#ChatGPT4 by
@OpenAI
VS
@ChatDKG by
@origin_trail
With generative
#AI models taking over pretty much all pores of our lives, it is worth comparing the performance of these two systems tackling the same question. The results are INTERESTING:
When asking ”What was the composition of
@Nike's global corporate leadership in terms of gender in 2021?”, the answer
#ChatGPT provides is vague, rather long and pretty much does not answer the question.
On the other hand,
@ChatDKG provides a concise and factually precise answer, based on verifiable source of knowledge discoverable within the
@origin_trail Decentralized Knowledge Graph (DKG).
Link:
bit.ly/3UAGDRe
Enter Decentralized RAG
The key to using AI without unwanted hallucinations and with greater precision is Decentralized Retrieval-Augmented Generation (RAG). RAG is an AI framework for retrieving facts from an external knowledge base to ground large language models (LLMs) on the most accurate, up-to-date information and to give users insight into LLMs' generative process.
In a 2020 paper,
@Meta came up with a framework called retrieval-augmented generation to give LLMs access to information beyond their training data:
👉
arxiv.org/abs/2005.11401v4
And to make RAG even more transparent, verifiable with ensured information provenance, and most importantly, inclusive by allowing more participants to publish and connect knowledge,
@origin_trail DKG provides a foundation for the Decentralized Internet to power even more capable, efficient and precise AI solutions.
Going back to our question ”What was the composition of Nike's global corporate leadership in terms of gender in 2021?”,
@origin_trail establishes the exact lineage of information and allows you to discover even more knowledge related to the topic, driving further exploration.
Link:
bit.ly/495u0C4
In our case, the original source may be tracked down to the
@Wikirate open data platform providing open access to sustainability reports corporations publish annually.
Further explorations into the knowledge source reveals, that the answer provided by
@origin_trail Decentralized RAG - "In 2021, the composition of Nike's global corporate leadership in terms of gender was as follows: - Female: 43.0% - Male: 57.0%" - indeed corresponds to the original knowledge source - FY21
@Nike, Inc. Impact Report.
@origin_trail DKG has been most commonly used to successfully drive trust and transparency within enterprise knowledge exchange within spectrum on industries
origintrail.io and is now entering an exciting new phase of connecting the global wealth of knowledge that will drive Decentralized RAG for a more precise and inclusive AI.