Bhaumik Chokshi

Liked application of the methods and the analysis part of the projects;

Liked the idea of using anchor text for retrieving images.

Bob Leaman

Impressed by how AI techniques can be used in real applications; liked the overtones of the class discussion, such as the significance and impact of bias in machine learning.

Using domain knowledge is important to make real application work.

Chance Kirsch

Liked social networks for their application  to many web applications;

Liked information extraction;

Wondered how to come up with something better than the current search engines;

Garrett Wolf

Liked vector similarity ranking model and LSI;

Liked the connection between image processing and clustering/classification problems;

Interested in scalability of the techniques;

Hemal Khatri

Liked the discussion about DB/IR interaction;

Interested in how search engines can answer questions (as against retrieving documents)

Interested in topic specific search and personalized search;

Search engine should use more types of relevance feedback;

Madhu Datla

Felt that the course went significantly beyond search engine techniques;

Liked information integration/extraction discussion;

Liked the connections between topics;

Martin Lehner

Liked social network, folksonomies, NBC;

Felt the course is difficult since it brought together many different things such as AI and DB;

Interested in more AI techniques;

Mingyi Shu

Liked vector similarity ranking and PageRank, which are fundamental but very useful in practice;

Appreciated being able to see the value of the techniques behind google/amazon;

Liked the discussion of information extraction and social networks;

The course is a combination of many different areas;

Homework and projects are difficult but helpful and interesting;

Munmun De Choudhury

Liked vector similarity ranking;

Liked LSI but wondered why Google doesn’t use it;

Enjoyed the discussion on DB/IR interaction and using source statistics in query processing;

Nicholas Radtke

Liked vector similarity ranking and LSI;

Felt that pageRank makes questionable assumptions on the random surfing model;

Would like to see more clustering techniques beyond K-means;

XML and XML query languages are useful;

Liked folksonomy but the problem is how to get people to do it;

Impressed by how simple ideas work in real applications;

Niyati Parikh

Liked most topics covered.

Impressed by how simple ideas such as vector similarity ranking can be so effective in search;

Liked relevance feedback and wondered why it is not used in real search engines;

Liked social networks and wondered what other applications can take advantage of such techniques;

Siddharth Raghavan

Liked social networks and their applications in folksonomies etc.;

Felt the topics covered are very practical and liked to know the techniques behind google/amazon;

Liked the applicability of the techniques in a wide range of domains;

Syed Toufeeq Ahmed

Interested in research on how to extract events and rare patterns;

Interested in how to broadcast information fast – connection with social networks

Lei Tang

Impressed by how some unrealistic assumptions often works in real applications and wondered what the truth/true hypothesis is behind the text documents;

Would like to know more about natural language processing and multimedia information retrieval;