In this paper we deal with machine learning methods and algorithms applied in learning simple concepts by their refining or explication. The method of refining a simple concept of an object O consists in discovering a molecular concept that defines the same or a very similar object to the object O. Typically, such a molecular concept is a professional definition of the object, for instance a biological definition according to taxonomy, or legal definition of roles, acts, etc. Our background theory is Transparent Intensional Logic (TIL). In TIL concepts are explicated as abstract procedures encoded by natural language terms. These procedures are defined as six kinds of TIL constructions. First, we briefly introduce the method of learning with a supervisor that is applied in our case. Then we describe the algorithm 'Framework' together with heuristic methods applied by it. The heuristics is based on a plausible supply of positive and negative (near-miss) examples by which learner's hypotheses are refined and adjusted. Given a positive example, the learner refines the hypothesis learnt so far, while a near-miss example triggers specialization. Our heuristic methods deal with the way refinement is applied, which includes also its special cases generalization and specialization.

In this paper, we deal with the support in the search for appropriate textual sources. Users ask for an atomic concept that is explicated using machine learning methods applied to different textual sources. Next, we deal with the so-obtained explications to provide even more useful information. To this end, we apply the method of computing association rules. The method is one of the data-mining methods used for information retrieval. Our background theory is the system of Transparent Intensional Logic (TIL); all the concepts are formalised as TIL constructions.

This paper deals with an optimization of methods for recommending relevant text sources. We summarize methods that are based on a theory of Association Rules and Formal Conceptual Analysis which are computationally demanding. Therefore we are applying the ‘Iceberg Concepts’, which significantly prune output data space and thus accelerate the whole process of the calculation. Association Rules and the Relevant Ordering, which is an FCA-based method, are applied on data obtained from explications of an atomic concept. Explications are procured from natural language sentences formalized into TIL constructions and processed by a machine learning algorithm. TIL constructions are utilized only as a specification language and they are described in numerous publications, so we do not deal with TIL in this paper.

Navigation and an agent’s map representation in a multi-agent system become problematic when agents are situated in complex environments such as the real world. Challenging modifiability of maps, long updating period, resource-demanding data collection makes it difficult for agents to keep pace with rather quickly expanding cities. This study presents the first steps to a possible solution by exploiting natural language processing and symbolic methods of supervised machine learning. An adjusted algorithm processes formalized descriptions of one’s journey to produce a description of the journey. The explication is represented employing Transparent Intensional Logic. A combination of several explications might be used as a representation of spatial data, which may help the agents to navigate. Results of the study showed that it is possible to obtain a topological representation of a map using natural language descriptions. Collecting spatial data from spoken language may accelerate updating and creation of maps, which would result in up-to-date information for the agents obtained at a rather low cost.

No abstract