Machine Learning and Knowledge Discovery in Databases. Research Track : European Conference, ECML PKDD 2024, Vilnius, Lithuania, September 9–13, 2024, Proceedings, Part V / edited by Albert Bifet, Jesse Davis, Tomas Krilavičius, Meelis Kull, Eirini Ntoutsi, Indrė Žliobaitė.

Format:

Book

Contributor:

Bifet, Albert, editor.

Series:

Lecture Notes in Artificial Intelligence, 2945-9141 ; 14945

Language:

English

Subjects (All):

Artificial intelligence.

Computer engineering.

Computer networks.

Computers.

Image processing--Digital techniques.

Image processing.

Computer vision.

Software engineering.

Artificial Intelligence.

Computer Engineering and Networks.

Computing Milieux.

Computer Imaging, Vision, Pattern Recognition and Graphics.

Software Engineering.

Local Subjects:

Artificial Intelligence.

Computer Engineering and Networks.

Computing Milieux.

Computer Imaging, Vision, Pattern Recognition and Graphics.

Software Engineering.

Physical Description:

1 online resource (512 pages)

Edition:

1st ed. 2024.

Place of Publication:

Cham : Springer Nature Switzerland : Imprint: Springer, 2024.

Summary:

This multi-volume set, LNAI 14941 to LNAI 14950, constitutes the refereed proceedings of the European Conference on Machine Learning and Knowledge Discovery in Databases, ECML PKDD 2024, held in Vilnius, Lithuania, in September 2024. The papers presented in these proceedings are from the following three conference tracks: - Research Track: The 202 full papers presented here, from this track, were carefully reviewed and selected from 826 submissions. These papers are present in the following volumes: Part I, II, III, IV, V, VI, VII, VIII. Demo Track: The 14 papers presented here, from this track, were selected from 30 submissions. These papers are present in the following volume: Part VIII. Applied Data Science Track: The 56 full papers presented here, from this track, were carefully reviewed and selected from 224 submissions. These papers are present in the following volumes: Part IX and Part X.

Contents:

Intro

Preface

Organization

Invited Talks Abstracts

The Dynamics of Memorization and Unlearning

The Emerging Science of Benchmarks

Enhancing User Experience with AI-Powered Search and Recommendations at Spotify

How to Utilize (and Generate) Player Tracking Data in Sport

Resource-Aware Machine Learning-A User-Oriented Approach

Contents - Part V

Research Track

Fair Densest Subgraph Across Multiple Graphs

1 Introduction

2 Preliminary Notation and Problem Definition

3 Computational Complexity

4 Algorithms

4.1 Exact Algorithm for FDS

4.2 Exact Algorithm for SDS

4.3 Greedy Algorithm to Find a Good Solution for SDS

4.4 Greedy Algorithm to Find a Good Solution for FDS

4.5 Exact Algorithm for MDS

5 Related Work

6 Experimental Evaluation

6.1 Synthetic Dataset

6.2 Results for the Synthetic Dataset

6.3 Real-World Datasets

6.4 Results for the Real-World Datasets

6.5 Case Study

7 Concluding Remarks

References

A Human-Centric Assessment of the Usefulness of Attribution Methods in Computer Vision

2 Related Work

3 Conceptual Design of the Framework

3.1 Formal Definition

3.2 Assumptions and Discussion

4 Experimental Design

4.1 Idea, Goal, and Challenges

4.2 Dataset Creation

4.3 AI Model and Attribution Methods

4.4 Implementation of the Assessment Framework

5 Results

6 Discussion

7 Conclusion

A Details of the Dataset Creation Process

B Data Compliance with GDPR

Leveraging Plasticity in Incremental Decision Trees

2 Setting

3 Preliminaries

4 PLASTIC

4.1 Concept

4.2 Algorithm

5 Evaluation

5.1 Algorithms

5.2 Data Streams

5.3 PLASTIC vs. EFDT

5.4 Average Accuracy and Ranks

5.5 Runtime

6 Conclusion

References.

Fast-FedUL: A Training-Free Federated Unlearning with Provable Skew Resilience

3 Problem Formulation and Design Principles

4 Fast-FedUL

4.1 Overview

4.2 Efficient Sampling of Local Updates

4.3 Removal of Target Client Updates

4.4 Skew Estimation

4.5 Theoretical Analysis

5 Empirical Evaluation

5.1 End-to-End Comparison

5.2 Ablation Study

5.3 Qualitative Study

Input Compression with Positional Consistency for Efficient Training and Inference of Transformer Neural Networks

2 Data Augmentation Through Input Compression

3 Consistency-Aware Position Selection: Enabling ICPC in an Architecture-Agnostic Manner

4 Efficient Variable-Effort Inference Using ICPC

5 Experimental Results

5.1 Primary Results

5.2 Ablation: Evaluation of ICPC as a Data Augmentation Method

5.3 Improving Accuracy with Hardware-Aware Test-Time Augmentation

6 Related Work

Individual Fairness with Group Awareness Under Uncertainty

2.1 Fairness in ML

2.2 Survival Analysis

3 Notations and Problem Definition

4 Methodology

4.1 Quantifying Individual Fairness with Censorship

4.2 Quantifying Individual-Group Fairness with Censorship

4.3 Mitigating Bias Under Censorship

5 Experiment

5.1 Datasets

5.2 Baselines

5.3 Evaluation Metrics

5.4 Experimental Results

Generalizing Reward Modeling for Out-of-Distribution Preference Learning

2 Preliminaries

3 Meta-learning for Generalizing Reward Modeling

3.1 Preference Learning via Bilevel Programming

3.2 Meta-learning for Out-of-Distribution Preference Learning

3.3 Gradient-Based Algorithm

3.4 Convergence Analysis

4 Experiments.

4.1 Experimental Setup

4.2 Evaluation with Preference Data

4.3 Evaluation with a Learned RM

4.4 Evaluation with GPT-4 Judgement

4.5 Effects of on Reward Learning

5 Conclusion

FedPrime: An Adaptive Critical Learning Periods Control Framework for Efficient Federated Learning in Heterogeneity Scenarios

2.1 Client Selection in Federated Learning

2.2 Critical Learning Periods in Federated Learning

3 Problem Formulation

4 The Design of FedPrime Framework

4.1 The Overview of FedPrime Framework

4.2 Client Selection Module

4.3 Adaptive CLP Control Module

5.1 Experimental Setup

5.2 Test Accuracy

5.3 Generalization

5.4 Ablation Study

What Drives Online Popularity: Author, Content or Sharers? Estimating Spread Dynamics with Bayesian Mixture Hawkes

2.1 Hawkes Process

2.2 Dual Mixture Model

3 Bayesian Mixture Hawkes (BMH) Model

3.1 BMH-P, the Popularity Submodel

3.2 BMH-K, the Kernel Submodel

4 Predictive Evaluation

4.1 Datasets

4.2 Cold-Start Popularity Prediction

4.3 Temporal Profile Generalization Performance

5 What-If? Headline Style Profiling

6 Conclusion and Future Work

Disagreement Evaluation of Solutions for Math Word Problem

3 Method

3.1 Problem Statement

3.2 LoRA Fine-Tuning

3.3 Proposed Method

4 Theoretical Analysis

5.2 Results and Analysis

5.3 Case Study

Co-attention and Contrastive Learning Driven Knowledge Tracing

2.1 Knowledge Tracing

2.2 Contrastive Learning

3 Preliminary

4 Methodology.

4.1 Learning Activity Representation

4.2 Ability and Knowledge Co-attention Layer

4.3 Knowledge Internalization Encoder

4.4 Contrastive Learning

4.5 Performance Prediction

5.2 Baseline Methods

5.3 Experimental Settings

5.4 Performance Prediction Results

5.5 Ablation Study

5.6 Parameter Sensitivity Analysis

5.7 Visualizations

TiNID: A Transfer and Interpretable LLM-Enhanced Framework for New Intent Discovery

2 TiNID

2.1 Problem Formalization

2.2 Model Overview

2.3 Knowledge Generation

2.4 Knowledge Decoupling

2.5 Knowledge Transfer

2.6 Knowledge Interpretation

3 Experiments

3.1 Datasets and Baseline Models

3.2 Evaluation Metrics

3.3 Implementation Details

3.4 Main Results

4 Qualitative Analysis

Memory-Enhanced Emotional Support Conversations with Motivation-Driven Strategy Inference

2.1 Emotional Support Conversation Systems

2.2 Empathetic Dialogue Systems

3 Methodology

3.1 Task Definition and Model Overview

3.2 Context Encoder

3.3 Motivation-Driven Support Strategy Inference

3.4 Memory-Enhanced Strategy Modeling

3.5 Response Decoder

3.6 Model Optimization

4 Experiments

4.1 Datasets, Baselines and Evaluation Metrics

4.2 Implementation Details

4.3 Experimental Results

4.4 Discussions

4.5 Case Study

Semi-automated Construction of Complex Knowledge Base Question Answering Dataset Using Large Language Model

3 Enhancing the Movie Knowledge Graph (iMKG)

3.1 Assigning Wikidata URI to Movie Entities of MovieKG

3.2 Assigning Wikidata URI to Other Entities of MovieKG.

4 Leveraging LLM to Construct the Movie Complex Question Answering (MCQA) Dataset

4.1 Question Types Creation

4.2 Templates Creation

4.3 Data Sampling

5 Experiments

5.2 Evaluating KBQA Methods on MCQA

Graphical Model-Based Lasso for Weakly Dependent Time Series of Tensors

2 Methodology

2.1 Notation

2.2 Model Formulation

2.3 Near-Oracle Estimation Error Bound

3 Simulation Study

3.1 Experimental Setting

3.2 Results

4 Crypto-Asset Portfolio Construction

Multi-label Adaptive Batch Selection by Highlighting Hard and Imbalanced Samples

2.1 Multi-label Classification

2.2 Hard Sample and Batch Selection in Single Label Data

3 Proposed Method

3.1 Preliminaries

3.2 Rank-Based Batch Selection

3.3 Class Imbalance Aware Weighting

3.4 Incorporate Quantization Index

3.5 Variant of Adaptive Batch Selection Exploiting Label Correlations

3.6 Convergence Guarantee

4 Experiments and Analysis

4.1 Experiment Setup

4.2 Experimental Results

4.3 Convergence Analysis

4.4 Adaptive Batch Selection with Label Correlations

4.5 Investigation of Loss

4.6 Parameter Analysis

Quantification Over Time

2 Definitions, Notation and Background

2.1 Quantification

2.2 Quantification over Time (QoT)

2.3 QoT, Quantification and Time Series Forecasting

3 Related Work

3.1 QoT with Quantification only

3.2 QoT with Classify and Count and Moving Average

4.1 Adjustment Framework

4.2 KF-MA

5 Experimental Evaluation

5.2 Comparison with QoT Approaches in Literature

5.3 Results

Approximating the Graph Edit Distance with Compact Neighborhood Representations.

Notes:

Includes bibliographical references and index.

ISBN:

3-031-70362-6