MONTREAL--(BUSINESS WIRE)--Jul 24, 2023--
Over the last five years, AI Redefined has pioneered Reinforcement Learning with Human Feedback (RLHF), deployed award-winning human-AI platform Cogment enabling real-time human-in-the-loop fine-tuning for decision optimization AI agents, and is now employing that same expertise to address pressing Large-Language Model (LLM) needs. As corporations worldwide seek to embrace the transformative power of LLMs, this new offering on the market will greatly lower the barrier of entry when it comes to time, resources, and data requirements, while increasing accuracy, control, and trustworthiness.
This press release features multimedia. View the full release here: https://www.businesswire.com/news/home/20230724107785/en/
Human-AI collaboration product offers any organization a uniquely fine-tuned hyper-personalized, proprietary LLM that reflects their in-house subject-matter expertise. Fast, efficient and frictionless. (Photo: Business Wire)
While this solution is groundbreaking, its foundation relies on RLHF, which CEO of Open AI Sam Altman has credited as one of two key elements that transformed GPT3.5 into wildly successful ChatGPT and GPT4.
AI Redefined’s breakthrough in fine tuning comes from Cogment's unique ability to combine state-of-the-art AI technology with human expertise to frictionlessly leverage organizations' in-house subject matter expertise in new ways. The platform enables humans to interact with AI inside their environment, which uniquely allows AI to learn from implicit feedback (i.e. indirect data gathered from the user’s actions), the standard explicit (direct) feedback, and benefit from continuous expert-driven guidance, all in real time. The result is more precise, personalized, and efficient fine tuning in any enterprise context, thanks to Cogment's integration support for a wide range of tools used in the LLM ecosystem.
Unlike other solutions that rely on static training data, Cogment's active human-in-the-loop fine-tuning approach ensures an up-to-the-minute system that won’t depreciate over time due to drift and will in fact keep getting better. Using Cogment's fine tuning tool, companies can easily realign their fine tuning process as their needs and methods change, allowing for greater agility, resilience, and accuracy of the AI at all stages.
With faster results and more efficient fine tuning, companies can save both time and money as they reap the benefits of a hyper-personalized proprietary LLM that reflects their in-house subject-matter expertise.
CTO of LAVO Digital Michael Henson said, "We're one of the first organizations to trial Cogment's LLM fine tuning tool, which we'll employ to fine tune high-ROI planning and conversational recommendation agents. Our experience with AI Redefined and Cogment gives us the confidence we seek to ensure our optimizations deliver for our customers, and that our residential personalization offerings will scale as we require."
“Productivity gains ensure that every Fortune 5000 company will soon have a specific-to-the-entity LLM, derived from their own Content Management System and open-source tools. AI Redefined's platform enables real-time, continuous fine tuning to ensure accuracy, industry-specific terminologies, company-specific tone, all via an extremely modest amount of initial data needed," according to Craig Vachon, CEO of AI Redefined.
“Over the last five years, AI Redefined successfully targeted the orchestration of decision optimization algorithms in even the most complex industry contexts. Adding Cogment language fine tuning as a new offering to target LLMs is a natural extension of AI Redefined’s mission in human-AI alignment.” Vachon continued.
To learn more about Cogment and how it can transform your fine tuning process, visit www.ai-r.com/finetuner
View source version on businesswire.com:https://www.businesswire.com/news/home/20230724107785/en/
For media enquiries: Marie-Renée Brisebois, AI Redefined (AIR),contact@ai-r.com, (514-402-8108)
KEYWORD: NORTH AMERICA CANADA
INDUSTRY KEYWORD: SOFTWARE TECHNOLOGY ARTIFICIAL INTELLIGENCE OTHER TECHNOLOGY
SOURCE: AI Redefined
Copyright Business Wire 2023.
PUB: 07/24/2023 12:33 PM/DISC: 07/24/2023 12:31 PM
http://www.businesswire.com/news/home/20230724107785/en
