CS 498: Computational Advertising Homework 2 solution


Original Work ?


5/5 - (4 votes)

Question 1 (4 points)

In Section 9.8 (Exercises), solve Problem (3).

Question 2 (4 + 4 = 8 points)

In Section 15.10 (Exercises), solve Problem (2).
Repeat (a) to compute the socially optimal allocation using Generalized Second Price
(GSP) auctions.

Question 3 (3+ 5 = 8 points)

In Section 15.10 (Exercises), solve Problem (6).

Question 4 (5+5 = 10 points)

The objective of this question is to sketch frameworks to achieve privacy during web searching
and browsing. Since many websites restrict services with incognito browsing or ad-blockers,
the goal is to preserve user anonymity (as much as possible) without blocking ads (and hence

Note: This question does not have a single correct answer. Points will be awarded based
on creativity and clarity of answers.

1. We examine strategies to achieve privacy during web search. Towards this goal, an
influential work by [1] introduced a browser extension that obfuscates users’ actual
searches by programmatically generating a stream of decoy searches, with goal of confusing the search engine. Briefly describe the advantages and limitations of such an
approach to achieve privacy, and suggest a framework that improves upon this idea.

2. Consider a web browsing scenario where third-party websites track users across websites
through the use of cookies. Briefly outline a solution to avoid (or mitigate) the amount
of private information made available to trackers, while preserving anonymity as much
as possible. Hint: Analyze the mechanism of cookie synchronization.


[1] Nissenbaum, Helen, and Howe Daniel. ”TrackMeNot: Resisting surveillance in web
search.” (2009).