Models that can simulate how environments change in response to actions can be used by agents to plan and act efficiently. We improve on previous environment simulators from high-dimensional pixel observations by introducing recurrent neural networks that are able to make temporally and spatially coherent predictions for hundreds of time-steps into the future. We present an in-depth analysis of the factors affecting performance, providing the most extensive attempt to advance the understanding of the properties of these models. We address the issue of computationally inefficiency with a model that does not need to generate a high-dimensional image at each time-step. We show that our approach can be used to improve exploration and is adaptable to many diverse environments, namely 10 Atari games, a 3D car racing environment, and complex 3D mazes.

In this paper, we propose a new Recurrent Neural Network (RNN) architecture. The novelty is simple: We use diagonal recurrent matrices instead of full. This results in better test likelihood and faster convergence compared to regular full RNNs in most of our experiments. We show the benefits of using diagonal recurrent matrices with popularly used LSTM and GRU architectures as well as with the vanilla RNN architecture, on four standard symbolic music datasets.

This paper presents an end-to-end approach for tracking static and dynamic objects for an autonomous vehicle driving through crowded urban environments. Unlike traditional approaches to tracking, this method is learned end-to-end, and is able to directly predict a full unoccluded occupancy grid map from raw laser input data. Inspired by the recently presented DeepTracking approach [Ondruska, 2016], we employ a recurrent neural network (RNN) to capture the temporal evolution of the state of the environment, and propose to use Spatial Transformer modules to exploit estimates of the egomotion of the vehicle. Our results demonstrate the ability to track a range of objects, including cars, buses, pedestrians, and cyclists through occlusion, from both moving and stationary platforms, using a single learned model. Experimental results demonstrate that the model can also predict the future states of objects from current inputs, with greater accuracy than previous work.

The explosive increase of Electronic Medical Records (EMR) provides many opportunities to carry out data science research by applying data mining and machine learning tools and techniques. EMR contains massive and a wide range of information on patients concerning different aspects of healthcare, such as patient conditions, diagnostic tests, labs, imaging exams, genomics, proteomics, treatments, outcomes or claims, financial records [1]. Particularly, the extensive and powerful patient-centered data enables data scientists and medical researchers to conduct their research in the field of personalized (precision) medicine. Personalized medicine is defined as [2]: "the use of combined knowledge (genetic or otherwise) about a person to predict disease susceptibility, disease prognosis, or treatment response and thereby improve that persons health." In other words, the goal of precision medicine or personalized healthcare is to provide the right treatment to the right patient at the right time. Personalized medicine is a multi-disciplinary area that combines data science tools and statistics techniques with medical knowledge to develop tailor-made treatment, prevention and intervention plans for individual patients. In this study we focus on a vulnerable demographic subgroup (African-American) at high-risk for hypertension (HTN), poor blood pressure control and consequently, adverse pressure-related cardiovascular complications. We use left Figure 1.

Although stochastic approximation learning methods have been widely used in the machine learning literature for over 50 years, formal theoretical analyses of specific machine learning algorithms are less common because stochastic approximation theorems typically possess assumptions which are difficult to communicate and verify. This paper presents a new stochastic approximation theorem for state-dependent noise with easily verifiable assumptions applicable to the analysis and design of important deep learning algorithms including: adaptive learning, contrastive divergence learning, stochastic descent expectation maximization, and active learning.

In this paper we introduce a model of lifelong learning, based on a Network of Experts. New tasks / experts are learned and added to the model sequentially, building on what was learned before. To ensure scalability of this process, data from previous tasks cannot be stored and hence is not available when learning a new task. A critical issue in such context, not addressed in the literature so far, relates to the decision which expert to deploy at test time. We introduce a set of gating autoencoders that learn a representation for the task at hand, and, at test time, automatically forward the test sample to the relevant expert. This also brings memory efficiency as only one expert network has to be loaded into memory at any given time. Further, the autoencoders inherently capture the relatedness of one task to another, based on which the most relevant prior model to be used for training a new expert, with fine-tuning or learningwithout-forgetting, can be selected. We evaluate our method on image classification and video prediction problems.

We analyze dropout in deep networks with rectified linear units and the quadratic loss. Our results expose surprising differences between the behavior of dropout and more traditional regularizers like weight decay. For example, on some simple data sets dropout training produces negative weights even though the output is the sum of the inputs. This provides a counterpoint to the suggestion that dropout discourages co-adaptation of weights. We also show that the dropout penalty can grow exponentially in the depth of the network while the weight-decay penalty remains essentially linear, and that dropout is insensitive to various re-scalings of the input features, outputs, and network weights. This last insensitivity implies that there are no isolated local minima of the dropout training criterion. Our work uncovers new properties of dropout, extends our understanding of why dropout succeeds, and lays the foundation for further progress.

By considering more complex decisions (many more influencers), AI techniques extend traditional Machine learning strategies(like Anomaly detection).

Progress in artificial intelligence causes some people to worry that software will take jobs such as driving trucks away from humans. Now leading researchers are finding that they can make software that can learn to do one of the trickiest parts of their own jobs--the task of designing machine-learning software. In one experiment, researchers at the Google Brain artificial intelligence research group had software design a machine-learning system to take a test used to benchmark software that processes language. What it came up with surpassed previously published results from software designed by humans. In recent months several other groups have also reported progress on getting learning software to make learning software.

Summary: We are swept up by the rapid advances in AI and deep learning, and tend to laugh off AI's failures as good fodder for YouTube videos. But those failures are starting to add up. It's time to take a hard look at the weaknesses in AI and where that's leading us. Trying to spot new themes and directions I was trolling through some data and I found this. This is the Google Trends chart of searches for'Artificial Intelligence' for the last four years.