Convolutional Neural Networks (CNNs) have become the method of choice for learning problems involving 2D planar images. However, a number of problems of recent interest have created a demand for models that can analyze spherical images. Examples include omnidirectional vision for drones, robots, and autonomous cars, molecular regression problems, and global weather and climate modelling. A naive application of convolutional networks to a planar projection of the spherical signal is destined to fail, because the space-varying distortions introduced by such a projection will make translational weight sharing ineffective. In this paper we introduce the building blocks for constructing spherical CNNs. We propose a definition for the spherical cross-correlation that is both expressive and rotation-equivariant. The spherical correlation satisfies a generalized Fourier theorem, which allows us to compute it efficiently using a generalized (non-commutative) Fast Fourier Transform (FFT) algorithm. We demonstrate the computational efficiency, numerical accuracy, and effectiveness of spherical CNNs applied to 3D model recognition and atomization energy regression.

This paper presents a novel two-step approach for the fundamental problem of learning an optimal map from one distribution to another. First, we learn an optimal transport (OT) plan, which can be thought as a one-to-many map between the two distributions. To that end, we propose a stochastic dual approach of regularized OT, and show empirically that it scales better than a recent related approach when the amount of samples is very large. Second, we estimate a \textit{Monge map} as a deep neural network learned by approximating the barycentric projection of the previously-obtained OT plan. This parameterization allows generalization of the mapping outside the support of the input measure. We prove two theoretical stability results of regularized OT which show that our estimations converge to the OT plan and Monge map between the underlying continuous measures. We showcase our proposed approach on two applications: domain adaptation and generative modeling.

It is easy for people to imagine what a man with pink hair looks like, even if they have never seen such a person before. We call the ability to create images of novel semantic concepts visually grounded imagination. In this paper, we show how we can modify variational auto-encoders to perform this task. Our method uses a novel training objective, and a novel product-of-experts inference network, which can handle partially specified (abstract) concepts in a principled and efficient way. We also propose a set of easy-to-compute evaluation metrics that capture our intuitive notions of what it means to have good visual imagination, namely correctness, coverage, and compositionality (the 3 C's). Finally, we perform a detailed comparison of our method with two existing joint image-attribute VAE methods (the JMVAE method of Suzuki et al. (2017) and the BiVCCA method of Wang et al. (2016b)) by applying them to two datasets: the MNIST-with-attributes dataset (which we introduce here), and the CelebA dataset (Liu et al., 2015).

Supervised object detection and semantic segmentation require object or even pixel level annotations. When there exist image level labels only, it is challenging for weakly supervised algorithms to achieve accurate predictions. The accuracy achieved by top weakly supervised algorithms is still significantly lower than their fully supervised counterparts. In this paper, we propose a novel weakly supervised curriculum learning pipeline for multi-label object recognition, detection and semantic segmentation. In this pipeline, we first obtain intermediate object localization and pixel labeling results for the training images, and then use such results to train task-specific deep networks in a fully supervised manner. The entire process consists of four stages, including object localization in the training images, filtering and fusing object instances, pixel labeling for the training images, and task-specific network training. To obtain clean object instances in the training images, we propose a novel algorithm for filtering, fusing and classifying object instances collected from multiple solution mechanisms. In this algorithm, we incorporate both metric learning and density-based clustering to filter detected object instances. Experiments show that our weakly supervised pipeline achieves state-of-the-art results in multi-label image classification as well as weakly supervised object detection and very competitive results in weakly supervised semantic segmentation on MS-COCO, PASCAL VOC 2007 and PASCAL VOC 2012.

Due to recent technical and scientific advances, we have a wealth of information hidden in unstructured text data such as offline/online narratives, research articles, and clinical reports. To mine these data properly, attributable to their innate ambiguity, a Word Sense Disambiguation (WSD) algorithm can avoid numbers of difficulties in Natural Language Processing (NLP) pipeline. However, considering a large number of ambiguous words in one language or technical domain, we may encounter limiting constraints for proper deployment of existing WSD models. This paper attempts to address the problem of one-classifier-per-one-word WSD algorithms by proposing a single Bidirectional Long Short-Term Memory (BLSTM) network which by considering senses and context sequences works on all ambiguous words collectively. Evaluated on SensEval-3 benchmark, we show the result of our model is comparable with top-performing WSD algorithms. We also discuss how applying additional modifications alleviates the model fault and the need for more training data.

We're excited to announce a medical research partnership with the US Department of Veterans Affairs (VA), one of the world's leading healthcare organisations responsible for providing high-quality care to veterans and their families across the United States. This project will see us analyse patterns from historical, depersonalised medical records to predict patient deterioration. Patient deterioration is a significant global health problem that often has fatal consequences. Studies estimate that 11% of all in-hospital deaths are due to patient deterioration not being recognised early enough or acted on in the right way. Alongside world-renowned clinicians and researchers at the VA, we are analysing patterns from approximately 700,000 historical, depersonalised medical records in order to determine if machine learning can accurately identify the risk factors for patient deterioration and correctly predict its onset.

The idea behind using a neural network for image recognition is that you don't have to tell it what to look for in an image. You don't even need to care about what it looks for. With enough training, the neural network should be able to pick out details that allow it to make accurate identifications. For things like figuring out whether there's a cat in an image, neural networks don't provide much, if any, advantages over the actual neurons in our visual system. But where they can potentially shine are cases where we don't know what to look for.

This is a guest post by Adrian Rosebrock. Adrian is the author of PyImageSearch.com, a blog about computer vision and deep learning. Adrian recently finished authoring Deep Learning for Computer Vision with Python, a new book on deep learning for computer vision and image recognition using Keras. In this tutorial, we will present a simple method to take a Keras model and deploy it as a REST API. The examples covered in this post will serve as a template/starting point for building your own deep learning APIs -- you will be able to extend the code and customize it based on how scalable and robust your API endpoint needs to be.

My intuition tells me there's potential for breakthroughs there. People have focused on squeezing out more accuracy on big datasets like CIFAR and ignored the low sample size regime.

As the years go by, many possibilities of Artificial Intelligence application have been realized and many are yet to be. Despite the progress that several industries have made in the AI front, healthcare continues to be one sector where it has truly made a major impact that goes beyond convenience and essentially affects human lives. Artificial intelligence (AI) is defined as "the science of making computers do things that require intelligence when done by humans" by Turing Archive for the History of Computing. While we have not reached the level of sophistication in AI as in the Westworld, the AI technology is quickly developing. AI can re-design and improve healthcare in multiple ways.