Reinforcement learning algorithms that can reliably learn how to control robots, etc. Better generative models. Algorithms that can reliably learn how to generate images, speech and text that humans can't tell apart from the real thing. Learning to learn and ubiquitous deep learning. Right now it still takes a human expert to run the learning-to-learn algorithm, but in the future it will be easier to deploy, and all kinds of businesses that don't specialize in AI will be able to leverage deep learning. More cyberattacks will leverage machine learning to make more autonomous malware, more efficient fuzzing for vulnerabilities, etc. More cyberdefenses will leverage machine learning to respond faster than a human could, detect more subtle intrusions, etc. ML algorithms from opposing camps will fool each other to carry out both attacks and defensive actions.

Google-owned DeepMind is working on artificial intelligence (AI) that can imagine like humans and handle the unpredictable scenarios in real world. According to a report in Wired on Thursday, DeepMind, that was acquired by Google in 2014, is developing an AI capable of'imagination', enabling machines to see the consequences of their actions before they make them. "Its attempt to create algorithms that simulate the distinctly human ability to construct a plan could eventually help to produce software and hardware capable of solving complex tasks more efficiently," the report noted. DeepMind was successful in AI when it developed'AlphaGo' that recently beat a series of human champions at the tricky board game'Go'. But in case of'AlphaGo', there are a set of defined rules and predictable outcomes.

Modern films and TV shows are filled with spectacular computer-generated sequences computed by rendering systems that simulate the flow of light in a three-dimensional scene and convert the information into a two-dimensional image. But computing the thousands of light rays (per frame) to achieve accurate colour, shadows, reflectivity and other light-based characteristics is a labour-intensive, time-consuming and expensive undertaking. An alternative is to render the images using only a few light rays. That saves time and labour but results in inaccuracies that show up as objectionable "noise" in the final image. UC Santa Barbara electrical and computer engineering Ph.D. student Steve Bako and his advisor, Pradeep Sen, are advancing on a solution.

This articles was written by Andrej Karpathy. Andrej, PhD student at Stanford, is a Research Scientist at OpenAI working on Deep Learning, Generative Models and Reinforcement Learning. I still remember when I trained my first recurrent network for Image Captioning. Within a few dozen minutes of training my first baby model (with rather arbitrarily-chosen hyperparameters) started to generate very nice looking descriptions of images that were on the edge of making sense. Sometimes the ratio of how simple your model is to the quality of the results you get out of it blows past your expectations, and this was one of those times.

PCA is a classical statistical technique whose simplicity and maturity has seen it find widespread use as an anomaly detection technique. However, it is limited in this regard by being sensitive to gross perturbations of the input, and by seeking a linear subspace that captures normal behaviour. The first issue has been dealt with by robust PCA, a variant of PCA that explicitly allows for some data points to be arbitrarily corrupted, however, this does not resolve the second issue, and indeed introduces the new issue that one can no longer inductively find anomalies on a test set. This paper addresses both issues in a single model, the robust autoencoder. This method learns a nonlinear subspace that captures the majority of data points, while allowing for some data to have arbitrary corruption. The model is simple to train and leverages recent advances in the optimisation of deep neural networks. Experiments on a range of real-world datasets highlight the model's effectiveness.

In this work we present a novel system for PET estimation using CT scans. We explore the use of fully convolutional networks (FCN) and conditional generative adversarial networks (GAN) to export PET data from CT data. Our dataset includes 25 pairs of PET and CT scans where 17 were used for training and 8 for testing. The system was tested for detection of malignant tumors in the liver region. Initial results look promising showing high detection performance with a TPR of 92.3% and FPR of 0.25 per case. Future work entails expansion of the current system to the entire body using a much larger dataset. Such a system can be used for tumor detection and drug treatment evaluation in a CT-only environment instead of the expansive and radioactive PET-CT scan.

Single image super-resolution is the task of inferring a high-resolution image from a single low-resolution input. Traditionally, the performance of algorithms for this task is measured using pixel-wise reconstruction measures such as peak signal-to-noise ratio (PSNR) which have been shown to correlate poorly with the human perception of image quality. As a result, algorithms minimizing these metrics tend to produce over-smoothed images that lack high-frequency textures and do not look natural despite yielding high PSNR values. We propose a novel application of automated texture synthesis in combination with a perceptual loss focusing on creating realistic textures rather than optimizing for a pixel-accurate reproduction of ground truth images during training. By using feed-forward fully convolutional neural networks in an adversarial training setting, we achieve a significant boost in image quality at high magnification ratios. Extensive experiments on a number of datasets show the effectiveness of our approach, yielding state-of-the-art results in both quantitative and qualitative benchmarks.

In his introduction, Jacobs offers a brief history of Canada's pioneering contribution to the field of artificial intelligence (AI), explaining the significance of the shift between rules-based AI and machine learning that originated in Ontario. "Forty years ago, the prevalent form of AI involved programmers using IF/THEN statements to teach machines," Jacobs explains. "Then there were these outliers who believed that, 'no, you're not going to program anything, the machine is going to figure it out itself, and it's going to do this by using artificial neurons that mimic how the brain works.' The leader of that group was someone named Geoffrey Hinton, and for most of his career, people said that he was crazy…They couldn't really get any funding except for a couple small research organizations in Canada, including CIFAR." The Canadian Institute for Advanced Research (CIFAR) that Jacobs refers to, approved its first program, Artificial Intelligence & Robotics in 1982, while operating out of an Ontario government office just a few blocks from where Jacobs is sitting, and later recruited Geoffrey Hinton to Toronto.

Restricted Boltzmann Machines are described by the Gibbs measure of a bipartite spin glass, which in turn corresponds to the one of a generalised Hopfield network. This equivalence allows us to characterise the state of these systems in terms of retrieval capabilities, both at low and high load. We study the paramagnetic-spin glass and the spin glass-retrieval phase transitions, as the pattern (i.e. weight) distribution and spin (i.e. unit) priors vary smoothly from Gaussian real variables to Boolean discrete variables. Our analysis shows that the presence of a retrieval phase is robust and not peculiar to the standard Hopfield model with Boolean patterns. The retrieval region is larger when the pattern entries and retrieval units get more peaked and, conversely, when the hidden units acquire a broader prior and therefore have a stronger response to high fields. Moreover, at low load retrieval always exists below some critical temperature, for every pattern distribution ranging from the Boolean to the Gaussian case.