Trends and Applications in Knowledge Discovery and Data Mining : PAKDD 2021 Workshops, WSPA, MLMEIN, SDPRA, DARAI, and AI4EPT, Delhi, India, May 11, 2021 Proceedings

This book constitutes the refereed proceedings of five workshops that were held in conjunction with the 25th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2021, in Delhi, India, in May 2021. The 17 revised full papers presented were carefully reviewed and selected from a total of 39 submissions.. The five workshops were as follows: Workshop on Smart and Precise Agriculture (WSPA 2021) PAKDD 2021 Workshop on Machine Learning for Measurement Informatics (MLMEIN 2021) The First Workshop and Shared Task on Scope Detection of the Peer Review Articles (SDPRA 2021) The First International Workshop on Data Assessment and Readiness for AI (DARAI 2021) The First International Workshop on Artificial Intelligence for Enterprise Process Transformation (AI4EPT 2021) Preface Organization Contents Workshop on Smart and Precise Agriculture (WSPA 2021) Identification of Harvesting Year of Barley Seeds Using Near-Infrared Hyperspectral Imaging Combined with Convolutional Neural Network 1 Introduction 2 Materials and Methodology 2.1 Barely Samples and Near-Infrared Hyperspectral Imaging System 2.2 Model Development and Validation 3 Results and Discussion 4 Conclusions References Plant Leaf Disease Segmentation Using Compressed UNet Architecture 1 Introduction and Literature Survey 2 Proposed Model 2.1 Selection in DE 3 Experimental Setup and Results 4 Conclusion A Appendix I A.1 UNet architecture References PAKDD 2021 Workshop on Machine Learning for MEasurement INformatics (MLMEIN 2021) Hierarchical Topic Model for Tensor Data and Extraction of Weekly and Daily Patterns from Activity Monitor Records 1 Introduction 2 Model 3 Parameter Inference 4 Application to Step-Count Dataset 4.1 Dataset 4.2 Classification of Ambulatory Activities 4.3 Comparison with Tensor Train Decomposition 4.4 Effects of Weekly Activity Patterns on Body Weight Changes 5 Concluding Remarks References Convolutional Neural Network to Detect Deep Low-Frequency Tremors from Seismic Waveform Images 1 Introduction 2 Seismographs 3 Methods 3.1 Generation of Synthetic Images 3.2 Preprocessing of Synthetic Images 3.3 CNN for Tremor Detection 4 Results 5 Conclusion References Unsupervised Noise Reduction for Nanochannel Measurement Using Noise2Noise Deep Learning 1 Introduction 2 Related Work and Research Objective 3 Outline of Multistage Narrowing Nanochannel 4 Principle and Implementation of Noise2Noise Deep Learning 4.1 Principle 4.2 Implementation 5 Training of Noise Reduction Model 5.1 Training Data, Training Procedures, and Performance Verification 5.2 Methods for Performance Comparison 5.3 Verification Results 6 Verification for a Two-Particle Passing Waveform 7 Conclusion References Classification Bandits: Classification Using Expected Rewards as Imperfect Discriminators 1 Introduction 2 Problem Formulation 3 Problem Reduction 3.1 Reduction Theorem 3.2 Reducible Parameter Region 4 Algorithm 4.1 Decision Condition 4.2 Arm Selection Policies 5 Experiments 6 Conclusion and Future Works References The First Workshop and Shared Task on Scope Detection of the Peer Review Articles (SDPRA 2021) Overview and Insights from Scope Detection of the Peer Review Articles Shared Tasks 2021 1 Introduction 2 Problem Definition and Datasets 3 Participating Teams 4 Results and Discussion 4.1 Results 4.2 Code Reproducibility 5 Conclusion and Future Scope References Scholarly Text Classification with Sentence BERT and Entity Embeddings 1 Introduction 1.1 Task Description 2 Proposed Approach 2.1 Pre-trained Sentence Encoders 2.2 Training 3 Results 4 Conclusions References Domain Identification of Scientific Articles Using Transfer Learning and Ensembles 1 Introduction 2 Related Work 3 Dataset 4 System Description 4.1 System Architecture 4.2 Data Preprocessing 4.3 BERT 4.4 RoBERTa 4.5 SciBERT 4.6 Ensembling 5 Results and Analysis 6 Conclusion References Identifying Topics of Scientific Articles with BERT-Based Approaches and Topic Modeling 1 Introduction 2 Related Work 3 Shared Task 3.1 Task Definition 3.2 Dataset 4 Methodology 5 Experiments and Results 6 Conclusion References Using Transformer Based Ensemble Learning to Classify Scientific Articles 1 Introduction 1.1 Problem Description 1.2 Related Works 1.3 Our Contributions 2 Dataset 2.1 Data Description 2.2 Pre-processing and Feature Engineering 3 Methodology 3.1 System Description of Model-1 (RoBERTa) 3.2 System Description of Model-2 (RoBERTa+LDA) 3.3 System Description of Model-3 (RoBERTa on Sentences) 3.4 System Description of Model-4 (TF-IDF + Logistic Regression) 3.5 Submissions 4 Experiments, Results and Discussion 4.1 Experimental Setup 4.2 Results and Discussion 5 Conclusion and Future Works References The First International Workshop on Data Assessment and Readiness for AI (DARAI 2021) 1st International Workshop on Data Assessment and Readiness for AI 1 Objectives, Scope, and Contributions 1.1 Topics of Interest 2 Proposed Workshop Format and Attendees 3 Expertise and Experience of the Organizing Committee References Cooperative Monitoring of Malicious Activity in Stock Exchanges 1 Problem Statement 2 Cooperative Target Observation (CTO) 2.1 Generation of Alerts 2.2 Trade Level Alerts Generation 2.3 Trader Level Alerts Generation 3 Experiments 3.1 Data Generation 3.2 Attack Types 3.3 Parameters for Alerts Generation Module 3.4 Malicious Trader Identification 4 Conclusions and Future Work References Data-Debugging Through Interactive Visual Explanations 1 Introduction 2 Related Work 3 Data-Debugging Through Interactive Visual Explanations 4 Example: Label Purity 4.1 What Is Label Purity 4.2 Explanation 5 Summary and Future Work References Data Augmentation for Fairness in Personal Knowledge Base Population 1 Introduction 2 Related Work 3 Data Augmentation 4 KBP Evaluation 5 Fairness Analysis 6 Representative Set Sampling 7 Conclusion References The First International Workshop on Artificial Intelligence for Enterprise Process Transformation (AI4EPT 2021) ROC Bot: Towards Designing Virtual Command Centre for Energy Management 1 Introduction 2 Related Work 3 ROC Bot: Virtual Command Centre for Energy Management 3.1 Deep Learning Based Framework for Query Sketch Generation 3.2 Analytics, Automation and Anomaly Detection 3.3 Summarization 4 Results and Discussion 5 Conclusion and Future Work References Digitize-PID: Automatic Digitization of Piping and Instrumentation Diagrams 1 Introduction 2 Proposed Method: Digitize-PID 2.1 Detection 2.2 Comprehension 2.3 Reconciliation 3 Dataset 4 Experimental Results and Discussions 5 Conclusion References Author Index