Help Generation in a System for Learning Natural Language to First Order Logic Conversion

Isidoros Perikos, Foteini Grivokostopoulou, and Ioannis Hatzilygeroudis

Keywords

Automatic help generation, Personalized feedback, User error detection, Web based e-learning

Abstract

The NLtoFOL system is an interactive web-based system for learning to convert natural language (NL) sentences into first order logic (FOL) sentences, also called formalization. In this paper, we present the mechanism that helps students in learning the above conversion. It operates in two stages. In the first stage, it is used to recognize the errors that the user does. Error detection is based on an error categorization scheme. In the second stage, it is used for generating help messages to the students, based on the detected error(s). Help comes in two types, step-specific and answer-specific. Step-specific help consists in static hint messages, whereas answer-specific in dynamic feedback messages. Dynamic messages are created by filling in predefined message patterns. The choice of the help type is based on student's error history. The system uses a mix of immediate and on-demand feedback to the students. Evaluation showed promising results.

Important Links:

Go Back