scholarly journals Bayesian Exploration: Incentivizing Exploration in Bayesian Games

2021 ◽  
Author(s):  
Yishay Mansour ◽  
Alex Slivkins ◽  
Vasilis Syrgkanis ◽  
Zhiwei Steven Wu

In a wide range of recommendation systems, self-interested individuals (“agents”) make decisions over time, using information revealed by other agents in the past, and producing information that may help agents in the future. Each agent would like to exploit the best action given the current information but would prefer the previous agents to explore various alternatives to collect information. A social planner, by means of a well-designed recommendation policy, can incentivize the agents to balance exploration and exploitation in order to maximize social welfare or some other objective. The recommendation policy can be modeled as a multiarmed bandit algorithm under Bayesian incentivecompatibility (BIC) constraints. This line of work has received considerable attention in the “economics and computation” community. Although in prior work, the planner interacts with a single agent at a time, the present paper allows the agents to affect one another directly in a shared environment. The agents now face two sources of uncertainty: what is the environment, and what would the other agents do? We focus on “explorable” actions: those that can be recommended by some BIC policy. We show how the principal can identify and explore all such actions.

2020 ◽  
Vol 68 (4) ◽  
pp. 1132-1161
Author(s):  
Yishay Mansour ◽  
Aleksandrs Slivkins ◽  
Vasilis Syrgkanis

As self-interested individuals (“agents”) make decisions over time, they utilize information revealed by other agents in the past and produce information that may help agents in the future. This phenomenon is common in a wide range of scenarios in the Internet economy, as well as in medical decisions. Each agent would like to exploit: select the best action given the current information, but would prefer the previous agents to explore: try out various alternatives to collect information. A social planner, by means of a carefully designed recommendation policy, can incentivize the agents to balance the exploration and exploitation so as to maximize social welfare. We model the planner’s recommendation policy as a multiarm bandit algorithm under incentive-compatibility constraints induced by agents’ Bayesian priors. We design a bandit algorithm which is incentive-compatible and has asymptotically optimal performance, as expressed by regret. Further, we provide a black-box reduction from an arbitrary multiarm bandit algorithm to an incentive-compatible one, with only a constant multiplicative increase in regret. This reduction works for very general bandit settings that incorporate contexts and arbitrary partial feedback.


Anticorruption in History is the first major collection of case studies on how past societies and polities, in and beyond Europe, defined legitimate power in terms of fighting corruption and designed specific mechanisms to pursue that agenda. It is a timely book: corruption is widely seen today as a major problem, undermining trust in government, financial institutions, economic efficiency, the principle of equality before the law and human wellbeing in general. Corruption, in short, is a major hurdle on the “path to Denmark”—a feted blueprint for stable and successful statebuilding. The resonance of this view explains why efforts to promote anticorruption policies have proliferated in recent years. But while the subjects of corruption and anticorruption have captured the attention of politicians, scholars, NGOs and the global media, scant attention has been paid to the link between corruption and the change of anticorruption policies over time and place. Such a historical approach could help explain major moments of change in the past as well as reasons for the success and failure of specific anticorruption policies and their relation to a country’s image (of itself or as construed from outside) as being more or less corrupt. It is precisely this scholarly lacuna that the present volume intends to begin to fill. A wide range of historical contexts are addressed, ranging from the ancient to the modern period, with specific insights for policy makers offered throughout.


2008 ◽  
Vol 15 ◽  
Author(s):  
SueAnne Ware

Andreas Huyssen writes, ‘Remembrance as a vital human activity shapes our links to the past, and the ways we remember define us in the present. As individuals and societies, we need the past to construct and to anchor our identities and to nurture a vision of the future.’ Memory is continually affected by a complex spectrum of states such as forgetting, denial, repression, trauma, recounting and reconsidering, stimulated by equally complex changes in context and changes over time. The apprehension and reflective comprehension of landscape is similarly beset by such complexities. Just as the nature and qualities of memory comprise inherently fading, shifting and fleeting impressions of things which are themselves ever-changing, an understanding of a landscape, as well as the landscape itself, is a constantly evolving, emerging response to both immense and intimate influences. There is an incongruity between the inherent changeability of both landscapes and memories, and the conventional, formal strategies of commemoration that typify the constructed landscape memorial. The design work presented in this paper brings together such explorations of memory and landscape by examining the ‘memorial’. This article examines two projects. One concerns the fate of illegal refugees travelling to Australia: The SIEVX Memorial Project. The other, An Anti-Memorial to Heroin Overdose Victims, was designed by the author as part of the 2001 Melbourne Festival.


Author(s):  
George E. Mitchell ◽  
Hans Peter Schmitz ◽  
Tosca Bruno-van Vijfeijken

Chapter 5 explores how the foundations for TNGO legitimacy have changed over time, creating imperatives for TNGOs to invest in new capabilities and adopt new practices. In the past, TNGOs derived legitimacy from their espoused principles, representational claims, elite expertise, demonstrated financial stewardship, commitment to charity, and patterns of conformity. More recently, TNGOs themselves have helped to bring about a shift toward new bases for legitimacy that focus on effectiveness, strategy, leadership, governance, transparency, and responsiveness. However, transitioning to the legitimacy practices of the future is complicated by the persistence of an antiquated architecture that still demands that TNGO conform to legacy expectations. Nevertheless, new approaches to enhancing legitimacy provide a wide range of opportunities that invite organizations to proactively align their aspirations with emerging stakeholder expectations.


2018 ◽  
Vol 29 (1) ◽  
pp. 92-110
Author(s):  
Anja Danner-Schröder

This article examines how events from the past, present, and future form into event structures over time. This question is addressed by investigating the Great East Japan Earthquake in 2011 until the fifth anniversary in 2016. This allowed to analyze different events over time. The findings reveal that events can be used in two different ways. One process was meant to focus on events, whereas the other one backgrounded events. These different ways to use events revealed four different mechanisms of how event structures can be formed. Moreover, each mechanism has its own idiosyncratic temporal orientation toward either a nostalgic past, imagined future, “better” future or critical past. Second, the article contributes that the paradoxical ways of focusing on an event and backgrounding the very same event need to be embraced simultaneously to enable a greater sense of wholeness. Last, the article reveals multiple temporalities within and across temporal trajectories.


2003 ◽  
Vol 21 (2) ◽  
pp. 367-376 ◽  
Author(s):  
Piotr Górecki

Susan Reynolds's article is a culmination and a turning point. It builds on several approaches to medieval law and culture, of which two strike me as especially important. One is a study of legal history as a domain of human activity, especially habitual or routine activity, pursued by a wide range of social groups. The other is a search for the meaning and the criteria of the enormous transition during the central Middle Ages, which Christopher Dawson at the dawn of this subject, and Robert Bartlett in its currently definitive moment, have identified as “the making of Europe.” The first subject exists above all thanks to the work of Reynolds herself, while the second is an outcome of a number of quite distinct scholarly trajectories, spanning several generations. Apart from some suggestive and implicit links, those two subjects have, over the past quarter century, been pursued separately. Reynolds's article brings them together.


Legal Studies ◽  
1992 ◽  
Vol 12 (1) ◽  
pp. 1-19 ◽  
Author(s):  
A. I. Ogus

Regulation as a legal form of social engineering has been subjected to much analysis in the last decade or so. The importance of the topic to contemporary law cannot be overstated: on the one hand, it has been the avowed aim of government to ‘deregulate’ industry; on the other hand, and paradoxically, both the concomitant policy of privatisation and the evolution towards a Single European Market have increased the need for regulation in appropriate areas. The efforts to explore the strengths and weaknesses of different regulatory forms have brought together scholars from a wide range of disciplines. Administrative lawyers have been concerned with how the power of decision-making is allocated between institutions and the general problems of accountability and control of discretion to which this gives rise. Socio-legal researchers have critically examined the practices of regulatory agencies as regards rule formulation and enforcement.


2020 ◽  
Vol 10 (21) ◽  
pp. 7393
Author(s):  
Juan F. Carías ◽  
Saioa Arrizabalaga ◽  
Leire Labaka ◽  
Josune Hernantes

Due to the hazardous current cyber environment, cyber resilience is more necessary than ever. Companies are exposed to an often-ignored risk of suffering a cyber incident. This places cyber incidents as one of the main risks for companies in the past few years. On the other hand, the literature meant to aid on the operationalization of cyber resilience is mostly focused on listing the policies required to operationalize it, but is often lacking on how to prioritize these actions and how to strategize their implementation. Therefore, the usage of the current literature in this state is not optimal for companies. Thus, this study proposes a progression model to help companies strategize and prioritize cyber resilience policies by proposing the natural evolution of the policies over time. To develop the model, this study used semi-structured interviews and an analysis of the data obtained from the interviews. Through this methodology, this study found the starting points for each cyber resilience policy and their natural progression over time. These results can help companies in their cyber resilience building process by giving them insights on how to strategize the implementation of the cyber resilience policies.


Author(s):  
Emily W. B. Russell Southgate

There are many field techniques that take research beyond what can be found in written documents. Clues to the past are hidden in such subtle (and not so subtle) features as topographic modifications, soils, and tree trunks. Field studies search for evidence of conditions and for their resultant effects through a variety of techniques. Features may be evident from a ground survey, located precisely through GPS. Remote imaging, using a variety of techniques from simple aerial photography to lidar, reveal hidden patterns, that may then be studied on the ground. Archeological study relates the patterns found on the ground to human activities, as they have changed over time. Dendroecology interprets records left in tree rings. Some studies look at stands of different ages to study change over time, and long-term studies trace change in individual stands. In these field studies that look at the past, soil characteristics and materials hidden in the soil, such as DNA, stable isotopes and charcoal, can reveal details of past processes and species. These methods are illustrated with examples from a wide range of biomes.


1997 ◽  
Vol 60 (3) ◽  
pp. 134-138 ◽  
Author(s):  
Tricia Jacobs

Much has been written in the past 15 years by medical educators on problem-based learning (PBL) in medical schools. The other health professions, such as occupational therapy, social work and nursing, have been slow in taking up the challenge and developing PBL programmes to suit their professional needs. The purpose of this article is to summarise previous developments and offer a schema for the preparation of curricula in occupational therapy. No schema for PBL has previously been published, so it is proposed that this attempt will be able to be adapted to a wide range of courses using the acknowledged educational concepts implicit in integrated, problem-based programmes.


Sign in / Sign up

Export Citation Format

Share Document