Search

We present the EuroPat corpus of patent-specific parallel data for 6 official European languages paired with English: German, Spanish, French, Croatian, Norwegian, and Polish. The filtered parallel corpora range in size from 51 million sentences (Spanish-English) to 154k sentences (Croatian-English), with the unfiltered (raw) corpora being up to 2 times larger. Access to clean, high quality, parallel data in technical domains such as science, engineering, and medicine is needed for training neural machine translation systems for tasks like online dispute resolution and eProcurement. Our evaluation found that the addition of EuroPat data to a generic baseline improved the performance of machine translation systems on in-domain test data in German, Spanish, French, and Polish; and in translating patent data from Croatian to English. The corpus has been released under Creative Commons Zero, and is expected to be widely useful for training high-quality machine translation systems, and particularly for those targeting technical documents such as patents and contracts.

Across many different educational settings, course discussion forums allow students to learn from one another and connect socially with their peers and instructors. Content analysis of the messages that are exchanged has been used to model engagement using two well-established theoretical frameworks, Community of Inquiry and ICAP. However, manual content analysis is slow and expensive, and prior work on automation is limited. In addition, these two theoretical frameworks developed out of different disciplines, and little work has been done to bring them together. To address these issues, I will evaluate the use of advanced methods from natural language processing to automate the content analysis, considering both frameworks individually and together, and comparing the results with prior work in terms of accuracy and explanatory power. I will also contribute to the conceptual understanding of what characterises a high quality discussion forum contribution by identifying connections between the frameworks themselves and places where they offer complementary perspectives.

Shared decision making (SDM) is increasingly considered as the best way to reach a treatment decision in a clinical environment. However, the use of SDM in practice can be obstructed by a number of factors, such as time constraints or lack of applicability due to patient characteristics. Our project, PrepDoc, explores how a Virtual Training Doctor (VTD) can help patients overcome some of these obstacles to experiencing effective SDM during doctor's visits. In this paper, we report on user studies conducted with 19 participants in Scotland aged 65+. The goal of these studies was to identify the reactions of this audience to the PrepDoc system, evaluate its suitability within Scotland, and elicit suggestions to improve it. Our findings revealed that the idea of empowering people to participate in SDM using a virtual agent was positively received by all participants. However, the reactions to how this idea was implemented in the PrepDoc system varied greatly across participants. Based on this, our paper outlines recommendations for enhancing the user experience with VTDs, accommodating individual differences of older adults, and accounting for the national context.

We present an experiment aimed at improving interpretation robustness of a tutorial dialogue system that relies on detailed semantic interpretation and dynamic natural language feedback generation. We show that we can improve overall interpretation quality by combining the output of a semantic interpreterwith that of a statistical classifier trained on the subset of student utterances where semantic interpretation fails. This improves on a previous result which used a similar approach but trained the classifier on a substantially larger data set containing all student utterances. Finally, we discuss how the labels from the statistical classifier can be integrated effectively with the dialogue system’s existing error recovery policies.

We present BEETLE II, a tutorial dialogue system designed to accept unrestricted language input and support exper-imentation with different tutorial planning and dialogue strategies. Our first system evaluation used two different tutorial policies and demonstrated that the system can be successfully used to study the impact of different approaches to tutoring. In the future, the system can also be used to experiment with a variety of natural language interpretation and generation techniques.

Our research goal is to investigate whether previous findings and methods in the area of tutorial dialogue can be generalized across dialogue corpora that differ in domain (mechanics versus electricity in physics), modality (spoken versus typed), and tutor type (computer versus human). We first present methods for unifying our prior coding and analysis methods. We then show that many of our prior findings regarding student dialogue behaviors and learning not only generalize across corpora, but that our methodology yields additional new findings. Finally, we show that natural language processing can be used to automate some of these analyses.

We propose that a set of five dimensions forms a foundation underlying a number of prevalent theoretical perspectives on learning. We show how student contributions to instructional dialogue can be reliably annotated with these dimensions. Finally, we provide preliminary validation evidence for our coding scheme and illustrate the potential value of such an approach to analyzing student behavior in tutorial dialogue.

Our research goal is to investigate whether previous findings and methods in the area of tutorial dialogue can be generalized across dialogue corpora that differ in domain (mechanics versus electricity in physics), modality (spoken versus typed), and tutor type (computer versus human). We first present methods for unifying our prior coding and analysis methods. We then show that many of our prior findings regarding student dialogue behaviors and learning not only generalize across corpora, but that our methodology yields additional new findings. Finally, we show that natural language processing can be used to automate some of these analyses.

Understanding answers to open-ended explanation questions is important in intelligent tutoring systems. Existing systems use natural language techniques in essay analysis, but revert to scripted interaction with short-answer questions during remediation, making adapting dialogue to individual students difficult. We describe a corpus study that shows that there is a relationship between the types of faulty answers and the remediation strategies that tutors use; that human tutors respond differently to different kinds of correct answers; and that re-stating correct answers is associated with improved learning. We describe a design for a diagnoser based on this study that supports remediation in open-ended questions and provides an analysis of natural language answers that enables adaptive generation of tutorial feedback for both correct and faulty answers.

We describe two tutorial dialogue systems that adapt techniques from task-oriented dialogue systems to tutorial dialogue. Both systems employ the same reusable deep natural language understanding and generation components to interpret students ' written utterances and to automatically generate adaptive tutorial responses, with separate domain reasoners to provide the necessary knowledge about the correctness of student answers and hinting strategies. We focus on integrating the domain-independent language processing components with domain-specific reasoning and tutorial components in order to improve the dialogue interaction, and present a preliminary analysis of BeeDiff's evaluation.