Sourish Chaudhuri

Audio event detection from acoustic unit occurrence patterns

In most real-world audio recordings, we encounter several types of audio events. In this paper, we develop a technique for detecting signature audio events, that is based on identifying patterns of occurrences of automatically learned atomic units of sound, which we call Acoustic Unit Descriptors or AUDs. Experiments show that the methodology works as well for detection of individual events and their boundaries in complex recordings.

It's Not Easy Being Green: Supporting Collaborative Green Design Learning

We present the results of a study in which we contrast alternative forms of collaborative learning support in the midst of a collaborative design task in which students negotiate between increasing power and increasing environmental friendliness. In this context, we evaluated the instructional effectiveness of four alternative support conditions as well as a goal manipulation. Both manipulations yield surprising findings, which we are continuing to investigate.

Unsupervised hierarchical structure induction for deeper semantic analysis of audio

Current audio analysis techniques rely on fairly shallow analysis of audio content, using symbols or patterns extracted directly from the observed acoustics. We hypothesize that the observed acoustics actually map to semantics in a hierarchical manner, and that the higher levels of this hierarchy correspond to increasingly higher-level semantics. In this paper, we present a model for deeper analysis of the observed acoustics, that induces a probabilistic tree structure depending on estimated constituent identities and contexts. Audio characterization using the deeper structure outperforms the standard shallow-feature based characterizations.

Helping Agents in VMT

In this chapter we describe ongoing work towards enabling dynamic support for collaborative learning in the Virtual Math Teams (VMT) environment using state-of-the-art language technologies such as text classification and dialogue agents. The key research goal of our long-term partnership is to experimentally learn broadly applicable principles for supporting effective collaborative problem solving by using these technologies to elicit behavior such as reflection, help seeking, and help provision, which are productive for student learning in diverse groups. Our work so far has yielded an integrated system that makes technology for dynamic collaborative learning support—which has proved effective in earlier lab and classroom studies—available for experimental use within the “wild” VMT environment.

Motivation and collaborative behavior: an exploratory analysis

The motivating effects of collaborative learning have long been argued, however a careful analysis of the relationship between the motivation orientation of a student and perceptions of himself, his partners, his collaborative behaviors, and learning in a collaborative context have not been as thoroughly explored. In this paper we present an exploratory analysis of data from a collaborative learning study from the standpoint of motivation type of students and their partners. Overall, what we see is that a students own motivation orientation may color their perception of the exchange of help in the collaboration, sometimes obscuring the reality of the help actually exchanged.

Learning contextual relevance of audio segments using discriminative models over AUD sequences

Effective retrieval of multimodal data involves performing accurate segmentation and analysis of such data. With easy access to a number of audio and video sharing platforms online, user-generated content with considerably less than ideal recording conditions has increased rapidly. One major issue with such content is the presence of semantically irrelevant segments in such recordings. This leads to the presence of considerable contextual noise in such recordings that makes analysis difficult. In this paper, we present a discriminative large-margin based approach that uses annotated data to understand which parts of the audio are relevant (while noting that the notion of relevance could be extremely subjective and potentially challenging to define), and can automatically extract such segments from new audio.

SIDE: The Summarization Integrated Development Environment

In this type-II demo, we introduce SIDE (the Summarization Integrated Development Environment), an infrastructure that facilitates construction of summaries tailored to the needs of the user. It aims to address the issue that there is no such thing as the perfect summary for all purposes. Rather, the quality of a summary is subjective, task dependent, and possibly specific to a user. The SIDE framework allows users flexibility in determining what they find more useful in a summary, both in terms of structure and content. As an educational tool, it has been successfully user tested by a class of 21 students in a graduate course on Summarization and Personal Information Management.

Leveraging Structural Relations for Fluent Compressions at Multiple Compression Rates

Prior approaches to sentence compression have taken low level syntactic constraints into account in order to maintain grammaticality. We propose and successfully evaluate a more comprehensive, generalizable feature set that takes syntactic and structural relationships into account in order to sustain variable compression rates while making compressed sentences more coherent, grammatical and readable.

Evaluating the Syntactic Transformations in Gold Standard Corpora for Statistical Sentence Compression

We present a policy-based error analysis approach that demonstrates a limitation to the current commonly adopted paradigm for sentence compression. We demonstrate that these limitations arise from the strong assumption of locality of the decision making process in the search for an acceptable derivation in this paradigm.

Supporting the Guide on the SIDE

We present SIDE (the Summarization Integrated Development Environment), which is an infrastructure that facilitates the construction of reporting interfaces that support group learning facilitators in the task of getting a quick sense of the quality and effectiveness of a collaborative learning interaction. The SIDE framework offers flexibility in the specification of which conversational behavior to take note of as well as how noted behavior should be reported to instructors, making it a valuable research tool.