作者: Alain Mille Marty , None
DOI:
关键词: Computer science 、 Artificial intelligence 、 Parsing 、 Datalog 、 Process (engineering) 、 Human–computer interaction 、 Representation (mathematics) 、 TRACE (psycholinguistics) 、 Interpretation (logic) 、 Knowledge base 、 Semantics
摘要: This paper presents a general way to represent and manage computer based activity traces, which allow human computational agents parse, process assist users in their complex dynamic activities. To succeed, such system has provide mechanisms declare, evolve improve the sufficient knowledge observe, interpret reason about its observations of this environment. We propose unambiguous language capable expressing knowledge, together with some precise well understood manipulating sets draw inferences, answer queries, make interpretations update both base desired behaviour. consider traces user interaction as system's activities experiences. As knowledge-based (KBS) exploits an explicit representation different kinds we describe Trace-Based System (TBS) kind KBS whose main source is set trace subsuming user-system interactions evolving his/her The architecture TBS framework. formal concept modelled associated languages supporting reasoning them interpretation. Firstly, focus on formalization model M-Trace offline exploitation. Then, specify M-Traces patterns, transformations. extend formalisation online Finally, implementation datalog discuss our proposition future works.