The "Black Box" of algorithms, particularly deep learning algorithms make grasping how decisions are made virtually impossible. These include trading decisions, investment decisions, and risk management decisions. The communication mechanisms inside the AI System is not transparent. When money is lost, it's difficult for the hedge fund or the regulatory body to reconcile that loss to any foul-play. If the AI System is at the center, then it is the AI System that is responsible.

Nielsen (NLSN) and the University of Oxford today announced a two-year collaboration to advance the use of artificial intelligence (AI) to identify and classify consumer packaged goods (CPG) products on shelves in retail stores. Facilitated between Nielsen's Image Recognition group and the Visual Geometry Group (VGG) at the University of Oxford, this partnership brings together the world's largest pool of product reference data with industry-leading brainpower around AI technology to yield greater accuracy in product identification and discovery. Through this partnership, Nielsen is working directly with University of Oxford Professors Andrew Zisserman and Andrea Vedaldi (Department of Engineering Science), world-renowned computer scientists and pioneers in image recognition and AI research. Zisserman, Vedaldi and their team of research scientists will work together with Nielsen to more precisely and quickly identify and classify in-store products based on product images captured through Nielsen's eCollection solution. The Oxford researchers will focus on building and enhancing the eCollection algorithms with increasingly advanced deep learning capabilities, enabling a more automatic detection of store products, promotions and prices without the need for manual intervention.

Variations in the El Niño/Southern Oscillation (ENSO) are associated with a wide array of regional climate extremes and ecosystem impacts1. Robust, long-lead forecasts would therefore be valuable for managing policy responses. But despite decades of effort, forecasting ENSO events at lead times of more than one year remains problematic2. Here we show that a statistical forecast model employing a deep-learning approach produces skilful ENSO forecasts for lead times of up to one and a half years. To circumvent the limited amount of observation data, we use transfer learning to train a convolutional neural network (CNN) first on historical simulations3 and subsequently on reanalysis from 1871 to 1973.

THURSDAY, Sept. 26, 2019 (HealthDay News) -- Early evidence suggests that diagnostic performance of deep learning models is equivalent to that of health care professionals for interpreting medical imaging, according to a study published online Sept. 25 in The Lancet Digital Health. Xiaoxuan Liu, M.B.Ch.B., from the University Hospitals Birmingham NHS Foundation Trust in the United Kingdom, and colleagues conducted a systematic review and meta-analysis to assess the diagnostic accuracy of deep learning algorithms versus health care professionals in classifying disease using medical imaging. Binary diagnostic accuracy data were extracted and contingency tables were constructed to derive the outcomes of interest: sensitivity and specificity. Data from 82 studies, describing 147 patient cohorts were included. The researchers found that based on 69 studies, sensitivity ranged from 9.7 to 100 percent and specificity ranged from 38.9 to 100 percent.

Welcome to TechTalks' AI book reviews, a series of posts that explore the latest literature on AI. The media is rife with stories that warn of AI algorithms bringing people back from the dead, AI algorithms developing secret languages, mass technological unemployment, and a looming robot apocalypse. Movies and TV series like Her, The Circleand Westworld,which present a mystic portrayal of conscious machines and human-level AI being just around the corner. Rebooting AI is a refreshing read and a much-needed reality check on the current confusing state of artificial intelligence. Consider the following text, mentioned in Rebooting AI: "Elsie tried to reach her aunt on the phone, but she didn't answer." You don't need to be a genius to quickly make the following assumptions after reading this sentence: But even the most sophisticated AI algorithm would struggle to draw the same conclusions.

Researchers have found a way to marry human creativity and artificial intelligence (AI) creativity to dramatically boost the performance of deep learning. A team led by Alexander Wong, a Canada Research Chair in the area of AI and a professor of systems design engineering at the University of Waterloo, developed a new type of compact family of neural networks that could run on smartphones, tablets, and other embedded and mobile devices. The networks, called AttoNets, are being used for image classification and object segmentation, but can also act as the building blocks for video action recognition, video pose estimation, image generation, and other visual perception tasks. "The problem with current neural networks is they are being built by hand and incredibly large and complex and difficult to run in any real-world situation," said Wong, who also co-founded a startup named DarwinAI to commercialize the technology. "These on-the-edge networks are small and agile and could have huge implications for the automotive, aerospace, agriculture, finance, and consumer electronics sectors."

Artificial Intelligence (AI) is a broad term that can apply to various computing tasks, including machine learning, deep learning, and big data analytics. Many AI projects are in a proof of concept stage, but CIOs and IT Managers need to understand that in the future, almost every business outcome and workflow will use and depend upon some form of AI processing. The time is now to prepare the Infrastructure for that eventuality. As AI environments move into production and begin to grow in size and importance, organizations need a strategy to address challenges the AI at scale will create for both the compute and storage architectures. For the last decade, developing a cloud strategy was at the top of every CIO's to-do list.

AI heralds a step-change in the performance and capability of wireless networks and other critical infrastructures. However, it may also cause irreversible environmental damage due to their high energy consumption. Here, we address this challenge in the context of 5G and beyond, where there is a complexity explosion in radio resource management (RRM). On the one hand, deep reinforcement learning (DRL) provides a powerful tool for scalable optimization for high dimensional RRM problems in a dynamic environment. On the other hand, DRL algorithms consume a high amount of energy over time and risk compromising progress made in green radio research. This paper reviews and analyzes how to achieve green DRL for RRM via both architecture and algorithm innovations. Architecturally, a cloud based training and distributed decision-making DRL scheme is proposed, where RRM entities can make lightweight deep local decisions whilst assisted by on-cloud training and updating. On the algorithm level, compression approaches are introduced for both deep neural networks and the underlying Markov Decision Processes, enabling accurate low-dimensional representations of challenges. To scale learning across geographic areas, a spatial transfer learning scheme is proposed to further promote the learning efficiency of distributed DRL entities by exploiting the traffic demand correlations. Together, our proposed architecture and algorithms provide a vision for green and on-demand DRL capability.

October 14, 2019 A BSTRACT Expressivity is one of the most significant issues in assessing neural networks. In this paper, we provide a quantitative analysis of the expressivity from dynamic models, where Hilbert space is employed to analyze its convergence and criticality. From the feature mapping of several widely used activation functions made by Hermite polynomials, We found sharp declines or even saddle points in the feature space, which stagnate the information transfer in deep neural networks, then present an activation function design based on the Hermite polynomials for better utilization of spatial representation. Moreover, we analyze the information transfer of deep neural networks, emphasizing the convergence problem caused by the mismatch between input and topological structure. We also study the effects of input perturbations and regularization operators on critical expressivity. Finally, we verified the proposed method by multivariate time series prediction. The results show that the optimized DeepESN provides higher predictive performance, especially for long-term prediction. Our theoretical analysis reveals that deep neural networks use spatial domains for information representation and evolve to the edge of chaos as depth increases. In actual training, whether a particular network can ultimately arrive that depends on its ability to overcome convergence and pass information to the required network depth. K eywords Deep neural networks; expressivity; criticality theory; convergence; activation function; Hilbert transform 1 Introduction Deep neural networks (DNNs) have achieved outstanding performance in many fields, from the automatic translation to speech and image recognition [1, 2].

In the domain of machine learning, Neural Memory Networks (NMNs) have recently achieved impressive results in a variety of application areas including visual question answering, trajectory prediction, object tracking, and language modelling. However, we observe that the attention based knowledge retrieval mechanisms used in current NMNs restricts them from achieving their full potential as the attention process retrieves information based on a set of static connection weights. This is suboptimal in a setting where there are vast differences among samples in the data domain; such as anomaly detection where there is no consistent criteria for what constitutes an anomaly. In this paper, we propose a plastic neural memory access mechanism which exploits both static and dynamic connection weights in the memory read, write and output generation procedures. We demonstrate the effectiveness and flexibility of the proposed memory model in three challenging anomaly detection tasks in the medical domain: abnormal EEG identification, MRI tumour type classification and schizophrenia risk detection in children. In all settings, the proposed approach outperforms the current state-of-the-art. Furthermore, we perform an in-depth analysis demonstrating the utility of neural plasticity for the knowledge retrieval process and provide evidence on how the proposed memory model generates sparse yet informative memory outputs.