The 1st International Workshop on

Anomaly and Novelty detection in Satellite and Drones systems

(ANSD '23)

Welcome

2:00-2:05PM BST

Welcome message in name of the co-chairs

Presenting: Simon S. Woo

Keynote I

2:05-2:30PM BST

Operation, Application and AI of Satellite in NSOAC

Keynote speaker: Dr. Daewon Chung

Speaker’s Bio

Daewon Chung received B.S. and M.S. degrees in electronics engineering from Kyungbook National University in 1992 and 1994. He received Ph. D. degree in electronics engineering from Chungnam National University in 2007. He joined Korea Aerospace Research Institute, Republic of Korea in 1994. He is currently Executive Director of National Satellite Operation & Application Center at Korea Aerospace Research Institute. He is also a professor at University of Science & Technology. He was chairman of SpaceOps Committee from 2013 to 2017. His research interest is currently AI based satellite image processing such as image domain change, super-resolution and object detection, spacecraft operations and satellite situational awareness.

Abstract

The Korea Aerospace Research Institute (KARI) was established on October 10, 1989 as a government funded research institute under the Ministry of Science and ICT. Since its inception, The KARI has made remarkable achievements in the area of aerospace over the short period of time. The NSOAC(National Satellite Operation and Application Center) was established to promote the operation of national satellites and the use of satellite information. The NSOAC has been researching lots of works related on anomaly and novelty detection in the area of satellite, ground system and satellite image. Nowadays, many satellites have been launching and operating efficiently. Therefore, data coming from satellites, ground systems and kinds of payload is increasingly becoming bigdata. Focus of research on anomaly and novelty detection has been changing from traditional analysis and modeling into new artificial intelligence technology using bigdata. So, various kinds of AI algorithm have been researching appropriateness and effectiveness of use on anomaly and novelty detection in the area of space application.

Technica Papers

2:30-2:50PM BST

Light TranAD : Light Transformer Networks for Anomaly Detection in Orbit Dataset

Authors: SeungWon Jeong, Simon S. Woo, and Youjin Shin

2:50-3:10PM BST

Enhancing Satellite Orbit Propagation and Maneuver Detection

with Deep Learning Techniques

Authors: Jinbeom Kim, Sangyup Lee, Kangjun Lee, Okchul Jung, Daewon Chung, and Simon S. Woo

3:10-3:30PM BST

Challenges in Firmware Re-Hosting for Satellite Systems

Sungjae Hwang, and Haeun Eom

Authors: Sungjae Hwang, and Haeun Eom

Coffee Break

3:30-4:00PM BST

Keynote II

4:00-4:50 BST

Deep Spatiotemporal Point Processes for Anomaly Detection and Beyond

Keynote speaker: Dr. Rose Yu,

Speaker’s Bio

Dr. Rose Yu is an assistant professor at the University of California San Diego, Department of Computer Science and Engineering. Her research focuses on advancing machine learning techniques for large-scale spatiotemporal data analysis, with applications to sustainability, health, and physical sciences. A particular emphasis of her research is on physics-guided AI which aims to integrate first principles with data-driven models. Among her awards, she has won Army ECASE Award, NSF CAREER Award, Hellman Fellow, Faculty Research Award from JP Morgan, Facebook, Google, Amazon, and Adobe, Several Best Paper Awards, Best Dissertation Award at USC, and was nominated as one of the ’MIT Rising Stars in EECS’.

Abstract

Accurate modeling of spatiotemporal events is critical for disaster response, logistic optimization and human mobility. Compared to traditional sequence data such as texts or time series, spatiotemporal events occur irregularly with uneven time and space intervals. Existing statistical approaches based on Spatiotemporal point processes (STPP) often require strong modeling assumptions, feature engineering, and can be computationally expensive. In this talk, I will describe a framework that integrates fundamental concepts from STPP with deep learning, leading to flexible, interpretable, highly efficient models for spatiotemporal events . I will showcase the applications of our framework to event-based forecasting, inference and anomaly detection tasks.

Discussion and Q&A

5:00-5:20PM BST

THE END!

Page updated

Google Sites

Report abuse