Bias and Social Aspects in Search and Recommendation: First International Workshop, BIAS 2020, Lisbon, Portugal, April 14, Proceedings (Communications in Computer and Information Science)

This book constitutes refereed proceedings of the First International Workshop on Algorithmic Bias in Search and Recommendation, BIAS 2020, held in April, 2020. Due to the COVID-19 pandemic BIAS 2020 was held virtually. The 10 full papers and 7 short papers were carefully reviewed and seleced from 44 submissions. The papers cover topics that go from search and recommendation in online dating, education, and social media, over the impact ofgender bias in word embeddings, to tools that allow to explore bias and fairnesson the Web. Preface Organization Contents Facets of Fairness in Search and Recommendation 1 Introduction 2 Dimensions of Search and Recommendation Results Evaluation 3 Fairness Metrics in Non-personalized Recommendation Settings 3.1 Accuracy-Based Fairness Metrics 3.2 Error Based Fairness Metrics 3.3 Causal Approach for Mitigating Discrimination 4 Fairness Metrics in Crowd-Sourced Non-personalized Recommendation Settings 5 Fairness Metrics in Personalized Recommendation Settings 6 Fairness Metrics in Advertisement Settings 7 Fairness Metrics in Marketplace Settings 8 Conclusion References Mitigating Gender Bias in Machine Learning Data Sets 1 Introduction 2 Related Work 2.1 Uncovering Gender Bias 3 Methods 4 Findings and Analysis 4.1 Presence of Women in Text 4.2 Gender-Specific Terms 4.3 Trends in Use of Androcentric Generics and Gender Neutrals 4.4 Gendered Associations: Negative or Stereotypical Descriptions 4.5 Gender and Emotion 4.6 Gendered Action 4.7 Character Descriptions and Gender 4.8 Gendered Associations with Family 4.9 Ordering of Binomials 5 Conclusion References Why Do We Need to Be Bots? What Prevents Society from Detecting Biases in Recommendation Systems 1 Introduction 2 Black Box Analyses 3 Case Study Facebook 3.1 Page Owner Perspective 3.2 Appropriate Forms of Audit 3.3 Broader Scope 4 Demands for a Legal Framework for Black Box Analyses References Effect of Debiasing on Information Retrieval 1 Introduction 2 Related Work 3 Method 3.1 Debiasing Word Embeddings 3.2 Retrieval Model and Experimental Setup 4 Results 5 Conclusion References Matchmaking Under Fairness Constraints: A Speed Dating Case Study 1 Introduction 2 Case Study 2.1 Matchmaking 3 Related Work 4 Preferential Fairness 4.1 Background 4.2 Model 5 Re-ranking Methods 5.1 Knapsack 5.2 Tabu Search 6 Experimental Results 6.1 Racial Bias 6.2 Religious Bias 6.3 Discussion 7 Conclusion References Recommendation Filtering à la carte for Intelligent Tutoring Systems 1 Introduction 2 Background 2.1 Clustering 2.2 Item Response Theory 3 Methodology 3.1 Recommendation Process 3.2 Clustering Phase 4 Conclusion References bias goggles: Exploring the Bias of Web Domains Through the Eyes of Users 1 Introduction 2 Implementation Discussion 2.1 Back-End 2.2 Front-Ends 3 Crawled Data 4 Performance Discussion 5 Future Work References Data Pipelines for Personalized Exploration of Rated Datasets 1 Introduction 2 Data Model 2.1 Rated Datasets and Labeled Segments 2.2 Data Pipelines 3 Data Pipelines Implementation 4 Empirical Validation and Discussion 4.1 Validation 4.2 Discussion References Beyond Accuracy in Link Prediction 1 Introduction 2 Related Work 3 Notation 4 Social Network Analysis 4.1 Distance-Based Metrics 4.2 Structural Diversity 5 Novelty and Diversity 5.1 Novelty 5.2 Diversity 6 Empirical Observation 6.1 Data 6.2 Link Prediction Algorithms 6.3 Results 7 Conclusions References A Novel Similarity Measure for Group Recommender Systems with Optimal Time Complexity 1 Introduction 2 Related Work 3 Preliminaries and Notation 4 The Kolmogorov-Based Similarity 5 Complexity Analysis 6 The Group Recommender System 7 Experimental Setup 7.1 Evaluation Measure 7.2 Experimental Results 8 Conclusions References What Kind of Content Are You Prone to Tweet? Multi-topic Preference Model for Tweeters 1 Introduction 2 Related Work 2.1 Tweeters Modeling for Recommendation 2.2 Groups Formation and Recommendations 2.3 Tweets Classification 3 Approach 3.1 Tweets Modeling 3.2 Extraction of the Suitable Number of Topics 3.3 Tweets Classification: The EM Algorithm Applied over Tweets 3.4 Twitter Users Model: Extraction of the MUM 3.5 Grouping Like-Minded Users 4 Experimental Framework 4.1 Data Collection 4.2 Baseline Approach 4.3 Experimental Setup and Strategy 4.4 Validation of Results 5 Conclusions and Future Work References Venue Suggestion Using Social-Centric Scores 1 Introduction 2 Related Work 3 Venue Suggestion 3.1 Frequency-Based Score 3.2 Review-Based Score 3.3 Location Ranking 4 Experiments 4.1 Experimental Setup 4.2 Results and Discussions 5 Conclusions and Future Work References The Impact of Foursquare Checkins on Users' Emotions on Twitter 1 Introduction 2 Proposed Approach 2.1 Problem Definition 2.2 Metric Definition 2.3 Methodology 3 Experiments 3.1 Dataset Description and Experimental Setup 3.2 Study Findings References Improving News Personalization Through Search Logs 1 Introduction 2 Search-Enhanced News Personalization 3 Experiments 3.1 Results 4 Related Work 5 Conclusions References Analyzing the Interaction of Users with News Articles to Create Personalization Services 1 Introduction 2 Related Work 3 Methodology 3.1 Analyzed Aspects 3.2 Description of the Used Dataset 4 Results of the Performed Analyses 4.1 Demographic Analysis 4.2 Time Sensitive Users' Interests Analysis 4.3 Semantic Analysis 5 Conclusions and Future Work References Using String-Comparison Measures to Improve and Evaluate Collaborative Filtering Recommender Systems 1 Introduction 2 Basic Concepts 2.1 Recommender Systems Based on Collaborative Filtering 2.2 String-Based Similarity Measures 3 Proposed Recommender System Architecture 3.1 System Overview 3.2 Methods to Determine the Similarity Function 4 Experiments 4.1 Setup 4.2 Results 5 Conclusion and Future Work References Enriching Product Catalogs with User Opinions 1 Introduction 2 OpinionLink: Overview 3 Experiments Results 3.1 Identifying Direct Opinionated Sentences 3.2 Opinion Mapping 3.3 End-to-End Results 3.4 OpinionLink in Large Scale 4 Conclusions References Correction to: bias goggles: Exploring the Bias of Web Domains Through the Eyes of Users Correction to: Chapter “bias goggles: Exploring the Bias of Web Domains Through the Eyes of Users” in: L. Boratto et al. (Eds.): Bias and Social Aspects in Search and Recommendation, CCIS 1245, https://doi.org/10.1007/978-3-030-52485-2_7 Author Index