Author: SRI International
-
A Study in Machine Learning from Imbalanced Data for Sentence Boundary Detection in Speech
We have constructed a hidden Markov model (HMM) system to detect sentence boundaries that uses both prosodic and textual information. Since there are more non-sentence boundaries than sentence boundaries in the data, the prosody model, which is implemented as a decision tree classifier, must be constructed to effectively learn from the imbalanced data distribution.
-
Posture-based data protection
We introduce Posture-Based Data Protection (PBDP), which encrypts data using keys available to a device only when it has been verified to be in a known good state, and has not subsequently performed any actions which place it at risk.
-
Investigation on Mandarin Broadcast News Speech Recognition
This paper describes our efforts in building a competitive Mandarin broadcast news speech recognizer. We present two novel algorithms in smoothing pitch features and segmenting Chinese characters into word units.
-
Globe Year 10 Evaluation: Into The Next Generation
This is the 10th in a series of annual evaluation reports that SRI has submitted to the GLOBE Program since the Program’s inception in 1995.
-
Speaker Clustered Regression-Class Trees for MLLR Adaptation
A speaker clustering algorithm is presented that is based on an eigenspace representation of Maximum Likelihood Linear Regression (MLLR) transformations and is used for training cluster-dependent regression-class trees for MLLR adaptation.
-
Improved Speech Activity Detection Using Cross-Channel Features for Recognition of Multiparty Meetings
We describe the development of a speech activity detection system using an HMM-based segmenter for automatic speech recognition on individual headset microphones in multispeaker meetings. We look at cross-channel features (energy and correlation based) to incorporate into the segmenter for the purpose of addressing errors related to cross-channel phenomena such as crosstalk.
-
Enriching Speech Recognition with Automatic Detection of Sentence Boundaries and Disfluencies
This paper describes a metadata detection system that combines information from different types of textual knowledge sources with information from a prosodic classifier. We investigate maximum entropy and conditional random field models, in addition to the predominant HMM approach, and find that discriminative models generally provide benefit over generative models.
-
Mystery Powders: An Application Of The Padi Design System Using The Four-Process Delivery System (Padi Technical Report 15)
This report illustrates how the general principles and structural components of the PADI framework were applied to a Mystery Powders assessment demonstration project that was computer-based. The body of this report addresses the design and technical issues that arose in the implementation of a computer-based interactive assessment, carried out with the PADI design system and…
-
Machine Translation Research at SRI
An outline of the many machine translation research programs at SRI in 2006.
-
Implications Of Evidence-Centered Design For Educational Testing (Technical Report 17)
-
QASR: Question Answering Using Semantic Roles for Speech Interface
In this paper, we evaluate a semantic role labeling approach to the extraction of answers in the open domain question answering task. We show that this technique especially improves the system performance when answers are communicated to the user by voice.
-
Within-Class Covariance Normalization for SVM-based Speaker Recognition
This paper extends the within-class covariance normalization (WCCN) technique described for training generalized linear kernels.