This post is about taking stock of my Ph.D progress. I suppose I should do that now. Part I dealt with many sundry events in the summer, but I left out what I did for my Ph.D.
At the beginning of the summer, I started reading "Optimality and Approximate Optimality of Source-Channel Separation in Networks" by Tian, Chen, Diggavi, and Shamai. (OAOSCSN from here on). A really nice paper, it deals with issues of source, network and channel separation in networks.
After about a week of reading, I realized that I did not know enough about many aspects of network information theory to be able to effectively decode the paper. I was fortunate to obtain Gamal and Kim's Lecture Notes in Network Information Theory (LNIT) around May end during my Bangalore (and IISc) trip from Prof. Sundar Rajan, my advisor's (Bikash Dey's) advisor. (Hopefully, in about 2-2.5 years' time, my academic grandfather).
Much of June was spent decoding parts of LNIT, and I have made progress on the 1st fifteen chapters of the book. I have more or less successfully decoded OAOSCSN.
Currently, I find the whole topic of separation problems and theorems fascinating and have been reading up quite a bit on the topic. Besides the work of the above group, there are great papers by Effros's, Medard's groups and also some by Gastpar.
Hopes/ Plans for the coming semester
=======================
I have not been assigned T.A duty this semester, so I hope to be able to accomplish relatively more.
In decreasing order of priority:
1. Make sufficient progress on a problem to be able to write a conference paper for the International Symposium on Information Theory (ISIT) 2012. The deadline is Feb 2012. Some problems have been suggested in OAOSCSN itself.
2. Decode all of Gamal and Kim's LNIT by the end of the semester, with the exception of one or two chapters, which do not interest me much. That is a total of 24 chapters, less one or two. Many open problems are suggested at the chapter end sections, and also some great references.
3. Audit Prof. Rajbabu's course EE779 on "Advanced Topics in Signal Processing". Sit-through Prof. Diwan's lectures on Network Information Flow (and possibly Randomized Algorithms) in his course CS601 "Algorithms and Complexity". Time permitting, maybe learn a little bit of Saravanan's EE605 "Error-Correcting Codes".
Again, I wonder if I am being too ambitious.
At the beginning of the summer, I started reading "Optimality and Approximate Optimality of Source-Channel Separation in Networks" by Tian, Chen, Diggavi, and Shamai. (OAOSCSN from here on). A really nice paper, it deals with issues of source, network and channel separation in networks.
After about a week of reading, I realized that I did not know enough about many aspects of network information theory to be able to effectively decode the paper. I was fortunate to obtain Gamal and Kim's Lecture Notes in Network Information Theory (LNIT) around May end during my Bangalore (and IISc) trip from Prof. Sundar Rajan, my advisor's (Bikash Dey's) advisor. (Hopefully, in about 2-2.5 years' time, my academic grandfather).
Much of June was spent decoding parts of LNIT, and I have made progress on the 1st fifteen chapters of the book. I have more or less successfully decoded OAOSCSN.
Currently, I find the whole topic of separation problems and theorems fascinating and have been reading up quite a bit on the topic. Besides the work of the above group, there are great papers by Effros's, Medard's groups and also some by Gastpar.
Hopes/ Plans for the coming semester
=======================
I have not been assigned T.A duty this semester, so I hope to be able to accomplish relatively more.
In decreasing order of priority:
1. Make sufficient progress on a problem to be able to write a conference paper for the International Symposium on Information Theory (ISIT) 2012. The deadline is Feb 2012. Some problems have been suggested in OAOSCSN itself.
2. Decode all of Gamal and Kim's LNIT by the end of the semester, with the exception of one or two chapters, which do not interest me much. That is a total of 24 chapters, less one or two. Many open problems are suggested at the chapter end sections, and also some great references.
3. Audit Prof. Rajbabu's course EE779 on "Advanced Topics in Signal Processing". Sit-through Prof. Diwan's lectures on Network Information Flow (and possibly Randomized Algorithms) in his course CS601 "Algorithms and Complexity". Time permitting, maybe learn a little bit of Saravanan's EE605 "Error-Correcting Codes".
Again, I wonder if I am being too ambitious.
But, have you not already taken Error Correcting Codes of the level of Lin and Costello ? Is Dr. Saravanan's course based on van Lint's book? Let me know when you have time.
ReplyDeleteIt is good to be ambitious or over-ambitious otherwise one is just plain lazy. But, of course, what you are trying is to accomplish a non-trivial Ph.D. thesis. I admire you for that.
ReplyDeleteThe search for a non-trivial problem itself is a problem. Once, you get that in your focus, you can learn the other math tools or other sophisticated CS tools/techniques/algos.
It could happen that in the initial phase of a Ph.D. one moves in "circles". Re this I will give you an article of Steven Weiberg, Physics Nobel Laureate. Perhaps, it may help.
Cheers.