Working memory is an essential component of reasoning -- the capacity to answer a new question by manipulating acquired knowledge. Current memory-augmented neural networks offer a differentiable method to realize limited reasoning with support of a working memory module. Memory modules are often implemented as a set of memory slots without explicit relational exchange of content. This does not naturally match multi-relational domains in which data is structured. We design a new model dubbed Relational Dynamic Memory Network (RDMN) to fill this gap. The memory can have a single or multiple components, each of which realizes a multi-relational graph of memory slots. The memory is dynamically updated in the reasoning process controlled by the central controller. We evaluate the capability of RDMN on several important application domains: software vulnerability, molecular bioactivity and chemical reaction. Results demonstrate the efficacy of the proposed model.

U.S. Bank's Chief Analytics Officer Bill Hoffman thinks the term "artificial intelligence" is a bit of a misnomer – he prefers to think about AI as a customized experience. "The'A' in AI should be'augmented,' not artificial," he said in an interview last year with CMO Australia. "There's nothing artificial about building a high quality, personalized relationship with a customer." Hoffman's less artificial approach to AI is transforming how U.S. Bank interacts with its customers and builds quality relationships with people. Hoffman was instrumental in U.S. Bank's adoption of Einstein, an AI platform offered by Salesforce, late last year.

Today Dell EMC announced new Ready Solutions for AI. With specialized designs for Machine Learning with Hadoop and Deep Learning with NVIDIA, the Dell EMC Ready Solutions simplify AI environments, deliver faster, deeper insights than the competition1, and leverage Dell EMC's proven expertise to help organizations realize the full potential of AI. There's no doubt that AI is the future, and our customers are preparing for it now," said Tom Burns, senior vice president, Networking & Solutions, Dell EMC. "Our goal is to lead the industry with the most powerful and fully-integrated AI solutions. What we're announcing today allows customers at any scale to start seeing better business outcomes and positions them for AI's increasingly important role in the future." Emerging technologies such as AI will transform lives and how people work and conduct business over the next decade. According to Dell Technologies' research with 3,800 business leaders around the globe, conducted in partnership with VansonBourne, nearly 80% of organizations will be investing in advanced AI technologies within the next five years. AI is increasingly a strategic priority for most organizations. However, deploying and managing AI workloads is complex, costly, and requires extensive integration and testing of the hardware and software. The new Dell EMC Ready Solutions for AI were built to simplify AI, deliver faster, deeper insights, and leverage Dell EMC's proven AI expertise. Organizations no longer have to individually source and piece together their own solutions. Instead, they can rely on a Dell EMC-designed and validated set of best-of-breed technologies for software – including AI frameworks and libraries – with compute, networking and storage. Dell EMC's portfolio of services from consulting to deployment, support and education helps customers drive the rapid adoption and optimization of their AI environments. In this video from the Dell EMC HPC Community Meeting, Jay Boisseau from Dell EMC describes how the company is moving forward as a thought leader in Artificial Intelligence. Dell EMC and NVIDIA engineered this deep learning design to be built around Dell EMC PowerEdge servers with NVIDIA Tesla V100 Tensor Core GPUs. AI is being driven by leaps in GPU computing power that defy the slowdown in Moore's Law," said Ian Buck, vice president and general manager, Accelerated Computing Group, NVIDIA.

With all the information flying around in businesses today, M-Files has been helping businesses succeed by improving efficiency and maximizing the reusability of information by using M-Files' unique metadata-driven architecture, you can find the right document instantly with a simple keyword search. Recently, they have secured a EUR 27 million financing agreement which they plan to use for international growth, partner channel expansion along with accelerating their R&D. I was able to reach out to Miika Mäkitalo, CEO of M-Files to ask a few questions about what they plan to do with the funding. Miika: "M-Files will use this newly secured financing to fuel our continued international growth, partner channel expansions and accelerated R&D in Europe. As the need for intelligent information management solutions grows rapidly, we are expanding our global footprint and our offices in the UK, Germany, France, Australia and the U.S. M-Files grew revenue by almost 40 percent in 2017 over 2016, including strong growth in both direct sales and through our global partner network, now numbering more than 600 partners worldwide. This financing will continue to fuel that momentum. M-Files will also use this funding to advance our intelligent information management platform, so we can continue to redefine how companies around the world manage information and data."

There is considerable hype around data science. Websites and social media are flooded with articles on Big Data, Data Science, and Data Analytics. These fields are projected as top fields, while data scientists are considered as saviours of the world and hence are supposed to be highest paid professionals. Most articles project how, using data science, companies have become super-intelligent in understanding their customers and hence are now able to sell products and services in a sophisticated manner. The marketing is now based on customers' differential purchasing interest and buying behaviour.

I'm sitting in Anki's head office in San Francisco, trying to work out who's sawing logs. Turns out, it's coming from a palm-sized robot called Vector, sitting on a charging dock getting some shut-eye. Anki is the company behind Cozmo, an AI-equipped toy robot that's sold over 1.5 million units. Unlike the kid-friendly Cozmo, Vector is designed for the entire family to interact with in the home. Home robots are familiar to many of us from the realms of cartoons and science fiction, like Rosie from The Jetsons or even R2-D2. But the fiction is a long way from the technological reality today.

Bio: Shailendra is a keynote speaker, influencer and a thought-leader in the Artificial Intelligence space . With an experience of over 23 years working with Corporates, Software Vendors and Consulting companies, Shailendra has delivered over One Billion Dollars through advanced analytics. He has established and lead several data science businesses and teams to generate revenue and drive incremental growth by creating multiple Artificial Intelligence solutions across a variety of sectors, including Telcos, Financial Services, Retail and Public Sector. Shailendra has published multiple articles about Advanced Analytics, Machine Learning, IoT, Artificial Intelligence and Blockchain; and recently published an Amazon bestseller "Making Money Out of Data" which showcases five business stories from various industries on how successful companies make millions of dollars in incremental value using analytics. Shailendra has held senior executive level positions at SAP, IBM, Accenture, Woolworths and Coles with Asia and ANZ mandates.

Affective computing [31] is "computing that relates to, arises from, or influences emotions." It is very important in human-machine interaction, as humans cannot have longlasting intimate relationships with machines if they cannot understand our affects and respond appropriately. Both affect classification and regression have been extensively studied in the literature [24], [43], [45], [46], [48]. For affect classification, the most commonly used categories are the six basic emotions (anger, disgust, fear, happiness, sadness, and surprise) proposed by Ekman et al. [5]. For regression, affects are usually represented as numbers in the 2D space of arousal and valence [35], or in the 3D space of arousal, valence, and dominance [25]. Recently, Yannakakis et al. [50] also argued that the nature of emotions is ordinal, and hence preference learning [51] should also play an important role in affective computing. Various input signals could be used in affective computing, e.g., speech [21], [47], facial expressions [8], [29], physiological signals [7], [43], and multimodal combination [26], [53]. Numerous features could be extracted from each modality. For example, 6,373 acoustic features were extracted by OpenSMILE [6] in the InterSpeech 2013 Computational Paralinguistics Challenge.

Regression problems are pervasive in real-world applications. Generally a substantial amount of labeled samples are needed to build a regression model with good generalization ability. However, many times it is relatively easy to collect a large number of unlabeled samples, but time-consuming or expensive to label them. Active learning for regression (ALR) is a methodology to reduce the number of labeled samples, by selecting the most beneficial ones to label, instead of random selection. This paper proposes two new ALR approaches based on greedy sampling (GS). The first approach (GSy) selects new samples to increase the diversity in the output space, and the second (iGS) selects new samples to increase the diversity in both input and output spaces. Extensive experiments on 12 UCI and CMU StatLib datasets from various domains, and on 15 subjects on EEG-based driver drowsiness estimation, verified their effectiveness and robustness.

Acquisition of labeled training samples for affective computing is usually costly and time-consuming, as affects are intrinsically subjective, subtle and uncertain, and hence multiple human assessors are needed to evaluate each affective sample. Particularly, for affect estimation in the 3D space of valence, arousal and dominance, each assessor has to perform the evaluations in three dimensions, which makes the labeling problem even more challenging. Many sophisticated machine learning approaches have been proposed to reduce the data labeling requirement in various other domains, but so far few have considered affective computing. This paper proposes two multi-task active learning for regression approaches, which select the most beneficial samples to label, by considering the three affect primitives simultaneously. Experimental results on the VAM corpus demonstrated that our optimal sample selection approaches can result in better estimation performance than random selection and several traditional single-task active learning approaches. Thus, they can help alleviate the data labeling problem in affective computing, i.e., better estimation performance can be obtained from fewer labeling queries.