The second, a Base AMI, available in Amazon Linux and Ubuntu versions, provides a high-performance foundational platform for power users to run their own customized deep learning models. The Conda-based AMI comes packaged with latest official releases of the following deep learning frameworks: Apache MXNet 0.12 with Gluon, TensorFlow 1.4, Caffe2 0.8.1, PyTorch 0.2, CNTK 2.2, Theano 0.9, Keras 1.2.2 and Keras 2.0.9. The Base AMI provides the foundation of following GPU drivers and libraries: CUDA 8 and 9, CuBLAS 8 and 9, CuDNN 6 and 7, glibc 2.18, OpenCV 3.2.0, Both of the new AMIs available from the AWS Marketplace include the following libraries and drivers for GPU acceleration on the cloud: CUDA 8 and 9, cuDNN 6 and 7, NCCL 2.0.5 libraries and. To assist with installation of the AMI version that best fits your needs, we have added wizard directly in the AWS console, created a step-by-step guide and provided additional how-to resources in our new documentation site.

Although it seemed like an afterthought, that introduction was the beginning of the mass movement within A.I. and Deep Learning. Artificial Intelligence (A.I., otherwise known as Machine Intelligence), is, as the name suggests: intelligence that is displayed by machines in comparison to our known, natural intelligence -- intelligence displayed by humans and other animals. From its root dating back to the summer of 1956 in Dartmouth College, the term "Artificial Intelligence" was coined by a group of scientists and mathematicians that was derived from a brainstorming session in which ways that robots and machines could simulate and potentially solve some issues in society. From then, the fascination with robots taking over the world (whether for good or evil) has been depicted in pop culture and movies, especially in the old movies in the 1960s and 1970s. A.I. has a wide range of technologies such as logic and rule-based systems that enables computers and robots to solve problems in ways that at least superficially resemble thinking. A subset of A.I. is a term, called -- Deep Learning. Deep Learning, believe it or not, is in our pockets! It is the training of our machines, software and applications. The best simplification of Deep Learning could be thought of as "A to B mappings", according Andrew Ng, the chief scientist at Baidu Research.

The dataset, released by the NIH, contains 112,120 frontal-view X-ray images of 30,805 unique patients, annotated with up to 14 different thoracic pathology labels using NLP methods on radiology reports. We label images that have pneumonia as one of the annotated pathologies as positive examples and label all other images as negative examples for the pneumonia detection task. We collected a test set of 420 frontal chest X-rays. Annotations were obtained independently from four practicing radiologists at Stanford University, who were asked to label all 14 pathologies, even though . We then evaluate the performance of an individual radiologist by using the majority vote of the other 3 radiologists as ground truth.

Researchers from our group at QUT and the Australian Centre for Robotic Vision have had six papers accepted to the upcoming Australasian Conference on Robotics and Automation to be held at The University of Technology Sydney. This year the conference trialed a dual submission process with the IEEE International Conference on Robotics and Automation, meaning work can be presented at both conferences but only published in the proceedings of one. The papers cover ongoing research in our lab spanning topics including robotics, positioning and AI for applications in mining, construction safety and autonomous vehicles. I'll give an overview here of the research we're doing, and a wrap up at the end. Despite very high safety standards, work sites of all varieties around Australia still cause large numbers of injuries and occasional fatalities.

Every year, 3 million newborns die within the first month of life. Birth asphyxia and other breathing-related conditions are a leading cause of mortality during the neonatal phase. Current diagnostic methods are too sophisticated in terms of equipment, required expertise, and general logistics. Consequently, early detection of asphyxia in newborns is very difficult in many parts of the world, especially in resource-poor settings. We are developing a machine learning system, dubbed Ubenwa, which enables diagnosis of asphyxia through automated analysis of the infant cry. Deployed via smartphone and wearable technology, Ubenwa will drastically reduce the time, cost and skill required to make accurate and potentially life-saving diagnoses.

Improving the quality of end-of-life care for hospitalized patients is a priority for healthcare organizations. Studies have shown that physicians tend to over-estimate prognoses, which in combination with treatment inertia results in a mismatch between patients wishes and actual care at the end of life. We describe a method to address this problem using Deep Learning and Electronic Health Record (EHR) data, which is currently being piloted, with Institutional Review Board approval, at an academic medical center. The EHR data of admitted patients are automatically evaluated by an algorithm, which brings patients who are likely to benefit from palliative care services to the attention of the Palliative Care team. The algorithm is a Deep Neural Network trained on the EHR data from previous years, to predict all-cause 3-12 month mortality of patients as a proxy for patients that could benefit from palliative care. Our predictions enable the Palliative Care team to take a proactive approach in reaching out to such patients, rather than relying on referrals from treating physicians, or conduct time consuming chart reviews of all patients. We also present a novel interpretation technique which we use to provide explanations of the model's predictions.

There is an increasing interest in exploiting mobile sensing technologies and machine learning techniques for mental health monitoring and intervention. Researchers have effectively used contextual information, such as mobility, communication and mobile phone usage patterns for quantifying individuals' mood and wellbeing. In this paper, we investigate the effectiveness of neural network models for predicting users' level of stress by using the location information collected by smartphones. We characterize the mobility patterns of individuals using the GPS metrics presented in the literature and employ these metrics as input to the network. We evaluate our approach on the open-source StudentLife dataset. Moreover, we discuss the challenges and trade-offs involved in building machine learning models for digital mental health and highlight potential future work in this direction.

Mapping the spatial distribution of poverty in developing countries remains an important and costly challenge. These "poverty maps" are key inputs for poverty targeting, public goods provision, political accountability, and impact evaluation, that are all the more important given the geographic dispersion of the remaining bottom billion severely poor individuals. In this paper we train Convolutional Neural Networks (CNNs) to estimate poverty directly from high and medium resolution satellite images. We use both Planet and Digital Globe imagery with spatial resolutions of 3-5 sq. m. and 50 sq. cm. respectively, covering all 2 million sq. km. of Mexico. Benchmark poverty estimates come from the 2014 MCS-ENIGH combined with the 2015 Intercensus and are used to estimate poverty rates for 2,456 Mexican municipalities. CNNs are trained using the 896 municipalities in the 2014 MCS-ENIGH. We experiment with several architectures (GoogleNet, VGG) and use GoogleNet as a final architecture where weights are fine-tuned from ImageNet. We find that 1) the best models, which incorporate satellite-estimated land use as a predictor, explain approximately 57% of the variation in poverty in a validation sample of 10 percent of MCS-ENIGH municipalities; 2) Across all MCS-ENIGH municipalities explanatory power reduces to 44% in a CNN prediction and landcover model; 3) Predicted poverty from the CNN predictions alone explains 47% of the variation in poverty in the validation sample, and 37% over all MCS-ENIGH municipalities; 4) In urban areas we see slight improvements from using Digital Globe versus Planet imagery, which explain 61% and 54% of poverty variation respectively. We conclude that CNNs can be trained end-to-end on satellite imagery to estimate poverty, although there is much work to be done to understand how the training process influences out of sample validation.

The lack of interpretability remains a key barrier to the adoption of deep models in many applications. In this work, we explicitly regularize deep models so human users might step through the process behind their predictions in little time. Specifically, we train deep time-series models so their class-probability predictions have high accuracy while being closely modeled by decision trees with few nodes. Using intuitive toy examples as well as medical tasks for treating sepsis and HIV, we demonstrate that this new tree regularization yields models that are easier for humans to simulate than simpler L1 or L2 penalties without sacrificing predictive power.

HE problem of training a machine learning model in the presence of different training and test distributions is known as domain adaptation [3], [6]-[9]. The goal of domain adaptation is to build a model that performs well on a target distribution while it is trained on a different but related source distribution. One important example is in the sentiment analysis of product reviews [1], where a model is trained on data of a source product category, e. g. kitchen appliances, and it is tested on data of a related category, e. g. books. A second example is the training of image classifiers on unlabeled real images by means of nearly-synthetic images that are fully labeled but which have a distribution that is different [2], [3]. Another example is the content-based depth range adaptation of unlabeled stereoscopic videos by means of labeled data from movies [4], [5]. It is shown in [10], that a classifier's error on the target domain can be bounded in terms of its error on the source domain and a difference between the source and the target domain distribution [10]. This motivated many approaches to first extract features that overcome the distribution difference and subsequently minimize the source error [8], [11]-[13]. With the recent developments in representation learning, approaches have been developed that embed domain adaptation in the feature learning process.