We propose a reference architecture for intelligent context-aware natural speech-enabled systems delivering complex functionality, with direct access to information, simplifying business processes and activities while providing domain-specific and task-specific depth in interactive banking, insurance, wealth management, finance, clinical, legal, telecom customer service, operations, supply chain, connected living room, and personal assistants. This system understands not just words, but intentions, and context of the interaction. We accomplish this through a marriage of speech recognition with advanced natural language processing techniques, scalable inference and semantic technologies. This architecture is expected to dramatically improve the quality of proactive decision support provided by virtual agents by enabling them to seek explanations, make predictions, generate and test hypothesis and perform what-if-analyses using scalable inference engines. The system can provide extreme personalization (N=1) by inferring user intent, making relevant suggestions, maintaining context, carrying out cost-benefit analysis from multiple perspectives, finding similar cases before they are searched for, finds relevant documents and answers, issues resolved by experts in similar situations. The architecture enables meaningfully correlating, finding and connecting people and information sources through discovery of causal, temporal and spatial relations. We present two examples of demonstrations of concept that we are in the process of building out.

We analyze several of the key technical and practical challenges involved in representing activity context across a large variety of knowledge, components, and applications. We present two novel broad methods that enable semantic knowledge capture and interchange, and suggest how they can be used for activity context-awareness. The first is knowledge representation and reasoning (KRR) in Rulelog, an expressively extended form of declarative logic programs that features defeasible higher-order logic formulas yet is computationally tractable, and is a draft dialect of W3C RIF. Rulelog's expressiveness enables representation of exceptions and change, and thus processes, agreements, and policies, e.g., for confidentiality. The second broad method is Textual Logic, an approach to mapping between natural language (text) and logic, where the mapping itself is logic-based. Textual Logic leverages Rulelog's expressiveness to enable relatively rapid text-based authoring of rich knowledge, reducing the knowledge acquisition bottleneck. Together, Rulelog and Textual Logic help address the potential for ontological and KRR Babel that lurks when representing activity context using previous semantic technologies.

Constraint Programming (CP) solvers classically explore the solution space using tree-search based heuristics. Monte-Carlo Tree-Search (MCTS) is a tree-search method aimed at optimal sequential decision making under uncertainty. At the crossroads of CP and MCTS, this paper presents the Bandit Search for Constraint Programming (BASCOP)  algorithm, adapting MCTS to the specifics of CP search trees. Formally, MCTS simultaneously estimates the average node reward, and uses it to bias the exploration towards the most promising regions of the tree, borrowing the multi-armed bandit (MAB) decision rule. The two contributions in BASCOP concern i) a specific reward function, estimating the relative failure depth conditionally to a (variable, value) assignment; ii) a new  decision rule, hybridizing the MAB framework and the spirit of local neighborhood search. Specifically, BASCOP guides the CP search in the neighborhood of the previous best solution, by exploiting statistical estimates gathered across multiple restarts. BASCOP, using Gecode as the underlying constraint solver, shows significant improvements over the depth-first search baseline on some  CP benchmark suites. For hard job-shop scheduling problems, BASCOP matches the results of state-of-the-art scheduling-specific CP approaches. These results demonstrate the potential of BASCOP as a generic yet robust search method for CP.

A central goal of AI is to reason efficiently in domains that are both complex and uncertain. Most attempts toward this end add probability to a tractable subset of first-order logic, but this results in intractable inference. To address this, Domingos and Webb (2012) introduced tractable Markov logic (TML), the first tractable first-order probabilistic representation. Despite its surprising expressiveness, TML has a number ofsignificant limitations. Chief among these is that it does not explicitly handle existence uncertainty, meaning that all possible worlds contain the same objects and relations. This leads to a number of conceptual problems, such as models that must contain meaningless combinations of attributes (e.g.,horses with wheels). Here we propose a new formalism, tractable probabilistic knowledge bases (TPKBs), that overcomes this problem. Like TML, TPKBs use probabilistic class and part hierarchies to ensure tractability, but TPKBs have a much cleaner and user-friendly object-oriented syntax and a well-founded semantics for existence uncertainty. TML is greatly complicated by the use of probabilistic theorem proving, an inference procedure that is much more powerful than necessary. In contrast, we introduce an inference procedure specifically designed for TPKBs, which makes them far more transparent and amenable to analysis and implementation. TPKBs subsume TML and therefore essentially all tractable models, including many high-treewidth ones.

Patterns of life (POL) are emergent properties of complex social systems. Computational models of POL offer significant potential for practical application and theoretical study, but also important challenges for AI research. Computational POL models must achieve simultaneous scalability along three key dimensions: population size, intelligence, and automatic behavior specification. Three broad research areas that could support important improvements in POL modeling are pattern recognition, representational abstraction, and behavior generation with intelligent agents and the like. This paper describes challenges in POL modeling that AI researchers from many fields can help to meet.

Network medicine is an emerging paradigm for studying the co-occurrence between diseases. While diseases are often interlinked through complex patterns, most of the existing work in this area has focused on studying pairwise relationships between diseases. In this paper, we use a state-of-the-art Markov network learning method to learn interactions between musculoskeletal disorders and cardiovascular diseases and compare this to pairwise approaches. Our experimental results confirm that the sophisticated structure learner produces more accurate models, which can help reveal interesting patterns in the co-occurrence of diseases.

Network medicine is an emerging paradigm for studying the co-occurrence between diseases. While diseases are often interlinked through complex patterns, most of the existing work in this area has focused on studying pairwise relationships between diseases. In this paper, we use a state-of-the-art Markov network learning method to learn interactions between musculoskeletal disorders and cardiovascular diseases and compare this to pairwise approaches. Our experimental results confirm that the sophisticated structure learner produces more accurate models, which can help reveal interesting patterns in the co-occurrence of diseases.

Network medicine is an emerging paradigm for studying the co-occurrence between diseases. While diseases are often interlinked through complex patterns, most of the existing work in this area has focused on studying pairwise relationships between diseases. In this paper, we use a state-of-the-art Markov network learning method to learn interactions between musculoskeletal disorders and cardiovascular diseases and compare this to pairwise approaches. Our experimental results confirm that the sophisticated structure learner produces more accurate models, which can help reveal interesting patterns in the co-occurrence of diseases.

Network medicine is an emerging paradigm for studying the co-occurrence between diseases. While diseases are often interlinked through complex patterns, most of the existing work in this area has focused on studying pairwise relationships between diseases. In this paper, we use a state-of-the-art Markov network learning method to learn interactions between musculoskeletal disorders and cardiovascular diseases and compare this to pairwise approaches. Our experimental results confirm that the sophisticated structure learner produces more accurate models, which can help reveal interesting patterns in the co-occurrence of diseases.

Network medicine is an emerging paradigm for studying the co-occurrence between diseases. While diseases are often interlinked through complex patterns, most of the existing work in this area has focused on studying pairwise relationships between diseases. In this paper, we use a state-of-the-art Markov network learning method to learn interactions between musculoskeletal disorders and cardiovascular diseases and compare this to pairwise approaches. Our experimental results confirm that the sophisticated structure learner produces more accurate models, which can help reveal interesting patterns in the co-occurrence of diseases.