Intelligent conversational tutoring system - a customisable software for training and education using a natural-language user interface (like a human tutor!)

Kurzfassung

An Israeli SME offers innovative software to facilitate training and education using technology that generates a mixed-initiative dialog with users. The system is capable of explaining challenging concepts effectively to ensure high efficacy in cases that are currently addressed effectively only by human resources. This system is based on recent academic studies that have empirically proved exceptional efficacy. Business partners sought to realise commercial products based on this technology.

Details

An Israeli start-up offers innovative software to facilitate training and education by utilising algorithm development in the fields of Computational Linguistics and Artificial Intelligence. The technology generates a mixed-initiative dialogue with users, "understands" them and explains a subject matter in a free-form conversation just as a human tutor would. This system is capable of explaining challenging concepts effectively to ensure high efficacy in cases that are currently addressed effectively only by human resources. This system is based on recent academic studies that have empirically proved exceptional efficacy.

The company offers to facilitate training, learning and knowledge transfer objectives using human tutoring emulation. The technology emulates human tutoring strategies and uses natural language algorithms to converse with a user in a mixed-initiative dialogue in which both parties, the user and the tutor, address a knowledge domain. The tutor "understands" the user and offers intelligent, personalised path to achieve the learning objective.

The software tutor emulator is designed to provide feedback to the user on what the user types in, positive, neutral, negative feedback, pumps the user for more information, prompts the user to fill in missing words, gives hints, fills in missing information with assertions, identifies and corrects erroneous ideas, answers the user's questions, and summarises answers.

The technological strength is accomplished by realising two basic principles: using the most intuitive interface - natural language conversation; and exercising proficient human-tutor strategies to explain concepts and ensure comprehension. The software tutor can assist and in some cases replace human resources on declarative knowledge domains, and initiate an intelligent dialogue that directs the learner to achieve predetermined learning goals. It challenges two major barriers of computer-based training - (1) cope with concepts that require an extensive level of user comprehension of learning objectives and (2) ease the user's cultural resistance to training with computer by using a natural-language dialogue interface.

The software tutor can facilitate knowledge acquisition objectives for training, K-12 education, customer service and enterprise knowledge management. It is designed cost-effectively in order to save direct measurable costs and to provide a sound positive ROI (Return On Investment) and provide unique solutions in an engaging environment.

Innovative Aspects:
The infrastructure is based on algorithms that were developed to imitate what human tutors do when students are tutored. The tutor must set the agenda and introduce most of the topics, questions and problems. However it is noticed that a human tutor does not use most of the ideal tutoring strategies. Using this software tutor, some of the computationally feasible tutoring strategies are implemented:
- Executes intelligent algorithms and evaluates user input by analysing speech acts and by analysing content.
- Tutoring moves are modelled to imitate a human tutor dialog moves. For example, a less specific question is used as a hint when the student does not know the answer.
The technological competitive advantage over existing educational technologies is based on the capability to initiate a natural-language dialogue to exercise cost-effectively intelligent tutoring strategies.
A natural-language one-on-one intelligent conversation increases productivity, engagement and knowledge retention, and in consequence, increases the economic benefits of a learning service/product. Furthermore, it increases the evaluated value for the user/customer by handling challenging learning objectives on which other technologies are not as efficient, and in cases a human tutor is not accessible.