Mihai Pătraşcu (1982-2012)

Update: Visit the Memorial Website

Mihai Pătraşcu passed away Tuesday after a bout with brain cancer. Even though he hadn't reached his 30th birthday, he had already produced a number of major achievements in the algorithmic community. The theoretical computer science community lost one of its brightest young stars.

During STOC, Yevgeniy Dodis arranged to have Mihai at the business meeting while we announced him as a co-winner of the EAPresburger award which Mihai will receive posthumously at ICALP next month. The standing ovation that followed made for the most emotional moment I have ever seen at an academic conference.

We asked Mohammad Taghi Hajiaghayi to write some personal thoughts on Mihai:

Last night I heard the very sad news that Mihai Pătraşcu has passed away at the age of 29. Though I knew about his cancer from 1.5 years ago and several steps that he took to fight the cancer still the news was quite shocking to me. Mihai's carreer was short but very productive with his lots of great papers in e.g., STOC, FOCS, and SODA.


Indeed Mihai is the second close friend of mine and a bright star researcher from my MIT times that I missed in the recent years (the first one was Misha Alekhnovich who died in 2006 in a white-water kayaking accident in Russia).


I became familiar with Mihai since 2002 when he joined MIT. He was interested in working on date structure with Erik Demaine who was my Ph.D. advisor as well. As a result, I got know him much better esp. since both of us had common experiences in IOI (International Olympiad in Informatics). One of Mihai's early ground-breaking papers is

Erik D. Demaine, Dion Harmon, John Iacono, Mihai Patrascu: Dynamic Optimality - Almost. FOCS 2004: 484-490

In this paper, an O(lg lg n)-competitive online binary search tree is given which improves upon the best previous competitive ratio of O(lg n). This was the first major progress on Sleator and Tarjans dynamic optimality conjecture of 1985 that O(1)-competitive binary search trees exist. Indeed Mihai could formulate the O(1)-competitive conjecture as an approximation factor of some algorithm and we worked together for some time to prove the conjecture (that so far we could not:)). This was my first research experience with Mihai. Since then we always have talked about research (but we never had a joint paper).


In general Mihai and I were close friends and talked about all things. For example some times the people asked me about Mihai's personality especially because of some posts in his blog that were a bit controversial. I always told them Mihai is a very nice guy to work or chat with and those posts are indeed some concerns that everyone has in this mind, but Mihai is just brave enough to mention them loudly in his blog for the hope of some changes in the way that our community think (and anyone else including myself may agree or disagree with them).


After Mihai's graduation with Ph.D. from MIT which was only for 2 years (the same was true for Misha Alekhnovich), he applied for lots of places and got several very good offers. I was one of the people who encouraged him to join AT&T research labs, the place that I was there at that time. Finally Mihai decided to decline several faculty offers and join AT&T esp. to work with Mikkel Thorup (before joining AT&T he went as a Raviv Postdoctoral Fellow to IBM Almaden for one year). He has also chosen AT&T because of its place near to NYC that was good for his wife's career. As a result since July 2009, I was very happy to have Mihai, a star researcher, as my colleague at AT&T research labs and we talked more about everything during the past years. For example, later I wanted to teach hashing in my Introduction to Algorithms course at UMD; I asked Mihai and he gave me a very good overview about the field including an overview of his recent ground-breaking paper on simple tabulation hashing.


I first knew about Mihai's cancer in Feb 2011. At that time he was taking chemo before a brain surgery in March 2011. Just a day after the surgery he sent an email to David Johnson from the hospital that he already recovered and started thinking about his research problems. After this surgery, doctors, Mihai, and all of us at AT&T were very hopeful that the bad days for Mihai have been passed and Mihai will recover fully and produce lots of papers again:). Indeed everything was fine since then and Mihai even applied for faculty positions in Fall 2011 (he always considered his AT&T (industry) position as a long-term post-doc position). He asked for my opinion about his research statement and I was happy to give him some comments (esp. since I was also a person who first took the path of the industry and then academia). Again everything was fine until mid Jan 2012 that Mihai understood the cancer is back. Unfortunately this time the cancer was very strong and in a week or two he went to a wheelchair from a completely healthy position. I saw him several times at AT&T recently (when I visited there) and I always was very sad to see him on a wheelchair. To my mind, the best good news for Mihai since January 2012 was the fact that Mihai was selected as a co-winner of the 2012 EATCS Presburger Young Scientist Award for his ground-breaking work on data structure lower-bounds. He was very happy about it as he posted in his blog about it as well.


At the end I pray that Mihai's soul rests in peace and the research area of data structure lower-bounds in which he had several ground-breaking works becomes more prosperous due to his work.