BROWSE TOPICS
RESOURCESABOUT THIS SITE |
Real-Time ReasoningWhen Time is Limited and Data are Streaming In AITopics > Reasoning > Real-Time Reasoning  "...real-time AI systems are required to work continuously over extended periods of time, interface to the external environment via sensors and actuators, deal with uncertain or missing data, focus resources on the most critical events, handle both synchronous and asynchronous events in a predictable fashion with guaranteed response times, and degrade gracefully." AAAI/KDD/UAI-2002 Joint Workshop on Real-Time Decision Support and Diagnosis Systems. "Workshop Description:While AI methodologies are being applied towards increasingly realistic domains that require timely responses, real-time systems are coming to incorporate decision-making tools that require more intelligent capabilities. Many real-world intelligent systems call for autonomous intelligent agents acting in the face of uncertain knowledge and limited computational resources. Real-time decision support and diagnosis systems are two such important application domains." Papers, abstracts, presentations and talks can be accessed. Demonstration of Music Plus One - A Real-Time System for Automatic Orchestral Accompaniment. Christopher Raphael. 2006. In Proceedings of the Twenty-First National Conference on Artificial Intelligence, 1951 - 1952. Menlo Park, Calif.: AAAI Press."We demonstrate a system that creates a real-time accompaniment for a live musician performing a non-improvisatory piece of music. The system listens to the live player by performing a hidden Markov model analysis of the player's acoustic signal. A belief network uses this information, a musical score, and past rehearsals, to create a sequence of evolving predictions for future note-onsets in the soloist and accompaniment. These predictions are used to guide the time-stretched resynthesis of prerecorded orchestral audio using a phase vocoder."
The Challenges of Real-Time AI. By David J. Musliner, James A. Hendler, Ashok K. Agrawala, Edmund H. Durfee, Jay K. Strosnider, and C. J. Paul. Computer Volume 28; Number 1: 58-66(January 1995). Excerpt from the abstract: "The research agendas of artificial intelligence and real-time systems are converging as AI methods move toward domains that require real-time responses, and real-time systems move toward complex applications that require intelligent behavior. They meet at the crossroads in an exciting new subfield commonly called 'real-time AI.' This subfield is still being defined, and the precise goals for various real-time AI systems are in flux. Traditionally, AI systems have been developed without much attention to the resource limitations that motivate real-time systems researchers. However, as these AI systems move from the research labs into real-world applications, they also become subject to the time constraints of the environments in which they operate. Rigorous design techniques developed by real-time systems re-searchers must be used to guarantee that a system will meet domain deadlines, even in worst-case scenarios, particularly for mission-critical assignments.". A Robot in Every Home - The leader of the PC revolution predicts that the next hot field will be robotics. By Bill Gates. Scientific American (January 2007). "Concurrency is a challenge that extends beyond robotics. Today as more and more applications are written for distributed networks of computers, programmers have struggled to figure out how to efficiently orchestrate code running on many different servers at the same time. ... The answer that Craig [Mundie]'s team has devised to the concurrency problem is something called the concurrency and coordination runtime (CCR). ... These technologies are a key part of Microsoft Robotics Studio, a new software development kit built by Tandy [Trower]'s team. Microsoft Robotics Studio also includes tools that make it easier to create robotic applications using a wide range of programming languages." Reasoning about Computational Resource Allocation: An introduction to anytime algorithms. By Joshua Grass. (1996). ACM Crossroads. "Anytime algorithms are important in Artificial Intelligence (AI) research for two reasons. The first reason is that although many AI algorithms can take a long time to generate complete results, they often can generate very good partial results in a much shorter time period. Having the ability to reason about how much time is needed to receive a result that is adequate can make many AI systems more adaptive in complex and changing environments. The second reason is that a technique for reasoning about allocating time isn't limited to the internal workings of an agent, it may also be applied to other agents working cooperatively." Real-time Active Inference and Learning (RAIL). From IBM Research - Artificial Intelligence. "The objective of this project is the development of efficient techniques for real-time inference (diagnosis and prognosis) and learning (model adaptation to system changes) in complex distributed systems." Real-Time FAQs. From the comp.realtime newsgroup.
Real-Time Goal-Orientated Behaviour for Computer Game Agents. By Nick Hawes of The Cognition and Affect Project at The University of Birmingham School of Computer Science Cognitive Science Research Centre. September 2000. "This paper discusses the CogAff architecture as the basis for an agent that can display goal orientated behaviour under real-time constraints. To aid performance in real-time domains (e.g. computer games) it is proposed that both the processes encapsulated by the architecture, and the information it must operate on should be structured in a way that encourages a fast yet flexible response from the agent. In addition, anytime algorithms are discussed as a method for planning in real-time." An Anytime Algorithm for Decision Making under Uncertainty. By Michael C. Horsch and David Poole. In Proc. 14th Conference on Uncertainty in Artificial Intelligence (UAI-98), Madison, Wisconsin, USA, July 1998, pages 246-255. "We present an anytime algorithm which computes policies for decision problems represented as multi-stage influence diagrams. Our algorithm constructs policies incrementally, starting from a policy which makes no use of the available information. The incremental process constructs policies which includes more of the information available to the decision maker at each step. While the process converges to the optimal policy, our approach is designed for situations in which computing the optimal policy is infeasible." - from the abstract. The Real-Time Research Respository. Maintained by the IEEE Computer Society's Technical Committee on Real-Time Systems (IEEE-CS TC-RTS). Links to research groups, courses, tutorials, journals, commercial products, and much more. Agent-Centered Search (Real-Time Heuristic Search). A research project conducted by Sven Koenig, Associate ProfessorComputer Science Department,University of Southern California. "Autonomous agents often need to operate in real-time. For example, computer-controlled agents in combat games need to move smoothly. Agent-centered search methods interleave planning and plan execution and often decrease the sum of planning and plan-execution time because gathering information early reduces the subsequent amount of planning needed. They have a solid theoretical foundation, are able to use heuristic knowledge to focus their search, can commit to actions in any given amount of time (as required for real-time animation or driving), are able to handle uncertainty (such as actuator, sensor, and domain uncertainty) and can be used by single autonomous agents as well as teams of agents with various degrees of communication. They only give advice about which actions to execute and fail gracefully if their advice is ignored from time to time. This makes it easy to integrate them into complete agent architectures."
Real-Time A.I. at the Parallel Understanding Systems Group, Computer Science Dept., University of Maryland at College Park. "Our research to date has focused on a new approach, the Cooperative Intelligent Real-time Control Architecture (CIRCA). In this architecture, an AI subsystem reasons about task-level problems that require its powerful but unpredictable reasoning methods, while a cooperating, parallel real-time subsystem uses its predictable performance characteristics to deal with control-level problems that require guaranteed response times."
|
