Study on the Integration of Speech Recognition with an Operation System:

October 1 Report

By Daniel Wilson

As I indicated in my last report, I have moved on to the Natural Language stage of the project. My efforts during the last two weeks have been directed toward gaining an understanding of NLP theory, as well as of Prolog.

I am making extensive use of Dr. Ray C. Dougherty’s book, Natural Language Computing: An English generative Grammar in Prolog. He provides dozens of sample programs with the book that illustrate the concepts he describes. Dr. Dougherty’s description of Prolog programming is fascinating. He discusses data in terms of facts and relations, much like a relational database. He explains that we are simulating intelligence, and emphasizes three aspects that intelligence must include. Unboundedness, rule-governed creativity, and judgement are all keys to good natural language processing. Dougherty believes that these aspects may be modeled in Prolog, and begins explaining Prolog programming.

As such, I have also worked on getting a Prolog interpreter to run on my Linux system. I had a little difficulty since most of the interpreters and compilers are available only via FTP access, and we do not have that on campus. I was finally able to download and install the SWI-Prolog 3.1 interpreter/compiler available from the RPM Repository at http://rufus.w3.org/linux/RPM/. It is running the simplest programs from Dr. Dougherty’s book, and I am experimenting with it to find out how to build an executable with it. If I can’t get SWI-Prolog to compile, I will try to get the GNU Prolog compiler.

The choice of a compiler should not seriously constrain what or how I write, so I intend to begin designing the NLP module Monday. By that time I should understand enough of Dougherty’s concepts to begin design. If I begin design Monday, however, finishing the unit by the 8th looks unlikely. I believe a one-week schedule slip will give me enough time, however to complete the unit and do a good job. My original schedule had three weeks built in at the end for writing a report and to accommodate a schedule slip if necessary. As the tables below indicate, this delay will not put the overall project behind.

September 3:

Project Starts

September 17:

SR unit written

October 8:

NLP unit written and integrated with SR

October 22:

Semantic analysis unit written and integrated

November 12:

OS integration

December 3:

Final deliverable

Table 1: Original Schedule

 

 

September 3:

Project Starts

September 17:

SR capabilities researched

October 15:

NLP unit written and tested

October 29:

Semantic analysis unit written and integrated

November 19:

OS integration

December 3:

Final deliverable

Table 2: New Schedule

1