The process of identifying changes or transformations in a scene along with the ability of reasoning about their causes and effects, is a key aspect of intelligence. In this work we go beyond recent advances in computational perception, and introduce a more challenging task, Image-based Event-Sequencing (IES). In IES, the task is to predict a sequence of actions required to rearrange objects from the configuration in an input source image to the one in the target image. IES also requires systems to possess inductive generalizability. Motivated from evidence in cognitive development, we compile the first IES dataset, the Blocksworld Image Reasoning Dataset (BIRD) which contains images of wooden blocks in different configurations, and the sequence of moves to rearrange one configuration to the other. We first explore the use of existing deep learning architectures and show that these end-to-end methods under-perform in inferring temporal event-sequences and fail at inductive generalization. We then propose a modular two-step approach: Visual Perception followed by Event-Sequencing, and demonstrate improved performance by combining learning and reasoning. Finally, by showing an extension of our approach on natural images, we seek to pave the way for future research on event sequencing for real world scenes.

In this working notes paper, we describe IBM Research AI (Almaden) team's participation in the ImageCLEF 2019 VQA-Med competition. The challenge consists of four question-answering tasks based on radiology images. The diversity of imaging modalities, organs and disease types combined with a small imbalanced training set made this a highly complex problem. To overcome these difficulties, we implemented a modular pipeline architecture that utilized transfer learning and multi-task learning. Our findings led to the development of a novel model called Supporting Facts Network (SFN). The main idea behind SFN is to cross-utilize information from upstream tasks to improve the accuracy on harder downstream ones. This approach significantly improved the scores achieved in the validation set (18 point improvement in F-1 score). Finally, we submitted four runs to the competition and were ranked seventh.

Because of the rich dynamical structure of videos and their ubiquity in everyday life, it is a natural idea that video data could serve as a powerful unsupervised learning signal for training visual representations in deep neural networks. However, instantiating this idea, especially at large scale, has remained a significant artificial intelligence challenge. Here we present the Video Instance Embedding (VIE) framework, which extends powerful recent unsupervised loss functions for learning deep nonlinear embeddings to multi-stream temporal processing architectures on large-scale video datasets. We show that VIE-trained networks substantially advance the state of the art in unsupervised learning from video datastreams, both for action recognition in the Kinetics dataset, and object recognition in the ImageNet dataset. We show that a hybrid model with both static and dynamic processing pathways is optimal for both transfer tasks, and provide analyses indicating how the pathways differ. Taken in context, our results suggest that deep neural embeddings are a promising approach to unsupervised visual learning across a wide variety of domains.

We propose a novel deep learning method for local self-supervised representation learning that does not require labels nor end-to-end backpropagation but exploits the natural order in data instead. Inspired by the observation that biological neural networks appear to learn without backpropagating a global error signal, we split a deep neural network into a stack of gradient-isolated modules. Each module is trained to maximize the mutual information between its consecutive outputs using the InfoNCE bound from Oord et al. [2018]. Despite this greedy training, we demonstrate that each module improves upon the output of its predecessor, and that the representations created by the top module yield highly competitive results on downstream classification tasks in the audio and visual domain. The proposal enables optimizing modules asynchronously, allowing large-scale distributed training of very deep neural networks on unlabelled datasets.

Capturing long-range dependencies in feature representations is crucial for many visual recognition tasks. Despite recent successes of deep convolutional networks, it remains challenging to model non-local context relations between visual features. A promising strategy is to model the feature context by a fully-connected graph neural network (GNN), which augments traditional convolutional features with an estimated non-local context representation. However, most GNN-based approaches require computing a dense graph affinity matrix and hence have difficulty in scaling up to tackle complex real-world visual problems. In this work, we propose an efficient and yet flexible non-local relation representation based on a novel class of graph neural networks. Our key idea is to introduce a latent space to reduce the complexity of graph, which allows us to use a low-rank representation for the graph affinity matrix and to achieve a linear complexity in computation. Extensive experimental evaluations on three major visual recognition tasks show that our method outperforms the prior works with a large margin while maintaining a low computation cost.

According to Wikipedia, the earliest diagnostic biopsy was conducted by an Arab physician in the 11th century. The term "biopsy" derives from the Greek word bios which means life and the also Greek word "opsis" which means sight. It was first introduced in the 19th century and since then medical researchers and practitioners have developed several techniques to perform biopsies on living organs or tissues needed further observation. Modern instruments such as intestinal biopsy tubes can extract samples from parts of the body where it is difficult to have access or even hazardous. In the cases described above, this histologic procedure is invasive and quite complicated putting a financial burden, not to mention the extra strain on the patient.

Artificial intelligence has the potential to transform healthcare and the pharma industry; making health services more predictive, more effective and more equitable. But can AI's potential be realised, and what can chief information officers and clinical chief information officers do to get the most out of the new technology? The Healthcare Information and Management Systems Society (HIMSS) Executive Leadership summit last month attended by NHS CIOs and CCIOs, leaders from NHS England and NHS Digital and other health IT innovators, expressed the potential AI-driven data analytics has in supporting the NHS's Five Year Forward View, narrowing gaps in health provision. The summit discussed how AI could address the health and wellbeing gap by predicting which individuals most risk illness, allowing the NHS to target treatments accordingly, with AI giving health professionals and patients bespoke diagnostics and treatments. It was also noted that AI could help address the efficiency and funding gap by automating tasks, triaging patients to the most appropriate services and allowing them to self-care.

Here are the videos and slides of Workshop IV: Deep Geometric Learning of Big Data and Applications, Part of the Long Program Geometry and Learning from Data in 3D and Beyond at IPAM. The workshop took place May 20 - 24, 2019. And thank you to the organizing committee (Xavier Bresson, Yann LeCun, Stanley Osher, Rene Vidal, Rebecca Willett) for making this workshop happen!

Object detection is a task in computer vision that involves identifying the presence, location, and type of one or more objects in a given photograph. It is a challenging problem that involves building upon methods for object recognition (e.g. In recent years, deep learning techniques are achieving state-of-the-art results for object detection, such as on standard benchmark datasets and in computer vision competitions. Notable is the "You Only Look Once," or YOLO, family of Convolutional Neural Networks that achieve near state-of-the-art results with a single end-to-end model that can perform object detection in real-time. In this tutorial, you will discover how to develop a YOLOv3 model for object detection on new photographs. How to Perform Object Detection With YOLOv3 in Keras Photo by David Berkowitz, some rights reserved.

A three-dimensional printout of a human heart is seen at the Heidelberg University Hospital (Universitaetsklinikum Heidelberg) in Heidelberg, Germany, August 14, 2018. If artificial intelligence (AI) has a big say in what we should watch next on YouTube or any other multimedia streaming platform, then perhaps it can also recommend what we should eat or drink next--or more likely not eat or drink next--to stay in the best of health. Already the applications of AI are going beyond tips for maintaining a healthy lifestyle, and AI-powered software is becoming an integral part of some medical diagnosis procedures. Thanks to deep learning algorithms and neural networks, AI solutions have become very good at pattern recognition, which after all, lies at the core of what a human doctor does for figuring out the root cause of a patient's ailment. Doctors, in essence, examine all the symptoms a patient is exhibiting--often with the help of medical imaging, bloodwork, and pathological tests--and then compare the systems with the telltale signs of likely diseases and conditions that they have learned about as medical students or cases that they have come across in the medical literature.