- IBM Develops Analytics Technology For Telcos
- A USB Hard Drive That Asks For Your PIN Before Allowing Access
- An Information Security Health Check-up For IBM Clients
- Enterprise Applications And Mid-tier Caching
- India Needs More Homegrown PhDs In Computer Science
- IBM: An Education Tourism Programme For IT Professionals And Students
Page 2 of 2
There are many topics that crop up in AI for which LISP-based machines alone are not sufficient. Some of these are perception problems for vision and speech, motor problems of manipulation (and associated mobility issues in robotics), and the symbol manipulation problems of language. Though we have addressed many issues associated with the implementation of AI, there are many still left.
Matching relevant information from large knowledge bases is a serious issue when it comes to large search engines. Here, AI algorithms can be implemented effectively. Integle (www.integle.org) is an example of this type of intelligent search engine. If we stick to LISP alone, we may have similar problems when it comes to repetitive knowledge, recognition of rule-based inferences in expert systems, combinatorial searches (in game playing and puzzle solving), and massively parallel processing for instantaneous extraction. This problem has not yet been completely addressed.
The paper "Super Chips for Artificial Intelligence" (by R Reddy, professor, Carnegie-Mellon University, Pittsburgh, PA, USA) showed the practical side of AI implementation in a far more lucid way. Yet, it is difficult to go for a single architecture, as many of the problems mentioned above entail a wide variety of computational tasks. But this doesn't seem to be a major issue. We have the Intia processor to tackle definite processing challenges concerning graph processing. It has a specially designed graph processing core (GPC).
The business side
Today, it is easy to design and build a robot that is computationally powerful, easy to work on, and large enough to operate in a real world home environment. But hardly anyone thinks of the powerful AI algorithms for business. No one tries to apply the improvised algorithms of the AIML (Artificial Intelligence Mark-up Language) and the ALICE (the artificial linguistic Internet computer entity) bot features in a call centre, even though many of them are much cheaper and more efficient than other means. You can also try computer vision, speech recognition (or generation) and sophisticated artificial intelligence techniques in your business. I will be dealing with these aspects in future columns. Stay tuned!