Over the course of the last few days, we interviewed Dan Bikel (http://research.google.com/pubs/DanielBikel.html), who is a Senior Research Scientist here at Google, about the book he co-edited with Imed Zitouni (IBM) entitled Multilingual Natural Language Processing Applications: From Theory to Practice.

Research at Google: Hi, Dan.  How are you?

Dan: I’m good, thanks.  And you?

R@G: I’m also doing well, but lately I’ve been feeling more like an organization than a person.

Dan: That’s understandable.

R@G: Thanks.  So tell me about this new book of yours.  How did it come about?

Dan: Sure. A little more than three years ago, Imed Zitouni and I were both Research Scientists at IBM.  At the time, we were building all sorts of Natural Language Processing (NLP) tools that work on many different languages.  We felt that while there were several excellent NLP textbooks out there, we wanted to produce one that dealt squarely with the issue of multilinguality, which is something that most NLP researchers and engineers need to worry about.

R@G: I see, you felt that an NLP book focusing on how to build multilingual NLP systems was missing, and you wanted to create such a book?

Dan: Yes, that’s right.  Hey, you picked up on that pretty quickly.

R@G: Thanks. People often tell me I’m a quick study.  I noticed that the word “applications” appears in the title of your book.  Is that significant?

Dan: It sure is. Imed and I also felt that there was a lot of real-world knowledge and experience in building robust, multilingual NLP applications that often didn’t make it into existing textbooks, conference proceedings and journal papers.  We wanted to create a book that didn’t shy away from such practical knowledge, but embraced it and highlighted it.  But you can’t talk about the practical without first mastering the theoretical, so the book is divided into two halves: the first half is entitled “In Theory” and the second, “In Practice”.

R@G: Sounds like a good idea. So how did you go about creating this book?

Dan: Creating a book that aims to be a textbook-like resource is no trivial thing.  Imed and I decided to write two of the chapters, and we collected an array of experts in various subdisciplines of NLP to write the other 14 chapters.  We are very grateful to our brilliant chapter authors! For example, we have an excellent and thorough review of multilingual NLP approaches to syntax written by Prof. Anoop Sarkar of Simon Fraser University, and our chapter on machine translation was written by the renowned Prof. Philipp Koehn of the University of Edinburgh.

R@G: What happened next?

Dan: Two years ago I started as a Senior Research Scientist at Google, where I’ve been working on multilingual NLP and speech processing technologies. Since that time, even though we’ve been at different industrial research labs, Imed and I have continued editing this book.  We are proud to say that it went to press last month and is now for sale!

R@G: Congratulations!

Dan: Thanks! I’m also greatly appreciative that Google in general, and Corinna Cortes in particular, have been incredibly supportive of this book project the entire time I’ve been a Googler. The fact that Google supports special interest projects for its Software Engineers and Research Scientists is a wonderful thing. 

R@G: I couldn’t agree more.  So, could you tell me a bit about some of the topics you cover in the book?

Dan: Certainly.  The first half, “In Theory”, opens with a trio of chapters describing the structure of language at various levels of granularity: words (morphology), documents (topic segmentation and sentence breaking) and sentences (syntax).  We also have chapters on semantic parsing, language modeling and two important areas called recognizing textual entailment and sentiment analysis.  The second half, “In Practice,” has chapters on various multilingual NLP application areas, such as information extraction and machine translation (which is an inherently multilingual NLP application area).

R@G: So who is the “target audience” for this new book of yours?

Dan: We aim to serve advanced graduate students studying NLP or related disciplines, as well as researchers and engineers in NLP in academia, industry and government.  Now is as exciting a time as ever to be in the field of NLP, and we hope this book allows us to interact and collaborate with as many researchers as we can!

Multilingual Natural Language Processing Applications: From Theory to Practice
Edited by Daniel M. Bikel and Imed Zitouni.
Published by IBM Press/Pearson.
Available now in bookstores and online:
2 Photos - View album
Shared publiclyView activity