Advances in Bias and Fairness in Information Retrieval: Second International Workshop on Algorithmic Bias in Search and Recommendation, BIAS 2021, ... in Computer and Information Science)

Advances in Bias and Fairness in Information Retrieval: Preface
Organization
Contents
Towards Fairness-Aware Ranking by Defining Latent Groups Using Inferred Features
	1 Introduction
	2 Data Description
	3 Methodology
		3.1 Group Definition
		3.2 Pure Relevance with BM25
		3.3 Fairness-Aware Re-ranking Algorithm
	4 Results and Discussion
		4.1 Baselines
		4.2 Weighted Fairness-Aware Re-ranking Results
	5 Limitations and Future Work
	6 Conclusion
	References
Media Bias Everywhere? A Vision for Dealing with the Manipulation of Public Opinion
	1 Introduction
	2 Field Analysis and Recommended Actions
		2.1 Field of Action 1: Evaluation Principles of Media Bias
		2.2 Field of Action 2: Information Presentation of Media Bias
		2.3 Field of Action 3: Transparency of Media Bias Evaluation
	3 Conclusion
	References
Users\' Perception of Search-Engine Biases and Satisfaction
	1 Introduction
	2 Background
	3 Method
	4 Results
	5 Discussion
	6 Conclusion
	References
Preliminary Experiments to Examine the Stability of Bias-Aware Techniques
	1 Introduction
	2 Datasets
		2.1 Generative Models for the Datasets
		2.2 Data Collection Procedure
		2.3 Evaluation of Cognitive Bias
	3 The Stability of Bias-Aware Techniques
		3.1 Causal Inference from the Viewpoint of Bias-Aware Classification
		3.2 Experiments to Examine the Stability of a Bias-Aware Technique
		3.3 Next Steps for Collecting Better Datasets
	4 Conclusion
	References
Detecting Race and Gender Bias in Visual Representation of AI on Web Search Engines
	1 Introduction
	2 Related Work: Race and Gender Bias in Image Search
	3 Case Study: Racial and Gender Representation of AI
	4 Methodology
	5 Findings
		5.1 AI and Antropomorphism
		5.2 AI and Race
		5.3 AI and Gender
	6 Discussion
	References
Equality of Opportunity in Ranking: A Fair-Distributive Model
	1 Introduction
	2 Related Work
	3 Problem Statement
	4 A Fair-Distributive Ranking Model
		4.1 Ranked Type Fairness
		4.2 Circumstances-Effort Based Ranking and Counterfactual-Ordered Ranking Utility
	5 Experiment
		5.1 Data and Settings
		5.2 Metrics
		5.3 Results and Discussion
	6 Conclusions
	References
Incentives for Item Duplication Under Fair Ranking Policies
	1 Introduction
	2 Related Work
		2.1 Fairness in Ranking
		2.2 Diversity in Ranking
	3 Duplicates and Fair Ranking
		3.1 Rewarding Duplicates
		3.2 Experimental Setup
		3.3 Results
	4 Conclusions
	References
Quantification of the Impact of Popularity Bias in Multi-stakeholder and Time-Aware Environments
	1 Introduction
	2 Related Work
		2.1 Multistakeholders
		2.2 Time-Aware Recommendations
		2.3 Bias and Unfairness
	3 Datasets
	4 Proposed Method
		4.1 Time Popularity Metrics
		4.2 Dynamic Grouping of Items and Stakeholders
		4.3 Measuring the Unfairness of Stakeholder Recommendations
	5 Experimented Recommender Systems
	6 Results
	7 Conclusions and Future Work
	References
When Is a Recommendation Model Wrong? A Model-Agnostic Tree-Based Approach to Detecting Biases in Recommendations
	1 Introduction
	2 Background and Related Work
	3 Problem Definition
	4 Proposed Approach
	5 Experimental Validation
		5.1 Simulating Biased Recommendations
		5.2 Detecting Biases for Collaborative Filtering Algorithms
	6 Conclusions and Future Work
	References
Examining Video Recommendation Bias on YouTube
	1 Introduction
	2 Related Work
		2.1 YouTube Algorithm, Radicalization, Misinformation
		2.2 Network Analysis Approaches and Topology of YouTube Recommendations
		2.3 Bias in Search and Recommendation
	3 Data and Methods
		3.1 Data Collection and Seed Lists
		3.2 Bias Evaluation Through Graph Topology and PageRank Distributions
	4 Experiment Results
	5 Conclusion
	References
An Information-Theoretic Measure for Enabling Category Exemptions with an Application to Filter Bubbles
	1 Introduction
	2 Related Work
	3 Problem Setup and Main Results
		3.1 Thought Experiments to Understand When to Use VCMI, CMI or MI
	4 Experimental Evaluation
		4.1 Study on Synthetic Data (Toy Example)
		4.2 Case Study on a Dataset Created from Real Data
	5 Conclusion and Discussions
	References
Perception-Aware Bias Detection for Query Suggestions
	1 Introduction
	2 Related Work
	3 Methodology
	4 Results of Bias Analysis of German Politicians Dataset
	5 Discussion
	6 Conclusion
	References
Crucial Challenges in Large-Scale Black Box Analyses
	1 Introduction
	2 Black Box Analysis
		2.1 Limits of a Black Box Analysis
	3 A Case Study: Eurostemcell Data Donation
		3.1 Study Design and Results
	4 Challenges in Conducting a Black Box Analysis
		4.1 Challenges in a Crowd-Sourced Approach
		4.2 Challenges in a Bot-Based Approach
		4.3 Arguments for Including a Pre-study
	5 From Experimental Studies to Establishing Accountability with the Help of Large-Scale Black Box Analyses
	6 Summary
	References
New Performance Metrics for Offline Content-Based TV Recommender System
	1 Introduction
	2 Content-Based TV Recommendation System
		2.1 Architecture
		2.2 Input Datasets
		2.3 Pre-processing
		2.4 Feature Engineering
		2.5 Recommendation Engine
	3 Evaluating the Recommendation System
		3.1 Traditional Metrics
		3.2 Proposed New Metrics
	4 Results Analysis and Concluding Remarks
		4.1 Conclusion
		4.2 Future Work
	References
Author Index