Projects

GroupFun

Title: Group recommender systems as a voting problem
Year: 2010-2012
Researchers: George Popescu and Pearl Pu
Keywords: Voting, Recommender systems, Group decision making, Social choice, Social websites, Dynamic selection, Preference aggregation, User incentives, Game Theory, Music preference, online user decision and behavior, social websites
Abstract: Aggregating users’ individual preference and recommending a common set of items for a group has become a challenging topic in group recommender systems and social websites. This issue is mainly concerned with the following three objectives: eliciting individual users’ preferences, suggesting outcomes that maximize the overall satisfaction for all users and ensuring that the aggregation mechanism is resistant to individual users’ manipulation. Firstly we show how our proposed probabilistic weighted-sum algorithm (PWS) works and emphasize on its advantages. Then we compare PWS with related approaches implemented in similar systems using the case of our music recommender, GroupFun. We describe an experiment design to study users’ perceptions of the algorithms, their perceived fairness and incentives to manipulate the final recommendation outcome. We expect our results to show that PWS will be perceived as fair and diversity- and discovery-driven, thus enhancing the group’s satisfaction. Our future work will focus on the actual evaluation of GroupFun using the experiment design presented here.
Sponsor: Swiss National Science Foundation
Links: Main page, Application

An Information Processing Model and Interaction Principles for Decision Tradeoff

Title: An Information Processing Model and Interaction Principles for Decision Tradeoff
Dates: 2004 – 2008
Researchers: Jiyong Zhang and Pearl Pu
Keywords: information processing model, interaction principles, tradeoff
Abstract: CritiqueShop is an online platform for designing and evaluating e-commerce product search tools based on the critiquing technique. It provides a unified user interface so that the performances of different recommendation algorithms can be evaluated under the same condition. This system is developed with Java/AJAX (see Google web toolkit for the detail of this technology). For more information about this system, please see our publications. Click here for more.
Sponsor: Swiss National Science Foundation
Links: Main pageApplication, image gallery

Preference Elicitation

Title: Preference Elicitation
Dates: 2003 – 2008
Researchers: Li Chen and Pearl Pu
Keywords: preference model, users’ beliefs, cognitive and emotional limitations
Abstract: As people increasingly rely on interactive decision support systems to choose products and make decisions about tradeoffs, we must understand how interface technologies influence user behaviors. Crucial to this issue is to develop the mathematical decision theory into a new theory of decision information processing model, which shows how to present the most meaningful information in relation to users’ task goals and supporting context, and how to adaptively change the information content when goals evolve. Building effective interfaces for interactive decision support systems is challenging because 1) users’ preference models are incomplete and it is hard to elicit value functions for preferences that do not exist; 2) users’ beliefs about desirability are ephemeral, uncertain and context dependent; and 3) users have cognitive and emotional limitations for processing information. To develop a useful theory of decision information processing, we need to further understand the process by which humans make tradeoff decisions, how information affects this process, and how to construct effective interface technologies to augment user performance.
Sponsor: Swiss National Science Foundation
Links: Home page

Interaction Models for Preference Elicitation

Title: Interaction Models for Preference Elicitation
Researchers: Pratyush Kumar and Pearl Pu
Keywords: focus + context, information visualization
Abstract: Human centered decision making principles have been the center of massive research activities in the recent past. While on one hand traditional decision theory claims to predict exact solution for a decision problem, provided a value function is defined. A value function is one which can predict accurately about decision values given precise value of the attributes. Researchers have long found the problem of defining a precise value function tough in real life scenarios. The problem becomes even more complicated when the decision task is a multi attribute function. A simple example of such a task is the selection of a multi attribute product. Our work centers around selection of a multi attribute product on an e-store. Many tools are currently available which help users in making the crucial decision. The major impediment with such tools is the fact that under uncertainty in preferences, they are only as good in helping the users in their decision task, as users themselves are in their preferences. Experiments done with users have suggested the fact that often preferences with users are fluid and uncertain, hence a proper method has to be designed to elicit those unformed and often unspecified preferences. It is these preferences that lead to a definite decision goal. Under the scenario of uncertain and unspecified preferences, the task of defining a precise value function also becomes highly complex and difficult. We believe that such hidden and unstated preferences can be elicited by providing users with example. This process is called example critiquing. The users are shown examples of the products they are looking for, and then they discover their preferences when attributes of the example violated any of their hidden preferences. The idea is that users do not know their preferences till they see them being violated. Thus, we believe that such example critiquing can be crucial in eliciting preferences of users and help them in better decision making. 
Sponsor: Swiss National Science Foundation
Links: Main page; Demo 1: RankedListUI; Demo 2: TweakingListUI

Human Factors and Interface Design

Title: Human Factors and Interface Design
Researchers: Gregory Delajoux and Pearl Pu
Keywords: human factors, interface design
Date: 2005
Sponsor: EPFL

Semantic Fisheye View

Title: Semantic Fisheye View
Dates: 2000 – 2004
Researchers: Paul Janecek and Pearl Pu
Keywords: focus + context, information visualization
Abstract: This research project is developing a context-aware multimedia information workspace based on Semantic Fisheye Views (SFEVs) to search through, browse, organize, and synthesize large personal collections of multimedia documents. This work focuses especially on annotated images, although the results are generalizable to other multimedia document types that are managed by their metadata. 
Sponsor: Swiss National Science Foundation
Links: Home page

Social Navigation

Title: Semantic Fisheye View
Dates: 2002 – 2003
Researchers: Punit Gupta and  Pearl Pu
Keywords: Social Navigation
Abstract: This project is on social navigation system, which involves investigation of the visual metaphors for social navigation in the digital space of scientific literature.In our case, social information is in the form of number of citations to a paper, year of publication, number of downloads, reputation of the research lab and the authors. Our project aims at using different information visualisation techniques to visually model social information and assist users to locate the resourses easily. We are looking for visual metaphors to render the social information about the papers which will help the users in interacting with the system in a better manner.
Sponsor: Swiss National Science Foundation
Links: Image gallery

Multimedia information retrieval

Title: An Information Processing Model and Interaction Principles for Decision Tradeoff
Dates: 1998 – 2003
Researchers: Zoran Pecenovic and Pearl Pu
Keywords: content-based image retrieval, active query, similarity analysis
Abstract: Dynamic Overview Techniques for Image Retrieval
Sponsor: Swiss National Science Foundation
Links: Home Page;  Image gallery

Scalable Intelligent Electronic Catalogs

Title: Scalable Intelligent Electronic Catalogs
Dates: 1998 – 2001
Researchers: Adriana Maria Jurca and Pearl Pu
Keywords: e-commerce, h/m interaction, visualization, artificial intelligence, planning systems
Abstract: In traditional commerce, merchants compete not only on price, but also on the services offerred to their customers. Good salespeople understand their customer’s needs and preferences, and are skilled at matching them to the products. In electronic commerce, such functions have to be carried out by the interaction model underlying the store front. Providing such an interaction model with catalog intelligence can provide more rich user experience, for example by offering: * personalized search functionalities for finding products according to complex criteria; * matching techniques that help the customer formulate his wishes by comparing them with those of others; * and configuration functions that compose products out of several parts, and filtering and visualizing tools to help users making purchase decisions.
Sponsor: Swiss National Science Foundation
Links: Home page Image gallery

Visualization and conceptualization for resource allocation

Title: Visualization and conceptualization for resource allocation
Dates: 1995 – 2000
Researchers: George Melissargos and Pearl Pu
Keywords: information visualization, human-computer interaction, constraint-based search
Abstract: This project was a collaboration of our group and the AI Lab at the Computer Science Department. The main target was to build an aircraft-to-flights allocation system for Swissair, Switzerland’s air carrier.
Sponsor: Swiss National Science Foundation
Links: Home page Image gallery