regions of Interest (ROI) are often used to isolate the most informative parts of the image. However, these are expensive to collect and may vary significantly across annotators. To overcome these issues, we propose a framework that employs saliency maps to obtain soft spatial attention masks that modulate the image features at different scales. We refer to our method as Adversarial Counterfactual Attention (ACAT). ACAT increases the baseline classification accuracy of lesions in brain CT scans from $71.39 %$ to $72.55 %$ and of COVID-19 related findings in lung CT scans from $67.71 %$ to $70.84 %$ and exceeds the performance of competing methods. We investigate the best way to generate the saliency maps employed in our architecture and propose a way to obtain them from adversarially generated counterfactual images. They are able to isolate the area of interest in brain and lung CT scans without using any manual annotations. In the task of localising the lesion location out of 6 possible regions, they obtain a score of $65.05 %$ on brain CT scans, improving the score of $61.29 %$ obtained with the best competing method.

The field of meta-learning, or learning-to-learn, has seen a dramatic rise in interest in recent years. Contrary to conventional approaches to AI where tasks are solved from scratch using a fixed learning algorithm, meta-learning aims to improve the learning algorithm itself, given the experience of multiple learning episodes. This paradigm provides an opportunity to tackle many conventional challenges of deep learning, including data and computation bottlenecks, as well as generalization. This survey describes the contemporary meta-learning landscape. We first discuss definitions of meta-learning and position it with respect to related fields, such as transfer learning and hyperparameter optimization. We then propose a new taxonomy that provides a more comprehensive breakdown of the space of meta-learning methods today. We survey promising applications and successes of meta-learning such as few-shot learning and reinforcement learning. Finally, we discuss outstanding challenges and promising areas for future research.

In few-shot learning, a machine learning system is required to learn from a small set of labelled examples of a specific task, such that it can achieve strong generalization on new unlabelled examples of the same task. Given the limited availability of labelled examples in such tasks, we need to make use of all the information we can. For this reason we propose the use of transductive meta-learning for few shot settings to obtain state-of-the-art few-shot learning. Usually a model learns task-specific information from a small training-set (the \emph{support-set}) and subsequently produces predictions on a small unlabelled validation set (\emph{target-set}). The target-set contains additional task-specific information which is not utilized by existing few-shot learning methods. This is a challenge requiring approaches beyond the current methods as at inference time, the target-set contains only input data-points, and so discriminative-based learning cannot be used. In this paper, we propose a framework called \emph{Self-Critique and Adapt} or SCA. This approach learns to learn a label-free loss function, parameterized as a neural network, which leverages target-set information. A base-model learns on a support-set using existing methods (e.g. stochastic gradient descent combined with the cross-entropy loss), and then is updated for the incoming target-task using a new learned loss function (i.e. the meta-learned label-free loss). This unsupervised loss function is optimized such that the learnt model achieves higher generalization performance. Experiments demonstrate that SCA offers substantially higher and state-of-the-art generalization performance compared to baselines which only adapt on the support-set.

In this paper the design, implementation, and evaluation of a general purpose smartphone based intelligent surveillance system is presented. It has two main elements; i) a detection module, and ii) a classification module. The detection module is based on the recently introduced approach that combines the well-known background subtraction method with the optical flow and recursively estimated density. The classification module is based on a neural network using Deep Learning methodology. Firstly, the architecture design of the convolutional neural network is presented and analyzed in the context of the four selected architectures (two of them recent successful types) and two custom modifications specifically made for the problem at hand. The results are carefully evaluated, and the best one is selected to be used within the proposed system. In addition, the system is implemented on both a PC (using Linux type OS) and on a smartphone (using Android). In addition to the compatibility with all modern Android-based devices, most GPU-powered platforms such as Raspberry Pi, Nvidia Tegra X1 and Jetson run on Linux. The proposed system can easily be installed on any such device benefiting from the advantage of parallelisation for faster execution. The proposed system achieved a performance which surpasses that of a human (classification accuracy of the top 1 class >95.9% for automatic recognition of a detected object into one of the seven selected categories. For the top-2 classes, the accuracy is even higher (99.85%). That means, at least, one of the two top classes suggested by the system is correct. Finally, a number of visual examples are showcased of the system in use in both PC and Android devices.