http://popups.lib.uliege.be/1373-5411/index.php?id=2298 Publications of Auteurs Yuichi Goto fr 0 http://popups.lib.uliege.be/1373-5411/index.php?id=4379 An anticipatory reasoning-reacting system anticipates based on anticipatory reasoning, which can draw new, previously unknown and/or unrecognized conclusions about some future event or events whose occurrence and truth are uncertain at the point of time when the reasoning is being performed. To perform anticipatory reasoning, we need to express the real world, predictive laws and behavioural patterns of the target domain as empirical theories represented by logical formulas which called world model, predictive model, and behavioural model correspondingly. However, there is no case to show what these models are and how to construct these models. To this end, this paper proposes a general procedure to construct these models, and presents a case study of runway incursion prevention. Besides, this paper also discusses the evaluation of the models. Thu, 10 Oct 2024 10:08:32 +0200 Thu, 10 Oct 2024 10:08:52 +0200 http://popups.lib.uliege.be/1373-5411/index.php?id=4379 http://popups.lib.uliege.be/1373-5411/index.php?id=4369 Anticipatory reasoning-reacting systems (ARRSs) were proposed as a new generation of reactive systems. Prediction and decision-making are important facilities of ARRSs. For the facilities, a prediction method and a decision-making method with forward reasoning based on strong relevant logic systems are proposed. On the other hand, practical reactive systems generally get sensory data and own internal status as character strings, but not as logical formulas. To implement facilities of prediction and decision-making based on the proposed methods, a transformation mechanism between observed data and logical formulas is demanded, but such a mechanism has not been proposed until now. This paper presents a transformation mechanism observed data and logical formulas for ARRSs. The mechanism can be applied to any computing anticipatory systems with logic-based reasoning. Thu, 10 Oct 2024 10:06:22 +0200 Thu, 10 Oct 2024 10:06:42 +0200 http://popups.lib.uliege.be/1373-5411/index.php?id=4369 http://popups.lib.uliege.be/1373-5411/index.php?id=3035 Anticipatory reasoning-reacting systems were proposed as a new generation of reactive systems with high reliability and high security. At present, there is no engineering environment to support development and maintenance of anticipatory reasoning-reacting systems. This paper presents a development and maintenance environment we are building for developing and maintaining anticipatory reasoning-reacting systems. By using our development and maintenance environment, developers and maintainers can develop and maintain various practical anticipatory reasoning-reacting systems more easily. Fri, 06 Sep 2024 16:03:29 +0200 Fri, 06 Sep 2024 16:03:39 +0200 http://popups.lib.uliege.be/1373-5411/index.php?id=3035 http://popups.lib.uliege.be/1373-5411/index.php?id=3027 Anticipatory reasoning engine is an indispensable component for computing anticipatory systems. In practical computing anticipatory systems, the efficiency of anticipatory reasoning engine is a key issue to satisfy the real-time requirements from applications. FreeEnCal: a forward reasoning engine for general-purpose purpose is a hopeful candidate for anticipatory reasoning engines. However, the current FreeEnCaJ is not efficient enough for practical computing anticipatory systems. This paper presents a new implementation of FreeEnCal improved by adopting fast algorithms, and shows its efficiency by comparing the improved FreeEnCal with the old one. The improved FreeEnCal can be used as an anticipatory reasoning engine of practical computing anticipatory systems. Fri, 06 Sep 2024 16:02:41 +0200 Fri, 06 Sep 2024 16:02:47 +0200 http://popups.lib.uliege.be/1373-5411/index.php?id=3027 http://popups.lib.uliege.be/1373-5411/index.php?id=2899 Persistent computing systems are an infrastructure of computing anticipatory systems. The reactability of a persistent computing system, which is how many reactions of the system are active at a certain time, is the most important property to characterize the system. On the other hand, to be anticipatory, the reactability of a computing anticipatory system must be continuous. This paper proposes the first method to measure the continuous reactability of a persistent computing system in a unified way. The continuous reactability of a persistent computing system is a new concept of computing systems, so that it will be raise new research problems of computing anticipatory systems as well as persistent computing systems. Tue, 03 Sep 2024 15:45:27 +0200 Tue, 03 Sep 2024 15:45:36 +0200 http://popups.lib.uliege.be/1373-5411/index.php?id=2899 http://popups.lib.uliege.be/1373-5411/index.php?id=2860 A computing anticipatory system must have the ability to make decision about its next action. To design and develop various computing anticipatory systems effectively, it is desirable to find a general methodology for decision making in computing anticipatory systems. This paper presents a deontic relevant logic approach to reasoning about actions in computing anticipatory systems. The paper discusses why the deontic relevant logic should be adopted as the fundamental logic to underlie reasoning about actions, presents a forward reasoning engine for reasoning about actions, and shows results of a case study to perform automated reasoning about actions based on deontic relevant logic. Tue, 03 Sep 2024 15:24:51 +0200 Tue, 03 Sep 2024 15:24:59 +0200 http://popups.lib.uliege.be/1373-5411/index.php?id=2860 http://popups.lib.uliege.be/1373-5411/index.php?id=2506 A practical anticipatory system with requirements of high reliability and high security must be able to perform any anticipatory reasoning to get enough effective conclusions anticipatorily within an acceptable time in order to satisfy the requirements from applications. This is a contradictory requirement since the execution time of anticipatory reasoning gets longer in proportion to the amount increasement of deduced conclusions. We are developing a forward deduction system for general-purposè entailment calculus, named EnCal. Although EnCal is a forward deduction engine for general-purpose entailment calculus, we expect that it can serve as the forward deduction engine in an anticipatory system to perform anticipatory reasoning based on temporal relevant logics. The key issue to achieve this goal is the efficiency of EnCal. This paper presents results and their implications of our experiences on improving the efficiency of EnCal by parallel processing techniques. Thu, 29 Aug 2024 11:00:11 +0200 Tue, 08 Oct 2024 13:49:58 +0200 http://popups.lib.uliege.be/1373-5411/index.php?id=2506 http://popups.lib.uliege.be/1373-5411/index.php?id=2292 An anticipatory reasoning-reacting system (ARRS) has been proposed as a highly reliable and highly secure reactive system. The most important component of an ARRS is its anticipatory reasoning engine (ARE). We have proposed temporal relevant logic (TRL) as a sound logical basis of anticipatory reasoning, and shown that parallel processing techniques are effective to efficient anticipatory reasoning. This paper presents a real ARE we are developing based on TRLs. We define basic requirements of an ARE, discuss implementation issues for an ARE, present our implementation techniques, and show and discuss some current experimental results obtained by using our ARE. Our ARE can also be used in other computing anticipatory systems where anticipatory reasoning plays a key role. Wed, 31 Jul 2024 12:53:25 +0200 Wed, 31 Jul 2024 12:53:33 +0200 http://popups.lib.uliege.be/1373-5411/index.php?id=2292