|
...known as the Turing Test (TT), most typically conducted as anonymous exchanges of English-language text between computers. A panel of judges poses questions to the contestants in order to determine which are human and which are programs. There have been many such experiments since Turing proposed the challenge, but there is considerable disagreement as to what passing the test means, and whether passing it tells us much at all. (2)
We use the TT here as a means of identifying artificial intelligence (AI) technologies that will have a pivotal role in creating more intuitive machine-human interactions. We have chosen six technologies supporting certain computer behaviors that could significantly increase the practical value of computers. In what follows we summarize and editorialize on where each of these technologies stands today, relying heavily on the findings of the conference and workshop "Machine Intelligence and the Turing Test" (3) held last year at the IBM Thomas J. Watson Research Center in Yorktown Heights, New York.
The first technology we address is natural language understanding (NLU). We confess to a particular passion for this part of AI, so fundamental for communication and yet still full of nuances, poorly understood, and hard to symbolize. (4) Even if computers could understand plain English, this would just be the beginning.
Our second technology is machine reasoning (MR). The TT judges ask the contestants questions intended to flush out the mere mechanical responses of a computer. To fool the human judges, a computer will need to provide reasonable answers, answers that are relevant within the context set by earlier exchanges.
Since TT questions can be about any subject, contestants need a very significant knowledge base covering a wide range of human activities such as sports, politics, health, and food. In addition to the rational, objective knowledge about ourselves and our society, the computer's knowledge base needs to also include "rules of thumb," myths, "old wives tales" and urban legends, as well as the complex relationships between facts, theories, conjectures, and judgments. We need knowledge representation (KR) technology to represent this information in all its complexity.
Creating the on-line body of knowledge would be itself a daunting task, and knowledge acquisition (KA) is our fourth AI technology with direct application for any would-be TT winner. The manual effort required to capture this knowledge would be enormous, so computers need to be programmed to listen and learn automatically.
Two less prominent technologies complete our set. There is growing interest in the dynamics of dialog and the role of identity in computer-human exchanges. This has led to theories and experiments in dialog management, (5) and to experiments in making computers react to human emotions. (6)
Natural language understanding
Natural language understanding (NLU) is the technology enabling computers to extract meaning from text--easy and natural for humans, but notoriously difficult as computation. NLU is a key component of software that can accept commands and queries from humans in their own language and produce...
NOTE: All illustrations and photos
have been removed from this article.
More articles from IBM Systems Journal
An architecture of diversity for commonsense reasoning. (Technical for..., September 01, 2002
Arguing A.I.: The Battle for Twenty-First Century Science.~(book revie..., September 01, 2002
Looking for additional articles?
Search our database of over 3 million articles.
Looking for more in-depth information on this industry?
Search our complete database of Industry & Market reports by text, subject, publication
name or publication date.
About Goliath
Whether you're looking for sales prospects, competitive information, company
analysis or best practices in managing your organization,
Goliath can help you meet your business needs.
Our extensive business information databases empower business
professionals with both the breadth and depth of credible,
authoritative information they need to support their business
goals. Whether it be strategic planning, sales prospecting,
company research or defining management best practices -
Goliath is your leading source for accurate information.
|
.%20%20&position=1)