Logo
Logo
HOME
this is alternative text
news-and-events
Harnessing the Power of AI: Unveiling Stellaris GPT's Advanced Capabilities in Real-Time Context Learning and Native Safety

Posted on: Fri Dec 01 2023

This is an image.

TOPIC
Sentiment Influences on Sentiment with Discussion in Football Subreddits

 

DATE AND TIME
Thursday, December 7, 2023, 2:00 p.m. - 3:00 p.m.

 

ABSTRACT
In this webinar, the speakers will discuss the groundbreaking achievements of Stellaris AI's flagship language model, Stellaris GPT (SGPT), including the innovative technologies and strategies that position SGPT at the forefront of the artificial intelligence industry. The session will explore SGPT's Real-time Context-Learning (RCL) capability, a pioneering approach that leverages advanced online learning techniques, such as Approximate Learning and Simultaneous Knowledge-Weights Editing. This unique feature enables SGPT to provide in-depth Q&A with the latest real-time information, surpassing traditional limitations of pre-trained datasets.

The webinar will also highlight SGPT's computational efficiency, demonstrating how it utilizes computational resources 10 times more efficiently than the industry norm. This efficiency not only enhances training flexibility but also signifies a leap forward in large language model (LLM) technology.

A key focus of the presentation will be on SGPT's Native-Safety Technology, which prioritizes the security of large models from the ground up. We'll discuss how safety is incorporated into SGPT's architecture, transforming it into a model that is 'Born to be safe'. This intrinsic safety mechanism defends against advanced model hacking, jailbreaking, and other attack methods, eliminating the need for subsequent patch-based moderation.

Lastly, the session will address SGPT's full local deployment feature, emphasizing a commitment to data protection and security. This webinar is designed for professionals, researchers, and tech enthusiasts keen on understanding the future of AI and the role of advanced large language models like SGPT in shaping that future.

 

SPEAKER BIOGRAPHY
Dr. Jacob Wu
Chief Executive Officer, Stellaris AI
Dr. Wu is an accomplished serial entrepreneur who, at the early age of 23, earned his Ph.D. in Computer Science from the University of Hong Kong. His academic focus on algorithm design, large-scale data analytics, and machine learning led him to undertake a substantial project in AI-based cancer treatment target prediction, establishing a strong foundation for his career in the field of AI solutions.

 

Dr. Liam Ning
Chief Technology Officer, Stellaris AI
Dr. Ning is an accomplished leader with a prodigious track record in the field of High-Performance Computing (HPC) and Artificial Intelligence (AI). Combining deep technical mastery with an extensive range of business experience, he brings a unique perspective to the table in both AI technology and commercialization.

 

ABOUT THE WEBINAR SERIES
This is the 4th session of a 5-part webinar series featuring leading academic and industry experts in risk management, artificial intelligence, information systems, big data analytics, machine learning, and large-scale data processing. Students interested in the CBSA-Wisers Analytics Challenge @HKUST 2021/22 are highly encouraged to attend the webinars.

NEWS
CBSA Director Prof. Yi Yang Showcases LayerZero AI at HKUST Unicorn Day 2025
Posted on: Mon Jun 16 2025

Prof. Yang highlights enterprise AI breakthrough at HKUST Unicorn Day, showcasing LayerZero AI's custom embedding technology.

Read more
NEWS
HKMA and Cyberport Launch Second Cohort of Gen A.I. Sandbox to Accelerate A.I. Innovation in the Financial Sector
Posted on: Thu May 01 2025

Prof. Yi Yang, Director of the Center, joins industry leaders in a panel discussion on cutting-edge GenAI risk management strategies during the launch of the second cohort of the Generative AI (GenAI) Sandbox initiative held on 28 Apr 2025.

Read more
NEWS
From Concept to Pitch: HKUST, IFEC & Cyberport Celebrate Final Round of Web3 Ideathon for Tertiary Students
Posted on: Mon Mar 31 2025

At the closing Pitch Fest of the first-ever Web3 Ideathon, top tertiary student teams showcased their groundbreaking Web3 concepts to industry experts.

Read more
CONTACT US
Center for Business and Social Analytics
Rm 6061, Lee Shau Kee Business Building, HKUST
+852 3469 2637
cbsa@ust.hk