In 2017, the "digital medicine" Spinraza was released to the public, after years of drug development to cure Spinal Muscular Atrophy (SMA), at a price of $750,000 initially and $375,000 annually after that. The cause of SMA was a simple mutation on the SMN1 gene on chromosome 5. One altered nucleotide sequence in the exon of the SMN1 gene changed the complete life trajectory for children born with this disease, many dying before the end of infancy. However, the price of pharmaceutical drug, which many government's and insurance companies refuse to pay, has left children unable to acquire treatment. All the medicine simply does, is takes the reverse compliment sequence of a neighbouring intronic sequence, and binds to it.

A group of researchers at Sandia National Laboratories have developed a tool that can cross-train standard convolutional neural networks (CNN) to a spiking neural model that can be used on neuromorphic processors. The researchers claim that the conversion will enable deep learning applications to take advantage of the much better energy efficiency of neuromorphic hardware, which are designed to mimic the way the biological neurons work. The tool, known as Whetstone, works by adjusting artificial neuron behavior during the training phase to only activate when it reaches an appropriate threshold. As a result, neuron activation become a binary choice – either it spikes or it doesn't. By doing so, Whetstone converts an artificial neural network into a spiking neural network.

NVIDIA's GPU Technology Conference (GTC) is the premier AI conference, offering hundreds of workshops, sessions, and keynotes hosted by organizations like Google, Amazon, Facebook as well as rising startups. In the video below, see the the top 5 deep learning sessions you can attend at the conference. TensorFlow is a flexible, high-performance software library for numerical computation using data flow graphs and NVIDIA TensorRT is a platform for high-performance deep learning inference. We'll describe how TensorRT is integrated with TensorFlow and show how combining the two improves efficiency of machine learning models. We'll also use examples to show how to use the integration.

Deep learning, which is a subset of AI (Artificial Intelligence), has been around since the 1950s. It's focused on developing systems that mimic the brain's neural network structure. Yet it was not until the 1980s that deep learning started to show promise, spurred by the pioneering theories of researchers like Geoffrey Hinton, Yoshua Bengio and Yann Lecun. There was also the benefit of accelerating improvements in computer power. Despite all this, there remained lots of skepticism. Deep learning approaches still looked more like interesting academic exercises that were not ready for prime time.

As part of my own learning, continuing from Part 1 and trying to improve our neural network model, we will use some of the well-known machine learning techniques mentioned on TensorFlow. In the previous article, we saw certain problems with our training. Here, we will address them and see if our results improve as we go. A model is considered to overfit when it performs with great accuracy on the training data (data used for training the model), but when evaluated against a test or unseen data set, it performs rather poorly. This happens because our model has overfit the data.

The inputs of deep neural network (DNN) from real-world data usually come with uncertainties. Yet, it is challenging to propagate the uncertainty in the input features to the DNN predictions at a low computational cost. This work employs a gradient-based subspace method and response surface technique to accelerate the uncertainty propagation in DNN. Specifically, the active subspace method is employed to identify the most important subspace in the input features using the gradient of the DNN output to the inputs. Then the response surface within that low-dimensional subspace can be efficiently built, and the uncertainty of the prediction can be acquired by evaluating the computationally cheap response surface instead of the DNN models. In addition, the subspace can help explain the adversarial examples. The approach is demonstrated in MNIST datasets with a convolutional neural network.

Electric Vehicle (EV) sharing systems have recently experienced unprecedented growth across the globe. During their fast expansion, one fundamental determinant for success is the capability of dynamically predicting the demand of stations as the entire system is evolving continuously. There are several challenges in this dynamic demand prediction problem. Firstly, unlike most of the existing work which predicts demand only for static systems or at few stages of expansion, in the real world we often need to predict the demand as or even before stations are being deployed or closed, to provide information and support for decision making. Secondly, for the stations to be deployed, there is no historical record or additional mobility data available to help the prediction of their demand. Finally, the impact of deploying/closing stations to the remaining stations in the system can be very complex. To address these challenges, in this paper we propose a novel dynamic demand prediction approach based on graph sequence learning, which is able to model the dynamics during the system expansion and predict demand accordingly. We use a local temporal encoding process to handle the available historical data at individual stations, and a dynamic spatial encoding process to take correlations between stations into account with graph convolutional neural networks. The encoded features are fed to a multi-scale prediction network, which forecasts both the long-term expected demand of the stations and their instant demand in the near future. We evaluate the proposed approach on real-world data collected from a major EV sharing platform in Shanghai for one year. Experimental results demonstrate that our approach significantly outperforms the state of the art, showing up to three-fold performance gain in predicting demand for the rapidly expanding EV sharing system.

We propose a simple yet effective technique for neural network learning. The forward propagation is computed as usual. In back propagation, only a small subset of the full gradient is computed to update the model parameters. The gradient vectors are sparsified in such a way that only the top-$k$ elements (in terms of magnitude) are kept. As a result, only $k$ rows or columns (depending on the layout) of the weight matrix are modified, leading to a linear reduction ($k$ divided by the vector dimension) in the computational cost. Surprisingly, experimental results demonstrate that we can update only 1-4% of the weights at each back propagation pass. This does not result in a larger number of training iterations. More interestingly, the accuracy of the resulting models is actually improved rather than degraded, and a detailed analysis is given. The code is available at https://github.com/lancopku/meProp

This work investigates multiple approaches to Named Entity Recognition (NER) for text in Electronic Health Record (EHR) data. In particular, we look into the application of (i) rule-based, (ii) deep learning and (iii) transfer learning systems for the task of NER on brain imaging reports with a focus on records from patients with stroke. We explore the strengths and weaknesses of each approach, develop rules and train on a common dataset, and evaluate each system's performance on common test sets of Scottish radiology reports from two sources (brain imaging reports in ESS -- Edinburgh Stroke Study data collected by NHS Lothian as well as radiology reports created in NHS Tayside). Our comparison shows that a hand-crafted system is the most accurate way to automatically label EHR, but machine learning approaches can provide a feasible alternative where resources for a manual system are not readily available.

With the great achievement of artificial intelligence, vehicle technologies have advanced significantly from human centric driving towards fully automated driving. An intelligent vehicle should be able to understand the driver's perception of the environment as well as controlling behavior of the vehicle. Since high digital map information has been available to provide rich environmental context about static roads, buildings and traffic infrastructures, it would be worthwhile to explore map data capability for driving task understanding. Alternative to commercial used maps, the OpenStreetMap (OSM) data is a free open dataset, which makes it unique for the exploration research. This study is focused on two tasks that leverage OSM for driving environment understanding. First, driving scenario attributes are retrieved from OSM elements, which are combined with vehicle dynamic signals for the driving event recognition. Utilizing steering angle changes and based on a Bi-directional Recurrent Neural Network (Bi-RNN), a driving sequence is segmented and classified as lane-keeping, lane-change-left, lane-change-right, turn-left, and turn-right events. Second, for autonomous driving perception, OSM data can be used to render virtual street views, represented as prior knowledge to fuse with vision/laser systems for road semantic segmentation. Five different types of road masks are generated from OSM, images, and Lidar points, and fused to characterize the drivable space at the driver's perspective. An alternative data-driven approach is based on a Fully Convolutional Network (FCN), OSM availability for deep learning methods are discussed to reveal potential usage on compensating street view images and automatic road semantic annotation.