"The key for progress is integration"

Research Interests

The goal of the research is to extend the PSI theory that will allow us integrate this approach with other areas of AI, such as vision and speech. My personal interests are in building a smart room that is able to know who enters the room, talk to them and understand the conversations that are ongoing in this room. Today, technology is far away from realizing such a smart room. My work will focus on identifying the exact boundaries of what we can do today. I want to integrate the best results from various efforts of the last decades in speech, vision, pattern recognition, cognitive agents and robotics as well as text retrieval, translation and summarization. Most of the work in these diverse fields have reached a level of maturity that they are ready to be integrated and pushed on to the next level. Only complex scenarios will emphasize which areas need real improvement. Perhaps it is exactly the interplay of all these related AI fields that will help us to identify our next research topics!

• If anyone would like to join in this effort, we are looking for collaboration partners in any of these topics. Currently, this lab is entirely without funding. Therefore, if you want to work with us and would like to visit, we can offer free housing and a lot of sunshine.


• We are aware that industry is very advanced in most of these areas. I will be happy to hear from industry who would be interested in remote advising of a student.

Activities

• Co-chair of Odyssey 2006
• Co-PI on project Window to the Caribbean
• Teaching Software Engineering (undergraduate, graduate, MBA)
• Teaching Software Engineering II (graduate CS only)
• Teaching Software Architecture (undergraduate, graduate, MBA)
• Teaching Software Testing (graduate)
• Teaching Speech Recognition for Master in Electrical Engineering
• Teaching Data Mining for Master in Computer Science
• Teaching Pattern Recognition for Master in Electrical Engineering
• Life-info at Pattern Recognition with Dr. Pedro A. Torres-Carrasquillo at Lincoln Laboratory
• Teaching C++

Other Interests

I am very interested to find out why most of our students that have a GPA higher than 3.5 in their core classes in Junior and Senior year graduate with an overall GPA of less than 3.0!!!!
As you know, most internships and scholarships have a cut off line at 3.0 GPA to be eligible. As a result, most of our students remain ineligible for important programs that entice students to go to graduate school or choose an academic career.
In terms of number of Hispanic Engineering students graduated, PUPR is one of the top two Engineering Schools in the country. We also have 25% female Engineering students, a % that is unheard of in any other school outside of the island.
We as a country are not addressing one of the root causes for the small number of minorities in academia, by taking care of the Seniors and Juniors that fit this profile.
I have spent some time designing a program for summer self study. This would allow students to study their "bad" subjects during the course of a summer on their own and take an exam (that could be standardized across several institutions). The new grade would figure in the final GPA - in the same way that any repeated course would affect the GPA in their home University.
Currently, there are efforts on the way to find funding for professors who will support these students. For this purpose I have written an NSF Proposal on Intensive Self-Study Summer Program. This was rejected. One of the key findings was that the students' would not be able to study on their own without tutors. We know that this is not true for the students that fit the profile in question and that is why we are now working on resubmitting this idea to the Coca Cola Foundation.
If you are a professor (at PUPR or any other University) that is interested in this topic, please email me. This issue is not at all specific to our University. We have a proven functional retention program as do most Universities. I am not talking about programs that address Freshman and Sophomore students. These exist! I am talking about programs that address Juniors and Seniors.
If you are a student that fits this profile, you can get in touch with me to find out more about progress in this project. In the meantime, please sign up to our yahoo group that is designed especially for PUPR students and alumni to find out and post about internships and scholarships. There is a big need to create awareness of the importance of the GPA in your life now and in the future!

Selected Publications

Additional Publications can be found at Lincoln Laboratory and OGI

• Kay Berkling, SCoPE, Syllable Core and Periphery Evaluation for Foreign Accent Identification and Automatic Syllabification. (Accepted for publication in Speech Communications Journal, special edition on multilinguality, extended version of workshop paper, below)
• Kay Berkling and Marc Zissman, Approaches to Automatic Language Identification of Speech, MIST: Multi-lingual Interoperability in Speech Technology, September 1999, Leusden, The Netherlands (Speech Communications Journal)
• Kay Berkling, Chris Cleirigh, and Julie Vonwiller, SCoPE, Syllable Core and Periphery Evaluation for Foreign Accent Identification and Automatic Syllabification, MIST: Multi-lingual Interoperability in Speech Technology, September 1999, Leusden, The Netherlands
• Kay Berkling, Doug Reynolds and Marc Zissman, Evaluation of Confidence Measures for Language Identification, Eurospeech '99, September 1999, Budapest, Hungary.
• Kay Berkling, Chris Cleirigh, Julie Vonwiller and Marc Zissman, Improving Accent Identification Through knowledge of English Syllable Structure, International Conference on Spoken Language Processing (ICSLP), Sydney, Australia, November 1998.
• Kay Berkling (available at http://cslu.cse.ogi.edu/publications) Automatic Language Identification with Sequences of Language-independent Phonemes PhD Thesis, 1996, Oregon Graduate Institute of Science & Technology
• Kay Berkling and Etienne Barnard, Language Identification with Inaccurate String Matching International Conference on Spoken Language Processing (ICSLP), Philadelphia, USA, October 1996.
• Kay Berkling and Etienne Barnard, Theoretical Error Prediction for a Language Identification System using Optimal Phoneme Clustering, Proceedings of the Fourth European, Conference on Speech Communiciation and Technology, Madrid, Spain, September 1995.
• Kay Berkling and Etienne Barnard, Language Identification of Six Languages Based on a Common Set of Broad Phonemes, Proceedings of International Conference on Spoken Language Processing (ICSLP), Yokohama, Japan, October 1994
• Yonhong Yan, Kay Berkling, and Etienne Barnard, Bigram Models and Phoneme Clusters for Language Identification, Proceedings of Speech Research Symposium SRS, Johns Hopkins, Baltimore USA, June 1994
• Kay Berkling, Takayuki Arai, Etienne Barnard and Ronald Cole, Analysis of Phoneme-Based Features for Language Identification, Proceedings of the International Conference on Acoustics, Speech and Signal Processing, Adelaide, South Australia, April 1994
• Yeshwant Muthusamy, Kay Berkling, Takayuki Arai, Ronald Cole and Etienne Barnard, A comparison of Approaches to Automatic Language Identification Using Telephone Speech, Proceedings 3rd European Conference on Speech Communication and Technology (Eurospeech 93), Berlin, Germany, September 1993.
• Y. K. Muthusamy and T. Arai and K. M. Berkling and R. A. Cole and E. Barnard, Two Approaches to Automatic Language Identification with Telephone Speech,Proceedings of Speech Research Symposium SRS, Johns Hopkins, Baltimore USA, June 1993