Models, Algorithms & Hardware

john

Fulbright Research Scholar John Iacono Teaches Why Speed Matters

New York University professor and current Fulbright scholar to Belgium John Iacono illustrated different models to explore the benefit of proper hardware for improving algorithms during his “Finding Things” lecture at the ULB campus.

audience

“We should be looking for ways for structures and applications to run faster,” Iacono said. Algorithm designers can help with hardware design, even creating “elegant algorithms” and thereby contribute to the speed of applications.

To begin his lecture, Iacono illustrated “How to Find a Name in a Phone Book,” through exploring two different methods: one in which you read it like a book or the second method through which you look in the middle and throw away half of the book. The second method is the better one, because at the core of algorithms, according to Iacono, is the principle that there are fundamentally good and bad ways of doing things.

john

Iacono says the way we think about developing algorithms to solve problems is critical as data gets bigger. But first, he says it’s necessary to get the simple questions right in order to solve bigger problems. For example, an issue Iacono and his associates (including Stefan Langerman Director of Research for FNRS) have been working on is the k—sum problem. For this problem, the theoreticians have a set of numbers and want to know the combination of 3 numbers that equals 0. Although the scholars have made progress, they don’t know if their method is the best or the easiest. Such a problem underscores the need to extract problems down to the simplest method. One of the most recurring questions is: Can you do better?

random

Iacono also modeled memory and locality diagrams, emphasizing how improved hardware leads to better algorithms. He says the past several months of his Fulbright period have put him in the company of “wonderful people. There are many visitors to the ULB campus, so there is always a constant stream of people in and out.” For him, conversation is the norm. “I talk a lot with my colleagues. Community is a big part of my work,” Iacono says. “We organize conferences for scholars in their field and we give opinions on publishable papers.”

speed

A typical morning for Iacono might include working alone to brainstorm, which is often succeeded by meetings, and finally lunch with colleagues or other guests. “On Fridays, we don’t go to the office but will have a working lunch at a local spot.”

As a theoretician, application is a different story. In other words, Iacono focuses on figuring out good questions to ask. “Some people are famous for asking the right questions,” Iacono says. “Every day you are doing what you can in random directions.”

thanks

— By Fran Djoukeng