Codetown ::: a software developer's community
Tags:
Commercial applications of computational linguistics have been growing by leaps and bounds. IBM created a new poster child for the field with their Jeopardy champion Watson, and most everyone has used Google Translate and/or Voice by now (and fewer yet will have escaped interacting with an Automated Voice Response system). Commercial segments making significant use of NLP include web marketing, medicine, biomedical research, finance, law, and customer call centers.
In the most general terms, computational linguistics is applying computational methods to problems in linguistics. Linguistics then is the study of human language in all its aspects. Although it hasn't received a lot of press until recently, computational linguistics has been around pretty much since the development of the computer. One of the first uses of digital computers (and a key impetus for their development) was in code breaking, which is an application of "compling" (I know, it looks like a typo for "compiling" but CL looks like Common LISP to me). The Association for Computational Linguistics (ACL), the largest and oldest scientific and profssional society in the field, will hold its 50th annual conference next July.
Being such a broad field there are of course many specializations and various communities with differing objectives and vocabularies. Folks primarily focused on engineering computer systems that process human language at a level deeper than simply character strings are generally under the Natural Language Processing (NLP) banner. The commercial success of CL applications and the focus on some problems other than the traditional NLP ones (translation, text understanding, and speech recognition and generation) has spawned Text Analytics as another subfield that is closely associated with Business Intelligence (itself largely commercial application of machine learning methods). Perhaps the biggest focus of Text Analytics has been on sentiment analysis, which assesses a speaker's attitude or mood in something they've said or written (we usually say "speaker" even when the medium is written or typed). There are many businesses that use sentiment analysis on the web to find out what folks are saying about them and their products, in call centers for quality control, and in finance to predict future prices. Applications in law and government include "e-discovery" and smart OCR systems. Lastly, and far from leastly, is compling in the medical field and the specialized domain knowledge it calls for which is known as bioinformatics. Bioinformatics may well be compling's "killer app" because of the tremendous opportunity to do good. Answering technical questions for medical practitioners is the application IBM has targeted as Watson's "day job".
This is a big topic and I have lots I would like to say, so I intend to drop in here often. So stay tuned!
~~~~ Jim White
On a similar note: This fall a colleague of mine informed me of a class being offered by two Standford professors. It is Introduction to Artifical Intelligence (ai-class.com). I signed up for the course along with 138,000 other students. It has been interesting to see some of the therories and formulas used to help the computer find the correct answer. Prior to starting the class, I did not realize it is mostly related to statistics. I just finished the midterm and will hopfully learn alot more as the course continues.
Next semester Stanford is offering a wider variety of online courses. One of them is specifically on natural language processing:
There are also some other courses that would make good follow-ups to the AI course.
Machine learning: http://jan2012.ml-class.org/
Probabilistic graphical models: http://www.pgm-class.org/ (These are the kinds of graphs you saw in the AI course - not about pictures)
Game theory: http://www.game-theory-class.org/
Excellent. Thanks, Eric!
Eric Lavigne said:
Next semester Stanford is offering a wider variety of online courses. One of them is specifically on natural language processing:
There are also some other courses that would make good follow-ups to the AI course.
Machine learning: http://jan2012.ml-class.org/
Probabilistic graphical models: http://www.pgm-class.org/ (These are the kinds of graphs you saw in the AI course - not about pictures)
Game theory: http://www.game-theory-class.org/
Codetown is a social network. It's got blogs, forums, groups, personal pages and more! You might think of Codetown as a funky camper van with lots of compartments for your stuff and a great multimedia system, too! Best of all, Codetown has room for all of your friends.
Created by Michael Levin Dec 18, 2008 at 6:56pm. Last updated by Michael Levin May 4, 2018.
Check out the Codetown Jobs group.
Amazon has released Strands Agents, an open source SDK that simplifies AI agent development through a model-driven approach. The framework enables developers to build agents by defining prompts and tool lists with minimal code.
By Vinod GojeThis week's Java roundup for May 26th, 2025 features news highlighting: the twelfth milestone release of GlassFish 8.0; four JEPs targeted for JDK 25; introducing the GPULlama3.java project powered by TornadoVM; and GA releases of Hibernate Reactive 3.0, Spring Modulith 1.4 and Spring Cloud 2025.0.
By Michael RedlichScaling a system is a hard problem to solve. Underinvesting in scalability leads to a shortened lifespan for the system, but overinvesting can kill the MVP business case because of cost.
By Kurt Bittner, Pierre PureurIn this episode, Thomas Betts speaks with Diana Montalion about how architecture is designing for knowledge flow. The conversation covers the differences between knowledge stock and knowledge flow and the importance of a growth mindset. If you’re trying to find new ways to solve problems, you have to start by thinking in new ways.
By Diana MontalionJEP 502 introduces the Stable Values API in JDK 25, enhancing application startup performance by allowing deferred immutability. This feature enables thread-safe, at-most-once initialization of complex objects, combining the best of final fields with lazy initialization. It empowers developers to improve efficiency and reduce initialization pitfalls while leveraging JVM optimizations.
By A N M Bazlur Rahman
© 2025 Created by Michael Levin.
Powered by