Experimental IR Meets Multilinguality, Multimodality, and Interaction. 13th International Conference of the CLEF Association, CLEF 2022 Bologna, Italy, September 5–8, 2022 Proceedings

Preface
Organization
Contents
Full Papers
``Meanspo Please, I Want to Lose Weight'': A Characterization Study of Meanspiration Content on Tumblr Based on Images and Texts
	1 Introduction
	2 Related Work
	3 Research Methodology
		3.1 Data Collection
		3.2 Image Analysis
		3.3 Text Analysis
	4 Results
		4.1 RQ1: Visual and Textual Characteristics
		4.2 RQ2: Classification
	5 Discussion
	6 Conclusion
	References
Tracking News Stories in Short Messages in the Era of Infodemic
	1 Introduction
	2 Related Work
	3 Datasets
		3.1 Represent Document with Multiple Features
	4 Tracking Documents Reporting the Same Stories
		4.1 Streaming Algorithm to Build News Stories
		4.2 Unsupervised News Tracking with K-Means
	5 Experiments
	6 The Issue of Short Message Similarities
	7 Discussion
	8 Conclusion
	References
Leveraging Wikipedia Knowledge for Distant Supervision in Medical Concept Normalization
	1 Introduction
	2 Related Work
	3 Distant Supervision Approach
		3.1 Data Description
		3.2 Identifying Medical Articles from Wikipedia
		3.3 Mapping Wikipedia Articles to SNOMED-CT
		3.4 Lay Medical Terms Extraction
	4 Experiment Setup
	5 Results
	6 Discussion
	7 Conclusion and Future Work
	References
Noise-Reduction for Automatically Transferred Relevance Judgments
	1 Introduction
	2 Related Work
	3 Identifying Potential Errors in the Training Instances
	4 Training Datasets with Potential Errors
	5 Review of Potential Errors in Positive Instances
	6 Experiments
	7 Conclusion
	References
The Effect of Prolonged Exposure to Online Education on a Classroom Search Companion
	1 Introduction
	2 Background and Related Work
	3 Experimental Set-Up
	4 Results and Discussion
		4.1 Stage 1: Children's Expectations
		4.2 Stage 2: Unintended Consequences of the Pandemic
		4.3 Discussion
	5 Conclusions, Limitations, and Future Work
	References
Rhythmic and Psycholinguistic Features for Authorship Tasks in the Spanish Parliament: Evaluation and Analysis
	1 Introduction
	2 Related Work
	3 Experimental Setting
		3.1 Dataset: ParlaMint
		3.2 Feature Extraction: BaseFeatures and Text Encodings
		3.3 Experimental Protocol
	4 Results
	5 Post-hoc Analysis of the AV Results
		5.1 One-Way ANOVA Test for Political Groups
		5.2 Spearman Coefficient Applied to Style Indices
	6 Conclusion and Future Work
	References
The Impact of Pre-processing on the Performance of Automated Fake News Detection
	1 Introduction
	2 Related Work
		2.1 Fake News Detection
		2.2 The Impact of Pre-processing in NLP Tasks
	3 Experiments
		3.1 Fake News Dataset
		3.2 Experiment Setting
	4 Results
	5 Conclusion and Future Work
	References
Short Papers
Business Document Information Extraction: Towards Practical Benchmarks
	1 Introduction
	2 Document Information Extraction Problems
		2.1 Key Information Extraction (KIE) and Localization (KILE)
		2.2 Table Extraction and Line Items
		2.3 One-Shot Learning for Information Extraction
		2.4 Other Document IE Problems and Tasks
	3 Semi-structured Business Document Datasets
	4 Where to Get More Documents
	5 Discussion and Future Work
	References
An Analysis of Logic Rule Dissemination in Sentiment Classifiers
	1 Introduction
	2 Logic Rule Dissemination Methods
		2.1 Network Architecture
		2.2 Iterative Rule Knowledge Distillation
		2.3 Contextual Word Embeddings
	3 Methodology
	4 Experimental Evaluation
		4.1 Dataset
		4.2 Performance Evaluation
	5 Conclusion
	References
Using Entities in Knowledge Graph Hierarchies to Classify Sensitive Information
	1 Introduction
	2 Related Work
	3 Hierarchy Modelling and Features
		3.1 Building Hierarchy Tree
		3.2 Feature Development
	4 Experiments
		4.1 Experimental Setup
		4.2 Results
	5 Conclusions
	References
Best of 2021 Labs
Evaluating Research Dataset Recommendations in a Living Lab
	1 Introduction
	2 Related Work
	3 Research Dataset Recommendations
		3.1 The GESIS Corpus
		3.2 Data Pre-processing
		3.3 Baseline Retrieval
		3.4 Re-ranking
	4 System Pre-testing
	5 Experimental Evaluation
	6 Conclusion
	References
Analysing Moral Beliefs for Detecting Hate Speech Spreaders on Twitter
	1 Introduction
	2 Feature Selection
		2.1 Hate Speech Detection
		2.2 Moral Values Detection
		2.3 Named Entity Recognition of HS Target
		2.4 Communicative Behaviour
	3 Error Analysis
	4 Features Analysis and Improvement
		4.1 MFT Values
		4.2 Named Entities
		4.3 Fine-Tuning
	5 Conclusions
	References
Query Expansion, Argument Mining and Document Scoring for an Efficient Question Answering System
	1 Introduction
	2 Related Work
		2.1 Argument Mining
		2.2 Comparative Question Answering
	3 Approach and Implementation
		3.1 Query Expansion
		3.2 Document Retrieval by ChatNoir API
		3.3 Document Aggregation
		3.4 Argument Mining
		3.5 Scoring
		3.6 Normalization and Scores Combination
		3.7 Sorting
	4 Evaluation
		4.1 On the Level of Argument Identification
		4.2 On the Level of Touché Shared Task 2
	5 Discussion
	6 Conclusion
	References
Transformer-Encoder-Based Mathematical Information Retrieval
	1 Introduction
	2 ARQMath 2021 Lab
	3 Related Work
	4 Transformer-Encoder-Based MIR
		4.1 Pre-training
		4.2 ALBERT Model
		4.3 ColBERT Model
		4.4 Evaluation Data
	5 Experiments
		5.1 Evaluation
	6 Conclusion
	A  Tokenization
	B  Hyperparameters
		B.1  ALBERT Models
		B.2  ColBERT Models
	References
ImageCLEF 2021 Best of Labs: The Curious Case of Caption Generation for Medical Images
	1 Introduction
	2 Background and Related Work
	3 Methodology
		3.1 Task Description and Dataset
		3.2 Metrics
		3.3 Models
		3.4 Medical Image Pre-processing
		3.5 Caption Formatting and Generation
		3.6 Fine-Tuning
	4 Results and Discussion
		4.1 Assumption 1: Pre-training Data
		4.2 Assumption 2: Encoder
		4.3 Assumption 3: Decoder
		4.4 Assumption 4: Formatting
		4.5 Similarity Between Predicted Captions
	5 Conclusion
	References
Data-Centric and Model-Centric Approaches for Biomedical Question Answering
	1 Introduction
	2 Methods
		2.1 Data-Centric Approach; Factoid Questions
		2.2 Model-Centric Approach; List Questions
		2.3 Yes/No Questions
		2.4 Ideal Answers
	3 Results
		3.1 Exact Answers
		3.2 Ideal Answers
	4 Conclusion
	References
Did I Miss Anything? A Study on Ranking Fusion and Manual Query Rewriting in Consumer Health Search
	1 Introduction
		1.1 Research Proposal
	2 Methodology
	3 Experimental Setting
		3.1 Dataset
		3.2 Topics
		3.3 Baselines
		3.4 Relevance Judgements
		3.5 Experiments
		3.6 Evaluation Measures
		3.7 Additional Notes
	4 Results
		4.1 Reciprocal Ranking Fusion
		4.2 Manual Query Rewriting
	5 Final Remarks and Future Work
	References
Overviews of 2022 Labs
Overview of eRisk 2022: Early Risk Prediction on the Internet
	1 Introduction
	2 Task 1: Early Detection of Pathological Gambling
		2.1 Decision-Based Evaluation
		2.2 Ranking-Based Evaluation
		2.3 Task 1: Results
	3 Task 2: Early Detection of Depression
		3.1 Task 2: Results
	4 Task 3: Measuring the Severity of Eating Disorders
		4.1 Task 3: Evaluation Metrics
		4.2 Task 3: Results
	5 Conclusions
	References
Overview of LifeCLEF 2022: An Evaluation of Machine-Learning Based Species Identification and Species Distribution Prediction
	1 LifeCLEF Lab Overview
	2 PlantCLEF Challenge: Identify the World's Flora
		2.1 Objective
		2.2 Dataset
		2.3 Evaluation Protocol
		2.4 Participants and Results
	3 BirdCLEF Challenge: Bird Call Identification in Soundscape Recordings
		3.1 Objective
		3.2 Dataset and Evaluation Protocol
		3.3 Participants and Results
	4 GeoLifeCLEF Challenge: Predicting Species Presence from Multi-modal Remote Sensing, Bioclimatic and Pedologic Images
		4.1 Objective
		4.2 Data Set and Evaluation Protocol
		4.3 Participants and Results
	5 SnakeCLEF Challenge: Automated Snake Species Identification on a Global Scale
		5.1 Objective
		5.2 Dataset and Evaluation Protocol
		5.3 Participants and Results
	6 FungiCLEF Challenge: Fungi Recognition as an Open Set Classification Problem
		6.1 Objective
		6.2 Dataset and Evaluation Protocol
		6.3 Participants and Results
	7 Conclusions and Perspectives
	References
Overview of ARQMath-3 (2022): Third CLEF Lab on Answer Retrieval for Questions on Math
	1 Introduction
	2 Related Work
	3 The ARQMath Stack Exchange Collection
	4 Task 1: Answer Retrieval
		4.1 Topics
		4.2 Participant Runs
		4.3 Baseline Runs
		4.4 Relevance Assessment
		4.5 Assessor Selection
		4.6 Evaluation Measures
		4.7 Results
	5 Task 2: Formula Search
		5.1 Topics
		5.2 Participant Runs
		5.3 Baseline Run: Tangent-S
		5.4 Assessment
		5.5 Evaluation Measures
		5.6 Results
	6 Task 3: Open Domain Question Answering
		6.1 Topics and Participant Runs
		6.2 Baseline Run: GPT-3
		6.3 Assessment
		6.4 Evaluation Measures
		6.5 Results
	7 Conclusion
	References
Overview of Touché 2022: Argument Retrieval
	1 Introduction
	2 Related Work
		2.1 Argument Retrieval
		2.2 Retrieval for Comparisons
		2.3 Image Retrieval
	3 Lab Overview and Statistics
	4 Task 1: Argument Retrieval for Controversial Questions
		4.1 Task Definition
		4.2 Data Description
		4.3 Participant Approaches
		4.4 Task Evaluation
		4.5 Task Results
	5 Task 2: Argument Retrieval for Comparative Questions
		5.1 Task Definition
		5.2 Data Description
		5.3 Participant Approaches
		5.4 Task Evaluation
		5.5 Task Results
	6 Task 3: Image Retrieval for Arguments
		6.1 Task Definition
		6.2 Data Description
		6.3 Participant Approaches
		6.4 Task Evaluation
		6.5 Task Results
	7 Conclusion
	References
Overview of BioASQ 2022: The Tenth BioASQ Challenge on Large-Scale Biomedical Semantic Indexing and Question Answering
	1 Introduction
	2 Overview of the Tasks
		2.1 Large-Scale Semantic Indexing - Task 10a
		2.2 Biomedical Semantic QA - Task 10b
		2.3 Task Synergy
		2.4 Medical Semantic Annotation in Spanish - DisTEMIST
	3 Overview of Participation
		3.1 Task 10a
		3.2 Task 10b
		3.3 Task Synergy
		3.4 Task DisTEMIST
	4 Results
		4.1 Task 10a
		4.2 Task 10b
		4.3 Task Synergy
		4.4 Task DisTEMIST
	5 Conclusions
	References
A Concise Overview of LeQua@CLEF 2022: Learning to Quantify
	1 Learning to Quantify
	2 The Rationale for LeQua 2022
	3 Setting up LeQua 2022
		3.1 Tasks
		3.2 The Evaluation Protocol
		3.3 The Evaluation Measures
		3.4 Data
		3.5 Baselines
	4 The Participating Systems
	5 Results
		5.1 T1A and T1B: The Vector Subtasks
		5.2 T2A and T2B: The Raw Documents Subtasks
		5.3 T1A and T2A: The Binary Subtasks
		5.4 T1B and T2B: The Multiclass Subtasks
	6 Final Remarks
	References
Overview of PAN 2022: Authorship Verification, Profiling Irony and Stereotype Spreaders, and Style Change Detection
	1 Introduction
	2 Authorship Verification
	3 Author Profiling
	4 Multi-author Writing Style Analysis
	References
Intelligent Disease Progression Prediction: Overview of iDPP@CLEF 2022
	1 Introduction
	2 Related Challenges
	3 Tasks
		3.1 Pilot Task 1: Ranking Risk of Impairment
		3.2 Pilot Task 2: Predicting Time of Impairment
		3.3 Position Papers Task 3: Explainability of AI Algorithms
	4 Dataset
	5 Lab Setup and Participation
		5.1 Guidelines
		5.2 Participants
	6 Evaluation Measures
		6.1 Pilot Task 1: Ranking Risk of Impairment
		6.2 Pilot Task 2: Predicting Time of Impairment
	7 Results
		7.1 Pilot Task 1: Ranking Risk of Impairment for ALS
		7.2 Pilot Task 2: Predicting Time of Impairment for ALS
		7.3 Approaches
	8 Conclusions and Future Work
	References
Overview of HIPE-2022: Named Entity Recognition and Linking in Multilingual Historical Documents
	1 Introduction
	2 Task Description
	3 Data
		3.1 Original Datasets
		3.2 Corpora Characteristics
		3.3 HIPE-2022 Releases
	4 Evaluation Framework
		4.1 Task Bundles, Tracks and Challenges
		4.2 Evaluation Measures
		4.3 System Evaluation, Scorer and Evaluation Toolkit
	5 System Descriptions
		5.1 Baselines
		5.2 Participating Systems
	6 Results and Discussion
	7 Conclusion and Perspectives
	References
Overview of JOKER@CLEF 2022: Automatic Wordplay and Humour Translation Workshop
	1 Introduction
	2 Background
		2.1 Typologies of Wordplay
		2.2 Translation of Wordplay
		2.3 Computational Humour
	3 Data
		3.1 Parallel Corpus of Puns
		3.2 Parallel Corpus of Term-Based Wordplay
	4 Pilot Task 1: Classify and Explain Instances of Wordplay
		4.1 Evaluation Metrics
		4.2 Results
	5 Pilot Task 2: Translate Single Terms Containing Wordplay
		5.1 Data
		5.2 Evaluation Metrics
		5.3 Results
	6 Pilot Task 3: Translate Entire Phrases Containing Wordplay
		6.1 Data
		6.2 Evaluation Metrics
		6.3 Results
	7 Conclusion
	References
Overview of the CLEF 2022 SimpleText Lab: Automatic Simplification of Scientific Texts
	1 Introduction
	2 Related Work
	3 Task 1: What Is in (or Out)?
		3.1 Evaluation Framework
		3.2 Results
	4 Task 2: What Is Unclear?
		4.1 Evaluation Framework
		4.2 Results
	5 Task 3: Rewrite This!
		5.1 Evaluation Framework
		5.2 Results
	6 Conclusion and Future Work
	References
Overview of the CLEF–2022 CheckThat! Lab on Fighting the COVID-19 Infodemic and Fake News Detection
	1 Introduction
	2 Previously on CheckThat!
	3 Description of the Tasks
		3.1 Task 1: Identifying Relevant Claims in Tweets
		3.2 Task 2: Detecting Previously Fact-Checked Claims
		3.3 Task 3: Fake News Detection
	4 Datasets
		4.1 Task 1: Identifying Relevant Claims in Tweets
		4.2 Task 2: Detecting Previously Fact-Checked Claims
		4.3 Task 3: Fake News Detection
	5 Evaluation
	6 Results for Task 1: Identifying Relevant Claims in Tweets
		6.1 Task 1A. Check-Worthiness Estimation
		6.2 Subtask 1B: Verifiable Factual Claims Detection
		6.3 Subtask 1C: Harmful Tweet Detection
		6.4 Subtask 1D: Attention-Worthy Tweet Detection
	7 Results for Task 2: Verified Claim Retrieval
		7.1 Subtask 2A: Detecting Previously Fact-Checked Claims in Tweets
		7.2 Subtask 2B: Detecting Previously Fact-Checked Claims in Political Debates and Speeches
	8 Results for Task 3: Fake News Detection
	9 Related Work
	10 Conclusion and Future Work
	References
Overview of ChEMU 2022 Evaluation Campaign: Information Extraction in Chemical Patents
	1 Introduction
	2 Related Work
	3 The ChEMU Chemical Reaction Corpus
		3.1 Data Selection
		3.2 Annotation Guidelines
		3.3 Annotation Process
		3.4 Data Partitions
	4 Task Definition
		4.1 Task 1: Named Entity Recognition
		4.2 Task 2: Event Extraction
		4.3 Task 3: Anaphora Resolution
	5 Evaluation Framework
		5.1 Evaluation Methods
		5.2 Coreference Linkings in Anaphora Resolution Task
		5.3 Baselines
		5.4 NER and EE Baseline
	6 Results and Discussions
		6.1 Task 1: Named Entity Recognition
		6.2 Task 2: Event Extraction
		6.3 Task 3: Anaphora Resolution
	7 Conclusions
	References
Overview of the ImageCLEF 2022: Multimedia Retrieval in Medical, Social Media and Nature Applications
	1 Introduction
	2 Overview of Tasks and Participation
	3 The Caption Task
		3.1 Task Setup
		3.2 Data Set
		3.3 Participating Groups and Submitted Runs
		3.4 Results
		3.5 Lessons Learned and Next Steps
	4 The Tuberculosis Task
		4.1 Task Setup
		4.2 Data Set
		4.3 Participating Groups and Submitted Runs
		4.4 Results
		4.5 Lessons Learned and Next Steps
	5 The Fusion Task
		5.1 Task Setup
		5.2 Data Set
		5.3 Participating Groups and Submitted Runs
		5.4 Results
		5.5 Lessons Learned and Next Steps
	6 The Aware Task
		6.1 Task Setup
		6.2 Data Set
		6.3 Participating Groups and Submitted Runs
		6.4 Results
		6.5 Lessons Learned and Next Steps
	7 The Coral Task
		7.1 Task Setup
		7.2 Data Set
		7.3 Participating Groups and Submitted Runs
		7.4 Results
		7.5 Lessons Learned
	8 Conclusion
	References
Author Index