Conveners
[EI6] To the Edge and Beyond AI in Computer Vision: [EI7] To the Edge and Beyond AI in Computer Vision
- Seul-Ki Yeom (Nota AI GmbH)
Description
Within the last decade, advances in Deep Learning, coupled with the large, freely available datasets as well as extremely high computing resources, have resulted in remarkable progress in the computer vision (CV), natural language processing (NLP), and broader artificial intelligence (AI) communities. Many academies and industries are also involved in developing/researching AI with Machine Learning to accelerate scientific discovery and engineering design in diverse application domains (e.g. ChatGPT, AlphaGo Zero, etc).
However, research in the AI field also shows that their performance on ranging from edge-device to high performance computing (e.g. Cloud server) is still far from practical towards open-world data and scenarios. Besides the accuracy that is widely concerned in deep learning, the phenomena are significantly related to the studies about AI model efficiency and robustness, which we abstract as “To the Edge and Beyond AI”.
In this reason, this workshop focuses on an emerging and impactful topic efficient artificial intelligence especially in computer vision field which is one of the most popular and practical fields in AI domain such as driving monitoring system, intelligence transportation system, etc. It aims to discuss and share the challenges in applying AI to specific science and engineering problem in computer vision field based on machine learning methods.
It will feature a host of invited talks covering a variety of topics in AI in CV through several domain experts as follows,
1. AI model compression techniques on edge device (pruning, quantization, etc.)
2. CV applications in AI
3. Medical image processing based on machine learning
The progress of deep learning (DL) and artificial intelligence is astonishing, and it attracts numerous researchers and practitioners from multidisciplinary domains. Although tremendous literature regarding DL applications has been published across domains, it is uncommon that DL applications are actually deployed in the daily routine. This study empirically investigated popular deep learning...
Global navigation satellite system-based navigation is sensitive to disturbances and jamming, hence the capability to provide reliable position accuracy without GNSS is a key element to develop the navigation systems. Recently, camera sensors have been widely used in navigation system due to the development and application of deep learning techniques. This paper proposes a deep learning based...
Magnetic resonance imaging (MRI) is a non-invasive medical imaging technique that provides detailed information about the internal structure and function of the human body. MRI signals are acquired and stored in k-space, which is the spatial frequency domain representation of these signals. The signals are then reconstructed into MR images using linear operators, such as Fourier Transform....
Neural networks performance has been significantly improved in the last few years, at the cost of an increasing number of floating point operations per second (FLOPs). However, more FLOPs can be an issue when computational resources are limited. As an attempt to solve this problem, pruning filters is a common solution, but most existing pruning methods do not preserve the model accuracy...
