IAAA 2021

KEYNOTE SPEAKERS

2:00pm-2:10pm Opening Remarks

2:10pm-2:45pm Keynote I

Speaker: Professor Leman Akoglu, Carnegie Mellon University

Title: Recent Trends in Outlier Mining: Automation and Fairness

2:45pm-3:20pm Keynote II

Speaker: Professor Christopher Leckie, University of Melbourne

Title: Anomaly Detection in Hostile Environments

3:20pm-4:30pm Presentations

Jing Ren, Feng Xia, Yemeng Liu, and Ivan Lee, Deep Video Anomaly Detection: Opportunities and Challenges

Ken-Yu Lin, Roy Ka-Wei Lee, Wei Gao, and Wen-Chih Peng, Early Prediction of Hate Speech Propagation

Zhiyue Wu, Hongzuo Xu, Yijie Wang, and Yongjun Wang, Surrogate Supervision-Based Deep Weakly-Supervised Anomaly Detection

Yiming Li, Da Sun Handason Tam, Siyue Xie, Xiaxin Liu, Qiu Fang Ying, Wing Cheong Lau, Dah Ming Chiu, and Shou Zhi Chen, Temporal Graph Representation Learning for Detecting Anomalies in E-payment Systems

4:30pm-5:30pm Break

5:30pm-6:05pm Keynote III

Speaker: Professor Longbing Cao, University of Technology Sydney

Title: Non-IID outlier detection: Learning abnormal couplings, interactions and heterogeneities

6:05pm-6:40pm Keynote IV

Speaker: Professor Edwin Hancock, University of York

Title: Entropic characterization of anomalies in financial time series

6:40pm-8:05pm Presentations

Chakkrit Termritthikun, Lin Xu, Yemeng Liu, and Ivan Lee, Neural Architecture Search and Multi-Objective Evolutionary Algorithms for Anomaly Detection

Chen Cao, Shihao Li, Shuo Yu, and Zhikui Chen, Fake Reviewer Group Detection in Online Review Systems

Bernat Coma Puig and Josep Carmona, A Human-in-the-Loop Approach based on Explainability to Improve NTL Detection

Mingliang Hou, Jing Ren, Falih Febrinanto, Ahsan Shehzad, and Feng Xia, Cross Network Representation Matching with Outliers

Andreas Lohrer, Jan Deller, Maximilian Hünemörder, and Peer Kröger, OAB - An Open Anomaly Benchmark Framework for Unsupervised and Semisupervised Anomaly Detection on Image and Tabular Data Sets

8:05pm-8:10pm Closing

Anomalies commonly exist in various real-world scenarios, such as fraud in finance and insurance, intrusion in cybersecurity, fault in safety-critical systems, bushfire early warning, disease outbreak control, fake news in social media, and medical diagnosis. Some anomalies could cause disasters that lead to immense economic loss or even deaths unless discovered and dealt with on time. These applications make anomaly analytics methods increasingly relevant in the modern world. Due to its foremost importance, the study of anomaly detection has a long history and has created a wealth of anomaly detection methods. With the advent of big data, new challenges and questions are introduced, which inspires novel ways of developing algorithms, methods and techniques to foster the analysis and interpretation of anomalies such as formal anomaly definition and specific anomaly localization.

Recent years have witnessed the rapid growth in the number of academics and practitioners interested in anomaly detection and closely related areas. In particular, various deep neural network models have been developed for anomaly detection. Without human in the loop, however, deep models are hard to tune and hard to interpret.

The workshop aims at bringing together researchers and practitioners to discuss how to detect, predict, and describe anomalies effectively and efficiently. Several core challenges, such as human-in-the-loop, intelligence augmentation, tools and methods for detecting, predicting, and describing anomalies will be the main center of discussions at the workshop. In this workshop, our goal is to contribute to the next generation of anomaly analytics and exploring it using intelligence augmentation, artificial intelligence, data mining, deep learning, and other appropriate technologies.

This workshop would like to share exciting techniques to solve critical problems such as:

• What are the next-generation anomaly analytics models and techniques?

• Why and how to use human-in-the-loop machine learning in anomaly analytics?

• How to leverage intelligence augmentation (and artificial intelligence) in anomaly analytics?

• Can we develop automated anomaly detection systems to identify anomalous objects without any human interventions?

Topics of interest include but not limited to:

• Foundations and understanding of anomaly analytics

• Novel models and algorithms for anomaly analytics

• Intelligence-augmented techniques for anomaly analytics

• Human-in-the-loop deep learning and interactive intelligence

• Trustworthy anomaly analytics

• Fairness, transparency, explainability, and robustness

• Datasets and benchmarking

• Automated anomaly detection systems

• Anomaly analytics in various domains

• Innovative applications of anomaly analytics and/or intelligence augmentation.

SUBMISSION INSTRUCTIONS

Authors are invited to submit original papers that must not have been submitted to or published in any other workshop, conference, or journal. The workshop will accept full papers describing completed work, work-in-progress papers with preliminary results, as well as position papers reporting inspiring and intriguing new ideas.

Paper submissions should be non-anonymous, limited to a maximum of eight (8) pages (plus 2 extra pages), in the IEEE 2-column format (https://www.ieee.org/conferences/publishing/templates.html), including the bibliography and any possible appendices. All submissions will be peer-reviewed by members of the Program Committee and be evaluated for originality, quality and appropriateness to the workshop.

By the unique ICDM tradition, all accepted workshop papers will be published in the dedicated ICDMW proceedings published by the IEEE Computer Society Press.

At least one author of each accepted paper must complete the ICDM 2021 conference registration and present the paper at the workshop.

ICDM 2021 will be a virtual conference, and hence IAAA 2021 will go virtual.