Format:
1 Online-Ressource (xv, 135 Seiten)
,
Illustrationen
ISBN:
1627056084
,
9781627056083
Series Statement:
Synthesis lectures on artificial intelligence and machine learning #35
Content:
Intelligent systems often depend on data provided by information agents, for example, sensor data or crowdsourced human computation. Providing accurate and relevant data requires costly effort that agents may not always be willing to provide. Thus, it becomes important not only to verify the correctness of data, but also to provide incentives so that agents that provide high-quality data are rewarded while those that do not are discouraged by low rewards. We cover different settings and the assumptions they admit, including sensing, human computation, peer grading, reviews, and predictions. We survey different incentive mechanisms, including proper scoring rules, prediction markets and peer prediction, Bayesian Truth Serum, Peer Truth Serum, Correlated Agreement, and the settings where each of them would be suitable. As an alternative, we also consider reputation mechanisms. We complement the game-theoretic analysis with practical examples of applications in prediction platforms, community sensing, and peer grading
Content:
1. Introduction -- 1.1 Motivation -- 1.1.1 Example: product reviews -- 1.1.2 Example: forecasting polls -- 1.1.3 Example: community sensing -- 1.1.4 Example: crowdwork -- 1.2 Quality control -- 1.3 Setting --
Content:
2. Mechanisms for verifiable information -- 2.1 Eliciting a value -- 2.2 Eliciting distributions: proper scoring rules --
Content:
3. Parametric mechanisms for unverifiable information -- 3.1 Peer consistency for objective information -- 3.1.1 Output agreement -- 3.1.2 Game-theoretic analysis -- 3.2 Peer consistency for subjective information -- 3.2.1 Peer prediction method -- 3.2.2 Improving peer prediction through automated mechanism design -- 3.2.3 Geometric characterization of peer prediction mechanisms -- 3.3 Common prior mechanisms -- 3.3.1 Shadowing mechanisms -- 3.3.2 Peer truth serum -- 3.4 Applications -- 3.4.1 Peer prediction for self-monitoring -- 3.4.2 Peer truth serum applied to community sensing -- 3.4.3 Peer truth serum in Swissnoise -- 3.4.4 Human computation --
Content:
4. Nonparametric mechanisms: multiple reports -- 4.1 Bayesian truth serum -- 4.2 Robust Bayesian truth serum -- 4.3 Divergence-based BTS -- 4.4 Two-stage mechanisms -- 4.5 Applications --
Content:
5. Nonparametric mechanisms: multiple tasks -- 5.1 Correlated agreement -- 5.2 Peer truth serum for crowdsourcing (PTSC) -- 5.3 Logarithmic peer truth serum -- 5.4 Other mechanisms -- 5.5 Applications -- 5.5.1 Peer grading: course quizzes -- 5.5.2 Community sensing --
Content:
6. Prediction markets: combining elicitation and aggregation --
Content:
7. Agents motivated by influence -- 7.1 Influence limiter: use of ground truth -- 7.2 Strategyproof mechanisms when the ground truth is not accessible --
Content:
8. Decentralized machine learning -- 8.1 Managing the information agents -- 8.2 From incentives to payments -- 8.3 Integration with machine learning algorithms -- 8.3.1 Myopic influence -- 8.3.2 Bayesian aggregation into a histogram -- 8.3.3 Interpolation by a model -- 8.3.4 Learning a classifier -- 8.3.5 Privacy protection -- 8.3.6 Restrictions on agent behavior --
Content:
9. Conclusions -- 9.1 Incentives for quality -- 9.2 Classifying peer consistency mechanisms -- 9.3 Information aggregation -- 9.4 Future work --
Content:
Bibliography -- Authors' biographies
Note:
Includes bibliographical references (pages 127-133)
Additional Edition:
ISBN 9781627057295
Additional Edition:
Print version ISBN 9781627057295
Language:
English
Keywords:
Electronic books
Bookmarklink