Data Mining track (9th December, 2010) @ Winter School, IIITH

Today we had lectures from 10 to 12.30 and 14.00 to 16.00 on Data Mining. Professor Vikram Pudi focussed on the following important topics -

* Association Rules
* Classifiers and no free lunch theorem
* Clustering

The theme of our work would be research paper mining. Given an input of research paper collection performing tasks like - extract hierarchy/ontology of topics, search portal for research papers, preprocess: extract metadata from pdfs, indexing, pagerank (based on citations + rank of conference + rank of author), given keywords, find matching topics, best conferences, best authors, best papers, landmark papers etc. We were introduced to the Weka toolkit, the demo of which we will see tomorrow. The rest of the day went in googling and reading mostly. I found another canteen here which offers decent food, so I guess I'll be well fed from now on. No LAN in room forced me to stick to the library here and is sort of inconvenient.

Will keep posting.

Day 1 and 2 - Winter School @ IIIT Hyderabad (7th and 8th December, 2010)

The first day at the winter school went quite decently I would say. The mess breakfast was good. We met Dr. Carolyn Rose, from Carnegie Melon University and she welcomed us to the winter school. We got to know how this idea of a winter school came up, as an attempt to bridge the gap between research in India and CMU. A lot of things she mentioned were quite thought provoking and insightful. Her focus was how this is time that we should decide what we ultimately want to do with our life. If we want to go into research, how to go about it.

Reading, as she said, was the key. She suggested that we should read papers in the field that interests us. To get a hang of how research papers are written and published, she suggested we should go ahead and compare a research paper on a particular topic with maximum citations with the one with very low citations to know what makes a research paper good.

Apart from the general introduction, we had lectures introducing the three tracks for the winter school – 1. Information Retrieval, 2. Data Mining, 3. Speech processing. There was also a lecture on VLSI Design and Embedded Systems by IIITH faculty (I didn't hear a syllable :) )

The most interesting part was the panel discussion about – Graduate school or not? India or US? A lot of pro's and cons of pursuing graduate studies both in India and abroad came up. One being that in India, productivity in general is quite low owing to the social repercussions one faces here. On a question about the perception of Indian students by the faculty in CMU, Carolyn tried to answer as diplomatically as possible that Indian students tend NOT to stick to their promises and usually drop out of the research, just after an MS when they had signed in for a Phd as well. She said it was looked down upon by the Profs and was enough reason one shouldn't get into research unless one's serious to finish it.

Lastly, we were assigned tracks in a rather unconventional way. We all were given one ribbon each – red, blue or yellow. Each color represented a track. In order to get what one likes, one had to negotiate and exchange for the desired color ribbon with a fellow colleague during the tea break.

So that was a smooth way to get equal number of people a track without the tedious job of assigning it to them. I took/ got Data Mining by the way.

Later I went out to see a lil of Hyderabad. Since IIITH is located quite away from the main city we had to go quite far. Didn't do much. Surfed through a mall, walked around. Found a very tempting and delicious looking 'Karachi Bakery'. The cakes on display – I have never seen anything like them. The pastry I had was unbelievably good. We took a bus back to campus.

That was the first day. From tomorrow onwards, there is going to be lot of track specific activities. Lets see how things turn out. Hoping for the best :)