Neela's blog

There are many online resources themed around grad school humor. But if you look closely, they are not really about fun in grad school, but rather about how a grad student is a source of amusement to others. And here is one of the major reasons why it is so -

In a nutshell, the definition of humor is different and very scholarly for grad students as compared to the rest of the population. In the first year of grad school, the jokes are pretty much funny to around 40 fellow grad students who are taking/ have taken the same courses. By the third year, the count decreases to 5 students who are working in the same research lab. And by the end of the fifth year, only the student is capable of getting some of his/her 'deep' jokes. Of course, I am still in my first year, but the above hypothesis seems to be true at least to the sample population I have observed so far.

I guess this phenomenon is mainly a result of the fact that once in grad school, each one of us starts relating everything to his/her research, even the things that are not even remotely connected. It may come from a belief that one's research is the answer to most, if not all, problems. So a math student starts computing probability of a surprize quiz given the TA's last paper was rejected (and of course the number of topics covered since the last quiz). A social-psychology student starts seeking examples of collaboration and social loafing among his/her peers for some group assignment. I guess this is what makes serious research fun for grad students. So the morale of the blog is, grad school can be as funny as one can make it. One just has to look in the right places!

Graduate school must be the most unique experience of academic life. At least it is the first time I am stepping out of the protective shell and chasing problems without an idea of where they will lead me. No more finding answers in the textbooks or even online, no more comparing course notes with friends - because what I as a PhD student have to do, is seek answers to yet unsolved problems. Probably I will be the only one in my department, my university studying that problem, working alone for the duration of four or more years untiI I make a significant contribution to the field.

I ask myself - Is this scary? Can I get through these years? What should I do when I am feeling low? Is there a mantra for success?

Well, I guess perhaps the most important thing to always remember is that every Ph.D student is just about in the same situation. Me and my friends totally agree on this point - so we now have a pact. No matter what happens, we all will get together every friday evening and leave worries of the week behind. We cannot work more than 8 hours a day so we have to make sure that we finish all tasks with peak efficiency without spilling over the alloted time. If we have to finish some writing assignments, we will coordinate the timings so that we sit together and get those assignments cracking. This way we can easily prevent each other from loafing around or wasting time on checking e-mails. We will also share our progress, be it in terms of research or paper publication or thesis writing.

I am sure these small things will make a big difference in sailing through the PhD years. It is not only about getting work done efficiently, but it is also about creating a support group, a group of  friends one can depend on and I truly believe that, that is the mantra for flourishing in graduate school.

 

I attended Dr. Andrew Zisserman's keynote talk on visual search, the VideoGoogle, in the Asian Conference on Computer Vision in 2006. With Google being an expert in text search, a similar application for videos is definitely the next-gen search application. The talk was magic, and novices like me were held captive when he presented a visual search demo on locating objects precisely in a movie, irrespective of the location, scale and viewing direction. It was finding a needle in the haystack. His infectious enthusiasm, willingness to discuss ideas with experts and novice students alike, and an extremely down-to-earth nature definitely won him many hearts at the conference. I, at least, was very inspired and drawn to the problem of visual search. Since then, I have been following upon Dr. Zisserman's work in detail and I find it very much fundamental, yet very applied to real-world problems.

Dr. Andrew Zisserman is currently a professor at the University of Oxford. To talk about his impact in a non-subjective manner is easy - perhaps 1 sentence long - "Andrew Zisserman has a Hirsch-index of 40". The Hirsch-index or the h-index as it is more commonly known, is a quantification of the actual scientific productivity and the apparent scientific impact of a scientist. A value of 40 means that 40 of Dr. Zisserman's publications have been cited by at least 40 other scientific publications each. Talk about impact! He has contributed to theoretical computer vision work as well as some very applied problems. He has co-authored a textbook "Multiple View Geometry in Computer Vision" with Dr. Richard Hartley. He continues to influence students and researchers across the world.

Long live Andrew Zisserman!

A number of people will argue that Ph.D. students, especially those in technical fields, lack a decent social life. Well, I am certainly not going to risk making any such controversial statements. On the other hand, I do believe that a large part of our social life stems from professional relationships that are formed at different conferences and gatherings of professional communities. I have made some great friends at conferences and I enjoy corresponding/ meeting with them every now and then.

I am already a student member of the ACM (Association for Computing Machinery) which is
the world's largest educational and scientific computing society. I am particularly involved in the activities of SIGMM which is a special interest group in Multimedia. This community strives to build cutting edge multimedia technology through publications, conferences and special workshops that provide a platform for researchers world-wide to interact.

Another leading association is the IEEE (Institute of Electrical and Electronic Engineers) which has similar structure of conferences, journals and workshops. What is different about the IEEE is that it also defines standards and measures for various technologies. The ACM and IEEE digital libraries are probably the highest visited sites by scientific community and are really a boon for students and researchers alike. By bringing global research to one's doorstep, they have truly enabled scientific progress like never before.

The impact of these organizations is even more visible in conferences. I attended the ACM Multimedia conference in 2006 and 2007 where I got to interact with leading researchers in multimedia community. ACM Multimedia is a conference that gives a flavor of all aspects of multimedia computing: from underlying technologies to applications, theory to practice, and servers to networks to devices. Other prominent conferences that I am interested in, are ICCV (International Conference on Computer Vision), ICPR (International Conference on Pattern Recognition), CVPR (Computer Vision and Pattern Recognition) and ACCV (Asian Conference on Computer Vision). I remember listening to Dr. Andrew Zisserman's keynote talk at the ACCV 2006 on VideoGoogle. This was the talk that inspired me to look into the problem of multimedia retrieval in the first place. Over last 3-4 years, I have closely followed the above conferences and become really interested in the aspects of pattern recognition and multimedia retrieval. I believe they have played an important role in making me want to become a researcher.

 

Well, this week's assignment is to blog about a senior IST graduate student from our respective labs. Unfortunately, Dr. Wang currently does not have a student who is in third or fourth year of Ph.D. So, I have decided to introduce to you Saurabh Kataria, a third year IST student co-advised by Dr. Giles and Dr. Prasenjit Mitra. Before coming to Penn State, Saurabh was working as a software developer in GlobalLogic, a startup based in India. He has a  Bachelors degree in Computer Science from Institute of Technology, Banaras Hindu University.

Saurabh is currently working on extraction of data points from two-dimensional plots in digital documents. He has published four papers around this topic, two of them as the first author. He, like me, is interested in understanding the digital media and extracting useful information that can help create more utility for a digital artifact. Best wishes to him for a great career!

 

Complementing my earlier blog post, this blog is about the professional activities and achievements of my adviser, Dr. James Wang. Dr. Wang received Bachelor's degree in Mathematics and Computer Science from University of Minnesota, an M. S. in Mathematics and another M. S. in Computer Science, both from Stanford University. He has a Ph. D. in Medical Information Sciences from Stanford University.  He is now an associate professor of the College of Information Sciences and Technology, the Computer Science and Engineering department, and the Integrative Biosciences (IBIOS) Program at the Pennsylvania State University. He is also the Vice Director of the Intelligent Information Systems Laboratory of which I am a member. He has been a Visiting Professor of the Robotics Institute, Carnegie Mellon University in 2007-2008. 
 
Dr. Wang specializes in visual database search and retrieval. He is the co-developer of two of the most famous image retrieval systems - the SIMPLIcity (semantics-sensitive image retrieval system) and the ALIPR (automatic linguistic indexing of pictures - real time).

Dr. Wang has authored two monographs, nearly 100 journal articles, book chapters, and many conference papers. But his emphasis has always been on publishing in top-tier journals. There is a Wikipedia entry on him. The best places to find more details about Dr. Wang are his website and the introduction in Penn State's directory.

One of the major attractions that IST building holds is a great collection of photographs on the second floor, often mistaken as professional wall papers downloaded from the web. But then, next to these photos is a short bio and a picture of the photographer himself. The photographer is James Wang, a researcher and professor at IST and also my adviser. James is an excellent photographer who happens to be doing fundamental research in the aesthetics and analysis of photographs. Check out some of his amazing collections clicked over the last ten years. Its truly great that his research is driven by one of his greatest passions.

Apart from photography, his other passions include traveling. He has traveled to more than 25 countries from Tunisia to Finland to India and so many others. He loves listening to classical music, collecting stamps and coins and also playing with high performance super computers. He currently guides the image analysis research at the Intelligent Information Systems lab  with his wife Jia Li who is a professor of Statistics at the Pennsylvania state university. Its really good fun interacting with him because of all the diversity.

The College of Information Sciences and Technology (IST) is Penn state's I-school and now a leading member of the I-school caucus. IST was established in 1999 as a conscious move to bring together the three founding aspects of a successful information system: the people, the technology and the information itself.

IST offers two bachelors degree (B.S.) programs - information sciences and technology and security and risk analysis. At graduate level, there are 5 major tracks which are more or less like guidelines. These tracks are computational informatics, AI and cognitive science, human-computer interaction, social and enterprise informatics and security informatics. Check out the centers and laboratories for more information.

IST is multi-disciplinary in nature. It provides a great platform for students as well as faculties from diverse backgrounds to interact. IST members come from many different cultures and countries like US, China, India, Korea, Thailand, just to name a few. There are seventeen formally listed research areas including AI and informatics, community informatics, crisis management and emergency response, globalization, health and bio informatics, information policy, social network analysis, etc. Faculty members specialize in many areas such as computer science, engineering, psychology, chemistry, statistics. Its absolutely wonderful to talk with them, to learn about something from broader and new angles than just one narrow aspect. I joined IST not knowing the meaning of an I-school. I joined here to be able to work in a specific research group, as I am sure many of my friends have. But the more I interact with different faculty and students, the more convinced I become of stumbling upon something rather amazing. Coming from a focused background of computer science, understanding problems from these other angles really has become "thinking out of the box".

Here is a picture of the IST building (taken from IST site) that became operational in 2004. The architecture of this 199000 square foot building is inspired by the Ponte Vecchio in Florence. IST building currently hosts the IST as well as the Computer Science department. In the words of the former IST Dean James B. Thomas, this building is "an intellectual bridge linking education and research to the new needs of society". How appropriate!

The 21st century will always be hailed as the era of information explosion. A significant portion of our every day life is spent surfing the web, reading news, articles, blogs as well as writing mails, blog posts and playing games. Online communities are being formed to cater to specific interests. Academic institutions as well as libraries are providing online material. Even the most common people are speedily creating and sharing information. It has become important like never before, to understand the technology enabled role of information in the lives of people.

A wonderful emergent movement to systematically study the interaction between information, people and technology has been initiated in the form of I-schools or the schools of Information. The I-school caucus consists of 21 I-schools across US and Europe with multi-disciplinary programs focusing on diverse fields like information sciences, computer science, human computer interaction, library sciences, policy, economics, etc. I-schools bring together people from different fields and encourage the interaction for building tomorrow's information systems. Please follow this wikipedia entry for more information and links on i-schools.

I feel that we can no longer escape the need for the confluence of information, people and technology. To create useful information we need to understand the users of information.  To create useful systems we need to understand  modeling of information with technology. I am really glad that I am a part of this movement as a student of IST, an i-school dedicated to this cause.

It was not until a very recent class of IST 501 (Integration of IST), that I found my niche. We were assigned a paper titled "The Montagues and the Capulets" that talks about the conflicting interests of computer scientists and social scientists resulting in complications with information systems development. Computers scientists are the "Montagues" or the theorists and social scientists are the "Capulets"  or the pragmatists. As I read this paper, it seemed to me that the authors of this paper were truly anti-computer scientists (or rather anti-all-the-theorists). Let me quote a few points from this paper along with my immediate reactions marked in color:

• "In the last four decades the Montagues have developed a startling and confusing number of languages" (evil laugh .. Muhahahahaha)
  

• "Whereas the Montagues see knowledge representation as an end in itself, the Capulets see it as a means to an end, and that end is Science". (How could you! Sure we like to create generic and well-behaved theories! We need to worry about `n' disciplines, and `n' can be really large! We have to make something that works for all kinds and even provide plug-ins for new requirements. And this is the thanks that we get in return? Are we expected to hard code domain-specific inputs in our model and spoon-feed? Oh, I am heart-broken :(( )
  

• "The operating timescale of the Capulets is immediate unlike the Motagues" (Now I am just plain angry!)
  

I have never read a paper that made my blood boil, but this one did a perfect job at that, at least initially. But the paper does raise some valid points regarding the necessity for better interaction between the pragmatists and the theorists. And not to forget, it also talks about a number of ways a Montague can frustrate a Capulet :D (and vice a versa).  I highly recommend this paper to everyone in the concerned fields. Truly an awesome read!

This paper got me thinking about what I am truly passionate about. I used to think it would be music or reading. But now I know what I love to do the most! I love to solve puzzles and problems. I like devising efficient algorithms and well-behaved theories that can be generically applied, and I would just love to be able to do some ground-breaking theoretical research. I know now that I am ... I am... a GEEK! I can't identify more with Cecilia from the PhDComics. Is it sad? Is it good? I don't know!