IBM's approach to AI is probably best described as "comprehensive." Watson, the company's flagship deep learning system, has been everywhere from the Grammy Awards to the International Space Station. In its spare time the company also uses AI to save the planet. AI has quickly become an important tool for environmental scientists and researchers trying to reverse climate change, develop clean energy, and revolutionize agriculture. Across the globe, IBM's researchers are creating solutions to some of the biggest problems our planet and species face today, and machine learning is a huge part of its efforts.

Due to the increasing deployment of Deep Neural Networks (DNNs) in real-world security-critical domains including autonomous vehicles and collision avoidance systems, formally checking security properties of DNNs, especially under different attacker capabilities, is becoming crucial. Most existing security testing techniques for DNNs try to find adversarial examples without providing any formal security guarantees about the non-existence of adversarial examples. Recently, several projects have used different types of Satisfiability Modulo Theory (SMT) solvers to formally check security properties of DNNs. However, all of these approaches are limited by the high overhead caused by the solver. In this paper, we present a new direction for formally checking security properties of DNNs without using SMT solvers. Instead, we leverage interval arithmetic to formally check security properties by computing rigorous bounds on the DNN outputs. Our approach, unlike existing solver-based approaches, is easily parallelizable. We further present symbolic interval analysis along with several other optimizations to minimize overestimations. We design, implement, and evaluate our approach as part of ReluVal, a system for formally checking security properties of Relu-based DNNs. Our extensive empirical results show that ReluVal outperforms Reluplex, a state-of-the-art solver-based system, by 200 times on average for the same security properties. ReluVal is able to prove a security property within 4 hours on a single 8-core machine without GPUs, while Reluplex deemed inconclusive due to timeout (more than 5 days). Our experiments demonstrate that symbolic interval analysis is a promising new direction towards rigorously analyzing different security properties of DNNs.

In mulsemedia applications, traditional media content (text, image, audio, video, etc.) can be related to media objects that target other human senses (e.g., smell, haptics, taste). Such applications aim at bridging the virtual and real worlds through sensors and actuators. Actuators are responsible for the execution of sensory effects (e.g., wind, heat, light), which produce sensory stimulations on the users. In these applications sensory stimulation must happen in a timely manner regarding the other traditional media content being presented. For example, at the moment in which an explosion is presented in the audiovisual content, it may be adequate to activate actuators that produce heat and light. It is common to use some declarative multimedia authoring language to relate the timestamp in which each media object is to be presented to the execution of some sensory effect. One problem in this setting is that the synchronization of media objects and sensory effects is done manually by the author(s) of the application, a process which is time-consuming and error prone. In this paper, we present a bimodal neural network architecture to assist the synchronization task in mulsemedia applications. Our approach is based on the idea that audio and video signals can be used simultaneously to identify the timestamps in which some sensory effect should be executed. Our learning architecture combines audio and video signals for the prediction of scene components. For evaluation purposes, we construct a dataset based on Google's AudioSet. We provide experiments to validate our bimodal architecture. Our results show that the bimodal approach produces better results when compared to several variants of unimodal architectures.

A restricted Boltzmann machine (RBM) is a generative neural-network model with many novel applications such as collaborative filtering and acoustic modeling. An RBM lacks the capacity to retain memory, making it inappropriate for dynamic data modeling as in time-series analysis. In this paper we address this issue by proposing the p-RBM model, a generalization of the regular RBM model, capable of retaining memory of p past states. We further show how to train the p-RBM model using contrastive divergence and test our model on the problem of predicting the stock market direction considering 100 stocks of the NASDAQ-100 index. Obtained results show that the p-RBM offer promising prediction potential.

We study critera for a pair $ (\{ X_n \} $, $ \{ Y_n \}) $ of approximating processes which guarantee closeness of moments by generalizing known results for the special case that $ Y_n = Y $ for all $n$ and $ X_n $ converges to $Y$ in probability. This problem especially arises when working with surrogate models, e.g. to enrich observed data by simulated data, where the surrogates $Y_n$'s are constructed to justify that they approximate the $ X_n $'s. The results of this paper deal with sequences of random variables. Since this framework does not cover many applications where surrogate models such as deep neural networks are used to approximate more general stochastic processes, we extend the results to the more general framework of random fields of stochastic processes. This framework especially covers image data and sequences of images. We show that uniform integrability is sufficient, and this holds even for the case of processes provided they satisfy a weak stationarity condition.

Structural data(age, gender, weight) from Electronic Health Records (EHRs) are exploited by Computer Aided Systems (CADS) as complementary information for disease prediction. Such systems, however, fail to weight the structural data based its relevance to the disease at hand. A model capable of evaluating the significance of every element of the structural data and performing the prediction task based on the selective and weighted procedure for elements of structural data is required. Such scheme will boost more semantic automatic disease prediction task Recently multi-modal data is processed using deep learning methods like Convolutional Neural Networks(CNNs)[9], Autoencoders[6], Modified Restricted Boltzman Machine[8] etc. These methods provide richer and discriminant feature space which helps to exploit the global complementary information from available modalities, however, fail to address the problem of unequal relevance. Structural data gives statistical information about the population as a whole. This is taken into consideration more recently using graphs, providing a more semantic way of using multi-modal data[7, 4]. These methods focus more on the association between the subjects with respect to either of the modalities and then solve the tasks such as disease prediction with features from other modalities.

Waze has struck a data-sharing agreement with Waycare, an artificial intelligence-based traffic management startup, the two companies announced today. The deal will allow them to combine anonymized navigation information crowdsourced from the 100 million drivers who use Waze with Waycare's proprietary traffic analytics. The collaboration is now active in Nevada, Florida, California and Nevada, with plans to expand over the next year. It is part of Waze's Connected Citizens Program, which gives cities around the world access to anonymized driver data to help them manage traffic and road infrastructure. A representative told TechCrunch that data supplied by Waycare to Waze will be incorporated into the app's usual interface, while data from Waze will be added to Waycare's platform alongside its other data sets. Founded in 2016, Waycare is a cloud-based platform that enables municipalities to gather data from many sources, including on-board devices, navigation apps, sensors and road camera feeds, and analyze them using proprietary deep learning algorithms to figure out how to improve traffic and road conditions.

Artificial intelligence is coming to a data center near you, and it will likely start performing many of the tasks that human operators spend the bulk of their time doing. But rather than view this inevitable development as a threat, today's IT worker would do better to learn the fundamentals of AI now so that when it does arrive it can be used as a tool to enhance the value of human effort to the organization, not replace it. First off, it helps to know that there are many different types of AI that serve various functions. Tech journalist Michael Copeland views the technology as a series of concentric circles, with AI as the outermost circle and more specialized forms like machine learning (ML) and deep learning falling within. Join nearly 200,000 subscribers who receive actionable tech insights from Techopedia. The differences lie in the levels of complexity exhibited by each form of AI and the specific functions they are designed to enable.

In March, an AI event was held at the Computer History Museum entitled --ASICs Unlock Deep Learning Innovation.-- Along with Samsung, Amkor Technology and Northwest Logic, eSilicon explored how these companies form an ecosystem to develop deep learning chips for the next generation of AI applications. There was also a keynote presentation on deep learning from Ty Garibay, CTO of Arteris IP. Over 100 people showed up, including myself, for an afternoon and evening of deep learning exploration and some good food, wine and beer as well. The audience spanned chip companies, major OEMs, emerging deep learning startups and research folks from both a hardware and data science/algorithm point of view.

Next time you're at King's Cross station, take a moment to think about this. Just yards from where you're standing, the world's most advanced artificial intelligence (AI) technology is being developed -- by a London company called DeepMind. You might assume that when it comes to AI, like lots of other new technologies, it's Silicon Valley or Israel that's out in front. The truth is that something incredibly special is happening in our city right now. Today London is the global powerhouse of AI -- the fiendishly difficult task of getting software and machines to perform tasks that require human intelligence to do.