ICD Research Projects at the School of Information
Table of Contents
Current ICD Projects
We are conducting field and lab experiments on the design of online mechanisms for charitable giving and fund-raising.
Researchers draw on theories and data from social psychology and public goods economics to drive design decisions about online communities. Their goal is to increase participants' contributions to the communal good.
Our current project is an NSF-funded effort to develop methods for empirical game-theoretic analysis of strategic settings that are intractable for traditional mathematical analysis. This grew out of a previous project concerning scheduling of complex time-dependent activities (DEXTER). In the past we applied worked on automated markets for complex transactions in a DARPA project investigating the use of markets to improve telecommunications network survivability in the face of disruptive attacks (MARX).
A reputation system gives people information about others' past performance. It can enhance an on-line interaction environment by:
- helping people decide who to trust;
- encouraging people to be more trustworthy;
- discouraging those who are not trustworthy from participating.
Del.icio.us is a website dedicated to 'social bookmarking.' We are studying how users make choices about which sites to bookmark, and which keywords to use to 'tag' these sites. We hope to develop an understanding of what incentives exist for users to use 'good' tags that help produce communities through the collaborative filtering capabilities of the inbox.
Digital Rights Management technologies are designed to increase the profitability of digital content such as music, movies, software, and books by decreasing piracy and increasing control over the uses of such content. We are investigating how (and if) these technologies have different effects on different types of goods, and the possible policy implications of such differences.
This stream of research incorporates bounded rationality and learning into the static mechanism design framework. In particular, the role of supermodularity in learning and convergence is examined.
We investigate the design of information goods. We focus on the configuration of the goods themselves (e.g., whether to bundle or sell separate items, and how much to combine information from different categories of interest), and on the pricing of the resulting information packages.
We are working on several alternatives to reduce spam and improve email service based on incentive-based strategies with markets:
- we proposed and analyzed an Attention-Bond Mechanism that allows email senders to signal the value that they believe their mail has, and allows users to receive selectively.
- we pointed out that by lowering the economic and information costs of switching between email service providers, we can stimulate competition to improve user-desired services such as strong spam filters.
- we offered a microeconomic model of the market for bulk commercial advertising email (the dominant form of spam) to develop a simple, feasible improvement to the current email system: an uncensored communication channel. Such a channel could be an email folder or account, to which properly tagged commercial solicitations are routed.
We are identifying barriers and developing alternative incentive mechanisms for data producers to deposit archive-ready data sets in an archive.
NEESGrid is a collection of large equipments that are used by the earthquake engineering community in the United States. It is funded by NSF and part of the mandate is for the resources to be shared in an efficient way. The goal of this research is to make recommendations on how to best allocate the resources using insights from the auction theory literature and results from the laboratory.
Computers have created many new security issues, particularly since their connection on the Internet. Many of these issues stem from choices made by people. My work involves applying incentive centered design to outstanding problems in information security: labelling of vulnerabilities, incentives for knowledge workers to follow security policies, and stopping botnets.
We study the efficient allocation of indivisible goods to individuals and propose computer-assisted allocation mechanisms superior to some used in the real world.
The open source phenonmenon seems to turn many traditional analyses across disciplines up side down. From this special case, we are trying to understand a broader class of phenonmenon in technology sharing and massive global collaboration, which we call open content. Our works so far include papers on open source, patent pools, standard setting organizations, interface and modularity. [Last Updated: Oct 24, 2006.]
We are developing an economic analysis of peer-to-peer file sharing systems, and designing incentive-aware mechanisms to improve social efficiency within such systems. Our initial goal is to characterize such systems in a general way that captures their most important features, and that allows us to analyze the degree of free-riding under different designs.
We are developing and pilot testing ride sharing services that dynamically match riders with rides. Our research focuses on how to motivate participation and reduce coordination costs in such services.
Past ICD Projects
In the early 1990s, economic analysis of Internet engineering came into being. ICD researchers at Michigan were pioneers in this area. We published a large body of work on the design of usage-based pricing to manage network congestion, multiple quality-of-service provisions, Web caching to reduce delays, and an economic approach to routing and address assignment.
Networked information technology, in particular the Internet, has created new opportunities for the exchange of digital information goods in an environment where the consumer population is changing and firms face a great deal of uncertainty. Firms must decide how much they should learn about the consumer population, given that this learning is costly. More broadly, we are interested in how agents in such an economy can form or locate niche markets, rather than competing for the mass market, and in identifying the conditions needed for niches to appeal to producers. This project involved joint work with researchers at IBM's T. J. Watson Labs, and generated about a dozen published papers.
The problem of achieving both information growth and information diffusion presents a difficult social tradeoff: policies that encourage investment in creating information frequently discourage widespread access. We applied this idea to address the question, "What software licensing schemes produce the greatest social welfare?"
We ask how information flows, access to technology, and social networks affect individual level output.
In the late 1990s, we undertook a 1.5 year experimental service for priced electronic journal delivery. This was the first wide-scale, large-volume commercial digital scholarly journal delivery service in the U.S. Our project was both a production service and a field experiment, and we conducted several studies of the results.
Package bidding is a common and difficult problem for electronic markets. We have developed theory for sealed bid auctions with ambiguity, and performed human-subject experiments on a number of auction designs.
Related SI Projects
AKKE: Conference on Advancing Knowledge and the Knowledge Economy
An international conference examined how the creation and organization of knowledge is changing and how this change affects the evolution and growth of the knowledge economy.
Cyberinfrastructure enables the creation, use, reuse, combination, organization, and sharing of knowledge within a virtually integrated environment of mixed resources. Cyberinfrastructure is constructed (using common applications and services) around specific projects, research communities, or unique resources. However, cyberinfrastructure can also support technology transfer, inter-sector collaboration, public education and participation, even commercial ventures. Full realization and optimization of cyberinfrastructure requires understanding organizational, economic, and legal context, including strategies and policies for advancing knowledge. As experience with the Internet demonstrates, choice of technology may influence business decisions and unbalance, or rebalance, laws and policies. Different forms of public and private ordering arise for addressing tensions between powerful enabling technologies and controls on access and use based on rights, secrecy, and security technologies. These ordering mechanisms, which can also be instantiated in cyberinfrastructure, may have further implications for knowledge-related strategies and policies.
to top