AI4AN 2020: The 1st Workshop on Artificial Intelligence for Anomalies and Novelties (co-located with IJCAI20) Yokohama, Japan, July 13, 2020 |
Conference website | https://sites.google.com/view/ai4an2020/home |
Submission link | https://easychair.org/conferences/?conf=ai4an2020 |
Submission deadline | April 26, 2020 |
Anomalies are referred to as observations or events which are rare or significantly different from the majority of observations we have in hand, while novelties are observations from novel classes that are unseen during learning. Recognition, detection and/or adaption of anomalies and novelties are some of the most active research areas in multiple communities, such as data mining, machine learning and computer vision. Some of the most relevant well-established research areas include anomaly detection, out-of-distribution example detection, adversarial example recognition and detection, curiosity-driven reinforcement learning, open-set recognition and adaption. The successful early detection of anomalies and novelties is of great significance in broad domains, e.g., it may prevent the loss of billions of dollars by its application to fraud detection and anti-money laundering in fintech, saves lives by millions through disease detection, safeguards large-scale network computers from malicious attacks by its use in intrusion detection, defenses AI systems from adversarial attacks, and equips AI systems with capabilities to work safely in the open world, etc. Due to this significance, some of these areas have been extensively explored for decades, with the other areas emerged in recent years. However, there are still significant challenges and many open problems in these area due to some unique nature of anomalies and novelties, such as rareness, heterogeneity, unknown and uncertainty.
This workshop aims to gather researchers and practitioners from diverse communities and knowledge background to largely promote the development and applications of anomaly and novelty recognition, detection and adaption techniques.
Submission Guidelines
We invite three types of submissions, including original research paper (6 pages plus references), demo paper (4 pages plus references), visionary papers (4 pages plus references). Submissions must be in PDF format, written in English, and formatted according to the IJCAI 2020 formatting guidelines available at https://www.ijcai.org/authors_kit. All papers will be peer reviewed and assessed based on their novelty, technical quality, potential impact, clarity, and reproducibility. All the papers are required to be submitted via EasyChair system at https://easychair.org/conferences/?conf=ai4an2020.
List of Topics
This workshop will feature the most recent artificial intelligence advances for recognition, detection and adaption anomalies and novelties. It targets both academic researchers and industrial practitioners from data mining, machine learning and computer vision communities, and solicits original research on but not limited to the following topics.
* Recognition/detection/adaption of anomalies/novelties in different types of data
- Anomaly/novelty detection
- Out-of-distribution example detection
- Adversarial example recognition and detection
- Curiosity-driven reinforcement learning
- Open-set recognition and adaption
* Deep learning for anomaly/novelty recognition/detection/adaption
- Feature learning specifically designed for anomalies/novelties
- End-to-end anomaly/novelty recognition/detection/adaption
* Anomaly/novelty-related theories/foundation
- Mathematical formalization
- Optimization
- Generalization bounds and learnability
- Anomaly/novelty explanation
* Relevant applications
- Fraud and risk analysis in finance and insurance
- Disease detection and diagnosis in healthcare
- Intrusion/malware detection in cybersecurity
- Malicious activity detection in social networks
- Misinformation and fake information detection
- Event detection in video surveillance
- Safety analysis in AI systems
Committees
Organizing committee
- Guansong Pang (University of Adelaide, Australia)
- Jundong Li (University of Virginia, United States)
- Anton van den Hengel (University of Adelaide, Australia)
- Longbing Cao (University of Technology Sydney, Australia )
- Thomas G. Dietterich (Oregon State University, United States)
Program committee
TBA
Contact
All questions about submissions should be emailed to guansong.pang@adelaide.edu.au