The Hindu Business Line : No dumb machine

The human brain is a very sophisticated machine. The human mind has a natural process of thinking, and has the capability to imagine as well. The basic aim of machine intelligence is to understand the mechanism of human thinking and make a machine think as intelligently and equip it to react faster and more efficiently than human beings.

For example, thousands of blood samples have to be tested urgently to detect a virus. Manual testing will take days before any action can be initiated. In such cases, machine intelligence can help speed up identification of the virus among the blood samples. Again, with regard to detailed census reports, character recognition on the basis of machine intelligence can facilitate speedy collation of detailed and specific data, which would have taken weeks to be collected and collated manually.

Just imagine an American businessman in Washington talking over the phone with his Japanese counterpart based in Tokyo. The American speaks in English but the Japanese gets to hear what is said in his own language. How does its happen ? The telephone exchange through which the call is routed has voice recognition capabilities and translates English into Japanese and Japanese into English on a real-time basis based on machine intelligence capabilities.

It was Alan Turing, a British mathematician who cracked the Nazi codes during World War II, who laid much of the groundwork for the creation of modern digital computing and machine intelligence. While his dream is yet to be fully realised, a group of scientists and academicians located in different geographies across the globe are working overtime to facilitate the fruition of Turing’s dream.

Towards this end, the Indian Statistical Institute (ISI) at Kolkata hosted the Second International Conference on Pattern Recognition and Machine Intelligence in December 2007.

Eighty-two participants from across the globe deliberated and discussed developments pertaining to pattern recognition and machine intelligence and related fields. They talked of encouraging academic and industrial interaction and promoting collaborative research and developmental.

Says Dr Malay Kumar Kundu, Professor in the Machine Intelligence Unit of ISI: “Machine intelligence is a paradigm by which researchers are trying to create a machine which has a human-like decision-making capabilities and can understand human-like correspondence and reacts to an evolving situation just as a human being would.”

Such a paradigm conveys the core concept of pattern recognition and machine learning with advanced technologies like fuzzy logic, artificial neural networks, evolutionary computation, particle swarm optimisation and rough sets, collectively called the soft computing paradigm, he adds.

Dr Kundu says that the soft computing paradigm provides techniques for flexible information processing to deal with real-life ambiguous situations.

The research and investigations currently being done in the Machine Intelligence Unit of ISI comprises developing these technologies individually and in an integrated manner and demonstrating their effectiveness in solving various problems of pattern recognition, machine learning, image and video processing, biometrics, data mining, bio-informatics etc, that are related to the design of intelligent systems. In fact, considerable work on machine intelligence is being done in the fields of web intelligence and data and web mining.

According to Dr Sankar K. Pal, Director of ISI, the applications of machine intelligence are infinite.

From applications in the medical, healthcare and defence sectors to sourcing of remote sensing data, voice recognition, dynamic imaging, weather forecasting, gathering mining intelligence inputs, etc, machine intelligence can find varied applications, says Dr Pal.

Dr Pal feels that a lot of research work remains to be done in the domain of machine intelligence. “There will be progress in the field but human expectations will always march ahead of technological progress. Machine intelligence is based on intelligent algorithms and facilitates faster decision-making. A human brain has its own limitations. And it is here that the machine scores”, says Dr Pal.

ambar_singhroy@rediffmail.com