Fuzzy Trigger Edge Multimodal Model

Project Description

Advanced developments in computing devices and artificial intelligence based on deep neural networks allow scientists use edge and cloud computing systems to perform complex data processing and analysis in a variety of research areas. Additionally, language models from small to large that are capable of human-like conversations extends accessibility of use of those computing techniques and infrastructure to inexpert users. To support scientists from diverse science domain to utilize edge and cloud computing resources to process and analyze real-world measurements, the project will explore how requests and questions from users asking fuzzy tasks can be specified to trigger actions at the edge and cloud using Language Models (LMs) and Multimodal models (MMs) to maximize user accessibility to an edge enhanced computing infrastructure for scientific research. This project will provide a baseline of user interact with edge-cloud infrastructure using human language to help understand how human language based services play a role in computing platforms for scientific research.

Project Objectives

• Topics: Fuzzy Trigger Edge-Cloud AI at Edge AI for Science
• Skills: Python, PyTorch, Linux, Containers
• Difficulty: Moderate
• Size: Large (350 hours)
• Mentors: seonghapark