Efficient Cross-Domain Learning of Complex Skills

Lecture Notes in Computer Science

et al. 2013

Self Cite

Abstract. Systems for smart authoring of automated tutors, like SimStudent, have been mostly applied in well-defined problem-solving domains where little real-world background knowledge is needed, like math. Here we explore the generality of these methods by considering a very different task, article selection in English, where little problem-solving is done, but where complex prior perceptual skills and large amounts of background knowledge are needed. This background knowledge includes the ability to parse text and the extensive understanding of semantics of English words and phrases. We show that good performance can be obtained by coupling SimStudent with appropriate broad-coverage linguistic tools. Performance can be improved further on this task by extending one of the learning mechanisms used by SimStudent so that it will accept less-accurate production rule conditions, and prioritize learned production rules by accuracy. Experimental results show that the extended SimStudent successfully learns the tutored article selection grammar rules, and can be used to discover a student model that predicts human student behavior as well as the human-generated model.

Section: A Brief Review Of Simstudentmentioning

confidence: 99%

“…It has been demonstrated in multiple domains such as fraction addition, equation solving, and stoichiometry [14]. Additionally, it has been shown that by integrating perceptual learning into skill learning, SimStudent can be used to find better student models than humangenerated models [3].…”

Section: Introductionmentioning

confidence: 99%

Integrating Perceptual Learning with External World Knowledge in a Simulated Student

Tian

Lecture Notes in Computer Science

et al. 2013

Self Cite

“…In previous work, we have developed a one-dimensional (1-D) pCFG learner to acquire representations of 1-D strings (e.g., the parse structure of -3x), and showed that the acquired representations yield effective learning, while reducing the amount of knowledge engineering required in building an intelligent agent [5]. Moreover, it has been shown that with this extension, the intelligent agent becomes a better model of human students [6], and can be used to better understand human student learning behavior [7].…”

Section: Related Workmentioning

confidence: 99%

“…where T i is the parse tree with root S for record R i given the current grammar G, and parse(T i ) denotes the parse structure of T i ignoring the symbols associated with the parse nodes 5 Since any subtree of a most probable parse tree is also a most probable parse subtree, we have…”

Section: Viterbi Trainingmentioning

confidence: 99%

See 1 more Smart Citation

Learning to Perceive Two-Dimensional Displays Using Probabilistic Grammars

Machine Learning and Knowledge Discovery in Databases

Koedinger

2012

Self Cite

Abstract. People learn to read and understand various displays (e.g., tables on webpages and software user interfaces) every day. How do humans learn to process such displays? Can computers be efficiently taught to understand and use such displays? In this paper, we use statistical learning to model how humans learn to perceive visual displays. We extend an existing probabilistic context-free grammar learner to support learning within a two-dimensional space by incorporating spatial and temporal information. Experimental results in both synthetic domains and real world domains show that the proposed learning algorithm is effective in acquiring user interface layout. Furthermore, we evaluate the effectiveness of the proposed algorithm within an intelligent tutoring agent, SimStudent, by integrating the learned display representation into the agent. Experimental results in learning complex problem solving skills in three domains show that the learned display representation is as good as one created by a human expert, in that skill learning using the learned representation is as effective as using a manually created representation.

Problem Order Implications for Learning

Int J Artif Intell Educ

Koedinger

2013

Self Cite

The order of problems presented to students is an important variable that affects learning effectiveness. Previous studies have shown that solving problems in a blocked order, in which all problems of one type are completed before the student is switched to the next problem type, results in less effective performance than does solving the problems in an interleaved order. However, we have no precise understanding of the reason for this effect. In addition to existing theoretical results, we use a machine-learning agent that learns cognitive skills from examples and problem solving experience, SimStudent, to provide a computational model of the problem order question. We conduct a controlled simulation study in three different math and science domains (i.e., fraction addition, equation solving and stoichiometry), where SimStudent is tutored by automatic tutors given problems that have been used to teach human students. We compare two problem orders: the blocked problem order, and the interleaved problem order. The results show that the interleaved problem order yields as effective or more effective learning in all three domains, because the interleaved problem order provides more or better opportunities for error detection and correction to the learning agent. Examination of the agent's performance shows N. Li ( )