23

What did you like most about this course?

What was there not to like.

The first half -- information retrieval, clustering,
crawling, dimensionality reduction.

I really liked the projects, they were very well thought out. They
were right in line with the topics of the class as well.

I thoroughly enjoyed the topics relating to collaborative and 
content based filtering, latent semantic indexing, and page rank. 
The papers that formed the bulk of the literature for the course 
provided stimulating reading. Class was always enjoyable because 
the instructor was very well prepared and gave powerful insights in 
the topics that were covered. He definitely added value to the 
written materials that the course covered. The projects were very 
amusing and the level of detail that was required provided ample 
opportunity for intellectual growth.

Learning about and applying "bleeding-edge" information retrieval 
techniques.
 
Learning about the merging of traditional structured databases and 
unstructured text and the management of such systems.

The topics covered are very closely related to reality and very 
interesting. The instructor shows great enthusiasm in teaching and 
pointed out many interesting direction for future work.

I liked the presentation of the course material. I am already 
familiar with some of the topics presented, but the instructor 
provided a different perspective to what I already know. I greatly 
appreciate the instructor's enthusiasm towards the course. He is 
highly motivated to make this course best and that really has a 
constructive impact on the students. I wish all the instructors 
were like this :) I also liked how he related the in-class material 
with what is happening outside in the real world. Time to time, he 
tried his best to keep the course material as up-to-date as 
possible. I also liked how he brought his phylosophy and principles 
into the homework and midterm questions. I love his homework and 
midterm questions (even though I wasn't able to answer them to 
perfection), because they not only cover the technical aspects of 
the course material but they also reflect the instructor's 
principles towards spiritual and psychological things. One more 
thing is the metaphors that he uses to explain certain concepts. I 
always get amazed by how clear his ideas are on any given topic. 
Wake him up in the middle of night and ask him a question on any 
topic from the course, I am 200% confident that he will explain 
immediately without any hesitation. Since he is very strong in his 
fundamentals and the course material, he was able to relate the 
topics to the real words aspects and explain very well. The last 
thing is the course projects. If not for this course, I wouldn't 
have sit and implemented vector space ranking, pagerank, auth/hubs 
and clustering algorithms. Even though one is clear about the 
algorithms, implementing them and testing on realworld data and 
analyzing the results yields a satisfying experience. 
 
Finally the TA's were awesome. They were very helpful and gave 
constructive feedback on the homeworks and projects. Their comments 
were very helpful. They were available for any questions outside 
the class time and office hours.

Amazing instructor. The instructor is VERY VERY reasonable and 
DEFINTELY understands the students issues. He is very enthusiastic in 
the class which is a motivating factor.
 
The course material was "CURRENT" and the material was excellently 
presented.

I liked the fact that there was no definitive text book for the 
course. I liked the approach of teaching from published papers. It 
was a very challenging course, and was tough.

24

What did you like least about this course?

Would of been nice to have a few more short examples/problems that
better explained, displayed the process of an algorithm. Conceptually
the algorithms/equations used were understandable, but sometimes the
actual calculations of numbers/results was not perfectly clear.

The second half -- information integration.
 
Note that "liked least" does not mean disliked.

Some of the homeworks went into the particulars of topics that in
class we only talked about on a higher level. It may halp to discuss
some of the lower level computations if those topics are going to
appear in the homework.

My only suggestion would be to include a project that used 
collaborative and content based filtering as part of the course 
work. Although, I am not sure that there would be enough time to do 
this.

Nothing... maybe how big the projects were, but they were worth it.

The course title is "Information Retrieval, Mining and Integration 
on the Internet", but the whole course talks about only text. 
Information doesn't only correspond to text! I would have liked to 
see how some of these concepts (retrieval, mining, and integration) 
would apply to other kinds of information, such as images, videos, 
and multimedia data. 

 
At some points of the semester, the instructor tried to push his 
ideas and bias towards certain topics to the class (for e.g., 
semantic Web). I somehow got an opinion that the instructor is 
biased towards wrapper sort of approach, rather than reasoning with 
the meaning of the content. Frankly some of the courses did not 
live up to his mark. He spent some of the classes talking about the 
phylosophical topics all the time. This is lot of fun but at the 
same time he should have a balance between the technical content and 
the social application of the material in each class. 

Amount of work..but it definitely is worthwhile...

The criteria and the method used for grading the projects should 
have been told early, especially the importance of analysis part. 
also I feel there is a huge loophole in the way projects are graded. 
It is possible for some students to fabricate results and do just 
the analysis instead of actually doing the project. Some portions of 
the course like database refresher, introduction to semantic web and 
the introduction to the course can be eliminated. They were nothing 
more than waste of time, and I feel though they are realted to the 
course, the instructor need not waste time on them. We can have them 
as pre-class reading.

25

Comments

Overall, I would rate this course as one of the best courses I have 
taken but not as good as the other course (Intro to AI) offered by 
the same instructor. The syllabus of the course is decent and I 
guess it would change according to the latest developments in the 
field in its next offering. I would like to see more lectures on 
databases/information retrieval intersection. The connection is 
interesting and possibly there could be one course project on this 
area. A project on db/ir connection would certainly elevate the 
students' understanding on database concepts and at the same time 
would provide good insight into ir concepts. 
 
Raspberry Bars are awesome!!! This tradition should continue...

Great Course

the instructor can afford to be much more polite and better behaved, 
when he meets or sees students outside the class. He can atleast nod 
and smile to acknowledge that he noticed his/her presence when their 
eyes meet.

It might be worth thinking about splitting the course into
two different courses. One on information retrieval and the
other on integration issues. Not only would this allow a more
detailed look at each topic, but the projects could be more
involved as well.

Thank you for a fantastic class. It was great and very enlightening.

Kambhampati is the best professor at ASU. Hands down. I've had him 
two semesters in a row.