Jaxon AI teams up with IBM watsonx in battle against AI hallucination

Key Points:

  • Jaxon AI’s DSAIL technology, powered by IBM’s watsonx foundation models, aims to combat hallucinations and inaccuracies in large language models, offering a novel approach to developing reliable AI solutions.
  • The DSAIL approach involves converting natural language inputs into a binary format and subjecting them to verification processes, along with utilizing the RAG model to limit non-determinism and enhance trustworthiness.
  • IBM collaborates with partners through programs like IBM Build, providing access to watsonx, technical support, and go-to-market assistance to offer reliable AI foundation models.

Summary:

Jaxon AI, initially focused on building AI systems for the U.S. Air Force, is now venturing into the enterprise market with Domain-Specific AI Language (DSAIL) to address inaccuracies and hallucinations in large language models (LLMs). Their technology utilizes IBM’s watsonx foundation models and emphasizes creating more reliable AI solutions. The DSAIL approach aims to reduce hallucination by converting natural language inputs into a binary format and subjecting them to verification processes before generating an AI response. Jaxon uses the Retrieval Augmented Generation (RAG) model to address the hallucination problem, integrating it into their DSAIL approach. They also leverage IBM’s StarCoder model for code generation. IBM, a major player in generative AI and LLM technology, collaborates with partners like Jaxon AI through IBM Build to provide access to watsonx, technical support, and go-to-market assistance, aiming to offer reliable AI foundation models with consistent pricing and performance.

DAILY LINKS TO YOUR INBOX

PROMPT ENGINEERING

Prompt Engineering Guides

ShareGPT

 

©2024 The Horizon