Few-shot object detection, the problem of modelling novel object detection categories with few training instances, is an emerging topic in the area of few-shot learning and object detection. Contemporary techniques can be divided into two groups: fine-tuning based and meta-learning based approaches. While meta-learning approaches aim to learn dedicated meta-models for mapping samples to novel class models, fine-tuning approaches tackle few-shot detection in a simpler manner, by adapting the detection model to novel classes through gradient based optimization. Despite their simplicity, fine-tuning based approaches typically yield competitive detection results. Based on this observation, we focus on the role of loss functions and augmentations as the force driving the fine-tuning process, and propose to tune their dynamics through meta-learning principles. The proposed training scheme, therefore, allows learning inductive biases that can boost few-shot detection, while keeping the advantages of fine-tuning based approaches. In addition, the proposed approach yields interpretable loss functions, as opposed to highly parametric and complex few-shot meta-models. The experimental results highlight the merits of the proposed scheme, with significant improvements over the strong fine-tuning based few-shot detection baselines on benchmark Pascal VOC and MS-COCO datasets, in terms of both standard and generalized few-shot performance metrics.

Just like other few-shot learning problems, few-shot segmentation aims to minimize the need for manual annotation, which is particularly costly in segmentation tasks. Even though the few-shot setting reduces this cost for novel test classes, there is still a need to annotate the training data. To alleviate this need, we propose a self-supervised training approach for learning few-shot segmentation models. We first use unsupervised saliency estimation to obtain pseudo-masks on images. We then train a simple prototype based model over different splits of pseudo masks and augmentations of images. Our extensive experiments show that the proposed approach achieves promising results, highlighting the potential of self-supervised training. To the best of our knowledge this is the first work that addresses unsupervised few-shot segmentation problem on natural images.

Automated Telling Machines (ATM) are one of the prominent services of the banks, which facilitate daily banking operations. Among the offered services, money withdrawal is a very basic functionality of ATMs. For the banks, it is important to manage the amount of money to be loaded in ATMs. Hence, prediction of the withdrawal amount is an important step in the ATM management. In this work, we investigate the performance of deep learning techniques for ATM money withdrawal amount prediction problem. The problem is defined in two ways: predicting the amount of money to be withdrawn, and predicting the class label of the withdrawal amount. Our data set includes daily total withdrawal amounts together with the date information. In our experiments, we further analyzed the effect of additional information, such as size of history window, weather condition, currency rate and location. For numeric value prediction task, we compared the prediction performance with statistical models of ARIMA and SARIMA. The experiments show that the neural architectures are feasible for the task of withdrawal amount prediction. They can especially provide high accuracy results when the problem is modeled as a class label prediction task.