Mastering Knowledge Bases: The Key to Transformative Chatbots | by Stefan Kojouharov

So recently, we have been experimenting with knowledge bases, and how information is structured is vital.

In some sense, it directly relates to the context and overall meaning. It also makes it easier for LLMs to answer questions accurately and reduces hallucination.

And so we created a few experiments, that you can play with and test it yourself.

But first, why does Information Structure Matter?

Imagine a library. One room has books strewn everywhere, titles mixed up, and no discernible order. In another room, the books are organized by subject, then by author, then by publication date. Which room would help you find the book you want more efficiently?

LLMs are similar. The way knowledge is presented and organized dictates not only their understanding but also their output.

Organization of Informational Taxonomies and Hierarchies:

This considers elements like URL structures, folders, and how information is interrelated. By defining the proper context, you can highlight what’s critical.

Organization Within a Document:

Delving deeper, this looks at the composition of individual pieces of information — from structure and semantics to formatting and summaries.

Let’s dive into our findings on the first aspect:

At its core, informational hierarchy is about context. Whether it’s a URL on a website or the structure of folders within a system, hierarchies set the scene and help LLMs understand the importance and relevance of different data points.

Consider this:
– ChatbotConferences.com/conferences/2019/nyc suggests there are multiple events across different cities.
– ChatbotConferences.com/new-york-city offers just a city, which, out of context, is ambiguous.
– ChatbotConferences.com/nyc/2019 indicates multiple NYC events but omits a broader context.

The Great Hierarchy Test

We started on a quest to understand the weightage and importance hierarchies. We build two chatbots with their primary differentiation being the organization of their hierarchy:

– Bot 1: Trained on multiple pages, each representing a distinct event and year.

– Bot 2: Trained on a consolidated page that collates all the agendas.

>>>> The result? Dive in and test our ‘Good Bot’ and ‘Bad Bot’ for yourself here.<<<

Stefan Kojouharov is a pioneering figure in the AI and chatbot industry, with a rich history of contributing to its evolution since 2016. Through his influential publications, conferences, and workshops, Stefan has been at the forefront of shaping the landscape of conversational AI.

Current Focus: Currently, Stefan is channeling his expertise into developing AI agents within the mental health and wellbeing sector. These projects aim to revolutionize the way we approach wellness, merging cutting-edge AI with human-centric care.

Join the Journey on Substack: For exclusive insights into the development process, breakthrough experiments, and in-depth tutorials, follow Stefan’s journey on Substack. Join a community of forward-thinkers and be a part of the conversation shaping the future of AI in mental health and wellbeing.

Subscribe now to Stefan’s Substack and unlock the full potential of AI-driven transformation.