Many data sharing systems are open to arbitrary users on the Internet, who are independent and self-interested agents. Therefore, in addition to traditional design goals such as technical performance, data sharing systems should be designed to best support the strategic interactions of these agents.Our research hypothesis is that designs that maximize the participants' autonomy can produce useful data sharing systems.We apply this design principle to both the system architecture and the functional design of a data sharing system, and study the resulting class of systems, which we call Decentralized Social Data Sharing ((DS) 2 ) systems.We formally define this class of systems and provide a reference implementation and an example application: a distributed wiki system called P2Pedia. P2Pedia implements a decentralized collaboration model, where the users are not required to reach a consensus, and instead benefit from being exposed to multiple viewpoints. We demonstrate the value of this collaboration model through an extensive user study.Allowing the users to autonomously control their data prevents the system architecture from being optimized for efficient query processing. We show that Regular Path Queries, a useful class of graph queries, can still be processed on the shared data:although in the worst case such queries are intractable, we propose a cost estimation technique to identify tractable queries from partial knowledge of the data.Through simulation, we also show that the users' control over network connections ii allows them to self-organize and interact with other users with whom their interests are best aligned. This may result in less data being available, and we study cases where this is in fact demonstrably beneficial to the users, as the available data to each user is the most relevant to them.This suggests that querying this reduced collection of shared data may lead to more tractable query processing without necessarily reducing the users' utility.
Abstract. Case-base reasoning in a real-time context requires the system to output the solution to a given problem in a predictable and usually very fast time frame. As the number of cases that can be processed is limited by the real-time constraint, we explore ways of selecting the most important cases and ways of speeding up case comparisons by optimizing the representation of each case. We focus on spatially-aware systems such as mobile robotic applications and the particular challenges in representing the systems' spatial environment. We select and combine techniques for feature selection, clustering and prototyping that are applicable in this particular context and report results from a case study with a simulated RoboCup soccer-playing agent. Our results demonstrate that preprocessing such case bases can significantly improve the imitative ability of an agent.
In the context of the RoboCup Simulation League, we describe a new representation of a software agent's visual perception ("scene"), well suited for case-based reasoning. Most existing representations use either heterogeneous, manually selected features of the scene, or the raw list of visible objects, and use ad hoc similarity measures for CBR. Our representation is based on histograms of objects over a partition of the scene space. This method transforms a list of objects into an image-like representation with customizable granularity, and uses fuzzy logic to smoothen boundary effects of the partition. We also introduce a new similarity metric based on the Jaccard Coefficient, to compare scenes represented by such histograms. We present our implementation of this approach in a case-based reasoning project, and experimental results showing highly efficient scene comparison.
Purpose The purpose of this paper is to report on empirical work conducted to open up algorithmic interpretability and transparency. In recent years, significant concerns have arisen regarding the increasing pervasiveness of algorithms and the impact of automated decision-making in our lives. Particularly problematic is the lack of transparency surrounding the development of these algorithmic systems and their use. It is often suggested that to make algorithms more fair, they should be made more transparent, but exactly how this can be achieved remains unclear. Design/methodology/approach An empirical study was conducted to begin unpacking issues around algorithmic interpretability and transparency. The study involved discussion-based experiments centred around a limited resource allocation scenario which required participants to select their most and least preferred algorithms in a particular context. In addition to collecting quantitative data about preferences, qualitative data captured participants’ expressed reasoning behind their selections. Findings Even when provided with the same information about the scenario, participants made different algorithm preference selections and rationalised their selections differently. The study results revealed diversity in participant responses but consistency in the emphasis they placed on normative concerns and the importance of context when accounting for their selections. The issues raised by participants as important to their selections resonate closely with values that have come to the fore in current debates over algorithm prevalence. Originality/value This work developed a novel empirical approach that demonstrates the value in pursuing algorithmic interpretability and transparency while also highlighting the complexities surrounding their accomplishment.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.