How do we acquire Language Skills?
Coming close to speech capability in many ways is linguistic capability. Humans (even illiterates) have phenomenal linguistic ability. To analyze and understand this is very difficult in case of grown ups. It is hopefully easier (and definitely much more interesting) to do language capability studies with growing children, while they are learning to speak. This can be done by studying what the growing child says at various stages - when he is one, one and a half or two.

We at Tata Infotech have done some very interesting studies in this area, of which we will talk later.

The 'Nuts and Bolts' of Human Intelligence
It would be tempting to understand human intelligence using a so called 'bottom up' approach. Intelligence is composed of a variety of sub components, each relating to one type of activity. Most types of intelligent tasks have in some sense been implemented as computer programs. One could claim that together, these amount to implementing human intelligence in machines.

We can also take a nuts and bolts view of the human brain. Can we implement or simulate in a computer, the nuts and bolts that together make up the human brain? We do know how the building blocks (neurons) work. We even know how they grow and connect up in case of insect brains!

How has artificial intelligence fared over the years, from the common person's point of view? Work in this area started in the 1950's. The first decade saw much optimism and enthusiasm - but little success. Later progress was much slower, even disappointing.

Automatic machine translation (of documents, from one language to another) was an eagerly awaited application. Despite much promise, there was little achievement even here. Automatic theorem proving attempts were quite successful, but this field merely remained a scientific curiosity.

Hand-eye systems (where a TV camera scans a scene and a robotic arm manipulates objects in that scene as desired by the operator) worked for a variety of situations, but not so well in noisy environments. Chess playing programs have been around for a long time. However, they remained at the level of competence of an experienced amateur (till quite recently).

Systems like IBM's DEEP BLUE, continuously improved over time with experience. However, being handcrafted, they remained confined to that application alone.

Even general and specialized problem solvers fell short of promise.

We can take a nuts-and-bolts view of the human brain. Can we implement or simulate in a computer, the nuts and bolts that together make up the human brain?

<<< previous page

TATA INFOTECH - TOTAL SOLUTIONS TOTAL COMMITMENT