Welcome to Klaus Zechner's homepage!

I am a research scientist in speech and NLP in the Automated Scoring and Natural Language Processing at Educational Testing Service in Princeton, NJ.

Since my arrival in 2002 I have been working mainly on the exciting and ambitious project of automatic scoring of non-native speech. The basic idea behind our approach is to first generate word hypotheses by means of an ASR system, then to compute a wide range of features describing speech proficiency and communicative competence, and finally to build a scoring model which generates a numeric proficiency score. As of now, no automatic speech scoring is planned in high-stakes test environments such as TOEFL®iBT.

My education includes four MS and MA degrees in computer science, linguistics, cognitive science and computational linguistics, as well as a Ph.D. in Language and Information Technologies at Carnegie Mellon University in Pittsburgh, PA.

Next to my current research in speech scoring, my other main research interests include:

• Summarization of Text and Speech
• Question Answering
• Information Extraction, Information Retrieval
• Shallow Analysis of Spoken and Written Language
• Text Normalization of Speech
• Machine Translation

The following is a list of my main projects in my scientific career:

• Automatic Scoring of Non-Native Speech (SpeechRater®)
• Rapid Deployment Statistical Machine Translation (RADD-MT)
• Multi-Modal Meeting Browser with Speech Summarization (Genoa)
• Automatic Discourse Annotation of Spontaneous Spanish (Clarity)
• Speech-to-Speech Machine Translation (C-Star, Verbmobil)
• Computer-Assisted Aphasia Rehabilitation (Auswege)

For a list of (most of) my publications please click the following link:

Klaus Zechner’s publications

Please contact me by email using kzechner as username before the at-sign and ets.org as domain name after it.

(I’m just trying to make it harder for automatic email harvesters :)J)

[---last update: November 18, 2007---]