Paul Erdos, the brilliant eccentric who was considered one of the century's greatest and most prolific mathematicians, died Friday while attending a conference in Poland. He was 83.

The Hungarian American scientist, winner of the World Prize, the highest-paying award in mathematics, died at the Medical Academy Hospital in Warsaw after suffering a heart attack, the Hungarian Embassy said.

Erdos was born in 1913 to a Jewish family in Budapest, the only child of two mathematicians. By the age of 3, he could amaze visitors by multiplying three-digit numbers in his head.

Never married, he spent most of his waking hours solving mathematical problems and rarely lived in a permanent home. Instead, he traveled from university to university, from conference to conference, living on honorariums and the kindness of other mathematicians.

"He may have been the most prolific mathematician of this century in terms of the amount of papers he wrote and the research he inspired," said Arthur Benjamin, associate professor of mathematics at Harvey Mudd College in Claremont. "He was a wandering minstrel of mathematics. We all thought he would live forever."

A member of Britain's Royal Society and the national academies of Hungary, the Netherlands, the United States, Australia and India, Erdos (pronounced ER-dish) published more than 1,500 articles. He was best known for his work in number theory, the theory of sets and probability theory. He helped found a branch of mathematics called combinatorics that deals with large numbers of objects that must be counted and classified and is widely used in computer science.

"Very often his proofs are quite elegant," a USC mathematician told the Times in 1983. "Sometimes, they involve just better and deeper techniques than other people normally have at their fingertips."

Erdos was also known for his great generosity (he often gave away his grant and prize money to needy mathematicians) and for his absent-mindedness--he often forgot his nine-digit Social Security number.

He lived and breathed mathematics and could fall asleep at a dinner table if mathematics wasn't the topic of conversation. Benjamin recalled a conference in February when Erdos fainted in the middle of a lecture in Baton Rouge, La. As he was being taken out on a stretcher, he regained consciousness and stood up, insisting on completing the proof he had started.

What drove a man to such obsession with a field many others considered baffling and arcane?

"This would be like if you asked Bach what pleasure does he get from his composing," Erdos told the Times in 1983. "Probably when you suddenly see something which was hidden. Something aesthetic."

Times wires services contributed to this story.