ETS: Domain Adaptation and Stacking for Short Answer Scoring

Automatic scoring of short text responses to educational assessment items is a challenging task, particularly because large amounts of labeled data (i.e., human-scored responses) may or may not be available due to the variety of possible questions and topics. As such, it seems desirable to integrate various approaches, making use of model answers from experts (e.g., to give higher scores to responses that are similar), prescored student responses (e.g., to learn direct associations between particular phrases and scores), etc. Here, we describe a system that uses stacking (Wolpert, 1992) and domain adaptation (Daume III, 2007) to achieve this aim, allowing us to integrate item-specific n-gram features and more general text similarity measures (Heilman and Madnani, 2012). We report encouraging results from the Joint Student Response Analysis and 8th Recognizing Textual Entailment Challenge.