Wednesday, July 29, 2015

Teacher Center A La Sports Center: Hilarious

Key and Peele, talented comedians with tight skits, have done many special pieces, but this is ultra special. A spoof on Sports Central, draft coverage, this one is on Teachers Draft (for High Schools):


Saturday, July 25, 2015

NYT on Terry Tao

Well written article. We need a 21st century media to make a movie out of this story. 

Labels:

Friday, July 24, 2015

Streaming and MapReduce algorithms: Researchers Wanted!

I would love to have a postdoc, senior graduate student or a faculty who is willing to spend 6 months to 1 year at DIMACS/Rutgers, anytime starting Sept '15, working on streaming algorithms, mapreduce/hadoop algorithms, Big Data, all the good theory stuff. Pl. contact me

Labels:

Alberto Apostolico

Alberto, a veteran researcher in algorithms and in particular, string algorithms, passed away recently. My sincere condolences. Alberto was a hub, bringing in many to research in string matching.

Alberto was also a consummate raconteur. If you have stories, please contact Ami.

Srinivas Aluru informs us: Through a private gift, Georgia Tech is establishing a lecture series in Alberto's memory.

Labels:

Big Data, Sublinear Algorithms, MapReduce Algorithms Workshop

Grigory Yaroslavtsev, Alexandr Andoni and I  are organizing the DIMACS workshop on Big Data through the Lens of Sublinear Algorithms, Aug 27--28, at Rutgers. There is a great set of talks, tutorials and a heap of intellect in attendance. Pl. join us if you are interested. Check out the fantastic workshop page!

Labels:

Wednesday, July 15, 2015

A New Direction in the World of Ads

Ad business in its looong history has seen very few transformative moments. Ads business has mostly consisted of the publisher-advertiser team showing things to the users they want to monolog with. Sponsored search was a true transformation: ads were shown based on what users asked--searched for, and it charged for users interacting with the ad, not for been shown one. Perhaps ad exchange was transformative, letting advertisers talk to specific users across contexts.

I am excited now to see some truly new steps in advertising, with miip.  Miip has following components:
  • miip is an InMobi branded ad gadget, 
  • it is customized to the user across sites, apps and aware of the history,
  • it shows curated, beautiful commercial content in a native way,
  • it handles transactions and payments, 
  • it lets users interact with the ad in a social way. 
What is new? 
  • An network that shows ads branding it, is new in ads business, as revolutionary as an Intel Inside sticker on your laptop. 
  • Ads which are primarily user centric gadgets that maintain user state across ad experiences is new. While some of the ``user signals'' are used in ads relevance and targeting, truly individualized ad experience over a prolonged period of time is a real innovation. 
  • Content in mobile is all about beautiful composition (instagram, houzz, pinterest, ..), and making ads beautifully composed is a natural evolution.
  • Handling transaction and payment is more than slapping a BuyItNow button on ads, it is making ads a buying experience. Again, here, content is going through this transformation (eg., video) and about time to make ads go through that experience too.
  • Letting users interact with ads is an ongoing exercise. In the old days, companies uploaded their TV ads to YouTube and let users comment; Or, Hulu used to let viewers tweet the video ads they see, etc. These responses were sparse and not very useful. With Instagram and Facebook, the response of users to ads is more natural, and miip makes it a first class object in the system, trying to make ads a dialog. 
What does it mean to do something transformative? It is a risk, and some of the above aspects may not stick, some maybe truly ahead of its time, etc. But I am glad someone  is instigating and taking the bold step, and even if some of it sticks, advertising business would have been moved to a new place.

ps: Naveen Tiwari, the CEO of InMobi, has a twinkle in his eyes, a big smile on his face, and I am glad to see that instead of declaring victory with doing an India-based B dollar company, he has inspired his people to arch their bows and shoot for something much bigger. 

Sunday, July 05, 2015

Status Update

Due to medical and personal issues in my orbit, I will be less than visible for next 2 months. FYI. 

EPFL IC Research Day

EPFL (Lausanne, Switzerland) School of Computer and Communication Sciences did a Research Day celebration on June 30, 2015.

Typically these research days tend to focus on big tangibles, maybe Big Data (research day in 2014), Drones, Brains, Digital Humanities or something of that ilk. What or who would instigate the Audacity of Theoretical Computer Science? Philippe Gillet  (wish I had more time to talk to him, his research on extraterrestial matter as well as formation of Alps seems interesting) who opened the event told us, Nisheeth passionately defended/projected Theory and that eventually led to this outcome, a research day on The Computational Universe. Dean James Larus followed, and noted that EPFL was voted the best young university in the world. I have served on committees with Jim and know that he will be a serious dean, he has a superb taste on what to showcase and support, quiet professionalism on what needs work in the background, and a solid academic and research sense.

Nisheeth's reach for the ethereal in fundamental theory (Avi W) and theory in flight (Christos P.) played out wonderfully. Avi is our spokesperson for what we know, dont know and want to know; Christos is our ambassador to new lands. Avi, his eyes in constant wonder, spoke about efficient computation, connection to Internet security and limits of human knowledge at the boundary of P vs NP. Avi described how computation arises everywhere,  developed P, NP classes with key examples,  emphasized the universality of NP completeness and ended with a picture of the computational universe, stars in the background. An earlier version of the talk is here.  Christos was introduced as a rock star, and he is. He put science under the computational lens, and discussed Physics (Led to it from Feynmann and is double sided questions about quantum phenomena <-> computing), Economics (Led to it from Nash and equilibria) and eventually to Life itself (neurons in computing, evolution, sex, and thus all of life). He discussed Valiant's vicinal algorithms and their recent extension with PJOINs. His performance was so total that someone in the audience was led to ask if we should be seeking help from theology and philosophy as computational supplements to make progress, and Christos solemnly said, we are and we will. 

It was great to visit EPFL, see friends like Matt Grossglauser (back in EPFL after heading Nokia research lab,  he keeps making important progress each year with prior work on understanding kickstarter to current work on predicting local elections), Pablo Rodriguez (continues to combine cuisine, soccer and telecom), Martin Vetterli (he is presiding over research at Swiss NSF, we spoke about what is achievable in the context of SFs, I am going to find  a way to spend a few days with him and work on wavelet problems, he is one of my favorites),  Elisa Celis (we have a lot in common, but mostly spoke about voluntary markets for user data and bandits with side info),  Della Chuang (am glad I got to watch the evolution of her art and design), and others. Among the projects I connected with in conversations was (a) how to market a local watchmaker,  (b) how to start an venture investment fund among academic folks, (c) TexelPoints app by Mintaton for the Island of Texel in Wadden Sea.

ps: I spoke about online advertising, or how to make billions doing theory. Here is a toned down version, I will be expanding this talk more over time. 

Labels:

Wednesday, July 01, 2015

Christos on the Greek Vote

I was sitting by the lake in Lausanne, winds and waves lapping at the sands, humanity in slim clothes, wine and meat on our plates, when I leaned over and asked Christos Papadimitriou, " Christos, When are you going to write something about the Greek crisis?", and he said in a way we have come to expect from him unfailingly, "I already did". Enjoy (use Google translate)!

Labels: