Whether we're ready for it or not, artificial intelligence (AI) is infiltrating the enterprise. From natural language processing systems that manage customer service inquiries to automated manufacturing plants staffed by robots, AI technologies, driven by machine learning, are having an impact. And with the accelerated rate of innovation--brought on by exponential increases in computer processing power and the sheer volume of data creation--AI clearly has the potential to transform just about every industry, from aerospace to retail. In a Harvard Business Review article, MIT researchers Erik Brynjolfsson and Andrew McAfee labeled AI "the most important general-purpose technology of our era." The effects of AI will be magnified in the coming decade, they say, as organizations "transform their core processes and business models to take advantage of machine learning."

Recurrent neural networks are nowadays successfully used in an abundance of applications, going from text, speech and image processing to recommender systems. Backpropagation through time is the algorithm that is commonly used to train these networks on specific tasks. Many deep learning frameworks have their own implementation of training and sampling procedures for recurrent neural networks, while there are in fact multiple other possibilities to choose from and other parameters to tune. In existing literature this is very often overlooked or ignored. In this paper we therefore give an overview of possible training and sampling schemes for character-level recurrent neural networks to solve the task of predicting the next token in a given sequence. We test these different schemes on a variety of datasets, neural network architectures and parameter settings, and formulate a number of take-home recommendations. The choice of training and sampling scheme turns out to be subject to a number of trade-offs, such as training stability, sampling time, model performance and implementation effort, but is largely independent of the data. Perhaps the most surprising result is that transferring hidden states for correctly initializing the model on subsequences often leads to unstable training behavior depending on the dataset.

At a recent conference in 2017, Microsoft CEO Satya Nadella used the analogy of a corn maze to explain the difference in approach between a classical computer and a quantum computer. In trying to find a path through the maze, a classical computer would start down a path, hit an obstruction, backtrack; start again, hit another obstruction, backtrack again until it ran out of options. Although an answer can be found, this approach could be a very time-consuming. They take every path in the corn maze simultaneously." Thus, leading to an exponential reduction in the number of steps required to solve a problem.

Rapid advances of hardware-based technologies during the past decades have opened up new possibilities for Life scientists to gather multimodal data in various application domains (e.g., Omics, Bioimaging, Medical Imaging, and [Brain/Body]-Machine Interfaces), thus generating novel opportunities for development of dedicated data intensive machine learning techniques. Overall, recent research in Deep learning (DL), Reinforcement learning (RL), and their combination (Deep RL) promise to revolutionize Artificial Intelligence. The growth in computational power accompanied by faster and increased data storage and declining computing costs have already allowed scientists in various fields to apply these techniques on datasets that were previously intractable for their size and complexity. This review article provides a comprehensive survey on the application of DL, RL, and Deep RL techniques in mining Biological data. In addition, we compare performances of DL techniques when applied to different datasets across various application domains. Finally, we outline open issues in this challenging research area and discuss future development perspectives.

Optimization of high-dimensional black-box functions is an extremely challenging problem. While Bayesian optimization has emerged as a popular approach for optimizing black-box functions, its applicability has been limited to low-dimensional problems due to its computational and statistical challenges arising from high-dimensional settings. In this paper, we propose to tackle these challenges by (1) assuming a latent additive structure in the function and inferring it properly for more efficient and effective BO, and (2) performing multiple evaluations in parallel to reduce the number of iterations required by the method. Our novel approach learns the latent structure with Gibbs sampling and constructs batched queries using determinantal point processes. Experimental validations on both synthetic and real-world functions demonstrate that the proposed method outperforms the existing state-of-the-art approaches.

Massive amount of electronic medical records accumulating from patients and populations motivates clinicians and data scientists to collaborate for the advanced analytics to extract knowledge that is essential to address the extensive personalized insights needed for patients, clinicians, providers, scientists, and health policy makers. In this paper, we propose a new predictive approach based on feature representation using deep feature learning and word embedding techniques. Our method uses different deep architectures for feature representation in higher-level abstraction to obtain effective and more robust features from EMRs, and then build prediction models on the top of them. Our approach is particularly useful when the unlabeled data is abundant whereas labeled one is scarce. We investigate the performance of representation learning through a supervised approach. First, we apply our method on a small dataset related to a specific precision medicine problem, which focuses on prediction of left ventricular mass indexed to body surface area (LVMI) as an indicator of heart damage risk in a vulnerable demographic subgroup (African-Americans). Then we use two large datasets from eICU collaborative research database to predict the length of stay in Cardiac-ICU and Neuro-ICU based on high dimensional features. Finally we provide a comparative study and show that our predictive approach leads to better results in comparison with others.

Editor: On "Learning Language" I was dismayed by the inclusion of William Katke's article ("Learning Language Using A Pattern Recognition Approach," Spring 1985). Usually you do an excellent job of representing "the current state of the art in Artificial Intelligence" (to quote your Editorial Policy), but I consider this article an exception. First of all, although the article claims to be on "Learning Language," what it presents is at best a knowledge-free approach to learning syntax. I saw no evidence that the induced syntax is useful for anything, and good reasons to believe that it is not, such as the unmnemonic category names and the intrinsic limitations of finite state grammars. Second, this kind of stuff has been done before, and it didn't work too well then either; for a useful overview of the field and pointers into the literature, see the article on "Grammatical Inference" in Volume 3 of The Handbook of The plete specifications and the verification of proposed impleideas and issues presented were firmly focused on a conven-mentations, we should concentrate more on incremental tional view of the design process-a view I can caricaturize development of specifications as a result of assessment of as the SPIV methodology: performance.

The Seventh International Conference on Intelligent User Interfaces (IUI-2003) was held 12-15 January 2003 in Miami Beach, Florida. The conference brought together researchers and practitioners to report on outstanding research and applications, examine emerging work, and delineate new avenues for intelligent user interfaces. The conference received an all-time record number of submissions, covering a wide range of areas and approaches. Oral presentation sessions were organized into seven major topics: (1) adaptive and collaborative interfaces, (2) affective interfaces, (3) agent-based interfaces, (4) knowledge acquisition and visualization, (5) model-based interface design, (6) multimodal input, and (7) natural language interfaces. The conference program included three invited talks, each reflecting a different direction for developing the intelligent interfaces of the future.

Hence, at a coarse-grained level of abstraction, KB-Ss can be characterized in terms of two components: (1) a knowledge base, encoding the knowledge embodied by the system, and (2) a reasoning engine, which is able to query the knowledge base, infer or acquire knowledge from external sources, and add new knowledge to the knowledge base. A knowledge-level account of a KBS (that is, a competencecentered, implementation-independent description of a system), such as Clancey's (1985) analysis of first-generation rule-based systems, focuses on the task-centered competence of the system; that is, it addresses issues such as what kind of problems the KBS is designed to tackle, what reasoning methods it uses, and what knowledge it requires. In contrast with task-centered analyses, Levesque and Lakemeyer focus on the competence of the knowledge base rather than that of the whole system. Hence, their notion of competence is a task-independent one: It is the "abstract state of knowledge" (p. This is an interesting assumption, which the "proceduralists" in the AI community might object to: According to the procedural viewpoint of knowledge representation, the knowledge modeled in an application, its representation, and the associated knowledge-retrieval mechanisms have to be engineered as As a result, they would argue, it is not possible to discuss the knowledge of a system independently of the task context in which the system is meant to operate.

They thought that the Chinese Room argument showed that computationalism could never fully account for the first-person perspective, that the "computer metaphor for the mind" might lead to some vital social questions being ignored, that passing the Turing Test They conducted 20 interviews with a rather idiosyncratic collection of people, largely on the east and west coasts, to find out what the consensus was in the field. One of their happy discoveries was that connectionism (about which they initially knew little) was expected to overcome many of these obstacles. Each interview begins with a brief personal history of why the interviewee became involved with the subject and what they take it to be, and then moves into a discussion of contemporary issues which the editors find interesting. While the interviews do not conform to a set pattern, they return regularly to a few favorite themes: the Chinese Room, the importance of the Turing Test, why "symbolic AI" has failed (a claim that is made repeatedly throughout the book), and the significance of connectionism as a replacement for it Wilensky, and Winograd could possibly be said to be active in mainstream AI; on the other hand there are seven or eight philosophers, of whom only Dennett has a sympathetic interest in AI; all the others have rejected its premises, and Dreyfus, Searle and Weizenbaum are notorious for their passionate and sustained attacks on the subject. This would be less important but for the fact that AI is the main subject matter of several of the interviews.