Information Retrieval

CS 6501 Fall 2014

HW1—What is Missing in Current Search Engine Services

Posted: September 03rd, 2014

Innovations are often motivated by a dissatisfaction with our current living environment and a vision about how to address this dissatisfaction and change our world for the better. For example, the visionary article As We May Think written by Vannevar Bush in 1945 stimulated research in information retrieval in early days. The Memex machine vision laid out in this article was based on an analysis of the world then, which led to the recognition of many inadequacies of the conventional library systems in helping a researcher to collect, select, and digest all the accumulated knowledge in the world. The Memex machine was then "designed" to remove these inadequacies.

Do a similar analysis of the world today in terms of information services provided to people by a search engine like Google, identify some major limitations, and write a short essay (with 400~600 words) to analyze the limitations and discuss possible ways to break these limitations and thus improve the current search engine services. While there are many ways to improve current search engines, you are also encouraged to envision a new system or function very different from the current search services with ideas that go beyond search to help people access information in other ways or better support users in decision making.

A good way to think about it is to consider the needs of yourself and people around you. Can we do a better job to improve the way we access, manage, and utilize information? For example, we all spend some time every day to browse news and social media, can we develop some tool to link those two sources of information and cross reference each other, in order to provide us a comprehensive understanding of the world? Trustworthiness is a serious problem in social media and search results, can we identify trustful source of information by some machine learning and information retrieval techniques? Can a system recommend activities for us based on our friends schedule and interest, so that we can quickly get to know more new friends.

In terms of functions, in general, we can think about how to move beyond search toward information access. Search is just one way to access information. What about information recommendation? What about allowing a user to find information flexibly in multiple ways (browsing + search + recommendation)? How can we better present search results? We can also further think about how to move beyond information access to task support or decision-making support. For example, many users use a Web search engine to find information about products to help optimizing their purchasing decisions. How can we best help such users more directly to finish their tasks? Systems like Google Products and Orbitz have been designed for this purpose. Can we do better? Many people want to compare products, can we pull out reviews about similar products and generate a "comparative summary" to help people compare these products? Information integration is another major line of directions. We currently have two extremes - either search the entire Web or search a specific website. Can we provide some federated search portal that would allow us to search all the computer science department websites? How do we integrate the search results and present them in a useful way to users in such a case? For example, can we generate a comprehensive summary of all the Ph.D. programs in Computer Science and then compare them?

If you cannot seem to come up with ideas, you can try to identify a specific (difficult) query that doesn't seem to work well on Google. You can then try to analyze the reason why the query doesn't work well, and discuss how you can possibly improve the search engine based on your analysis.

This exercise is intended to lead you to start thinking about what you can do for your course projects. Your real course project doesn't have to follow the idea(s) that you discuss in this essay, though; as time goes and you learn more about the topic of information retrieval, you might have other good ideas for your project or you might want to work with others on a joint project.

Submission

Your essay will be graded based on the novelty and practical/research value of the idea(s) you have discussed. It is your duty to convince the instructor about your idea(s)!

Please send a PDF version of your essay to the instructor before the deadline (no restriction of format or template to use). Your grade, in a scale of 0-100 points, will be notified by email after the deadline.

Deadline

The deadline for this assignment is 11:59pm, Friday, September 19th.

Selected essays from students:

Below is a list of selected essays from our students, from where you can find inspiring new ideas for information retrieval research and how such techniques can impact our life. I believe they are good starting points for preparing your course project proposals.