3 Reasoning engines
With recent advancements in AI, search technology made a leap forward. It evolved into reasoning engines that can understand and process complex queries in a more humanlike manner. Reasoning engines can provide more accurate and contextually relevant answers to make search experiences more intuitive and efficient.
In this video, you explore how these advancements transformed search engines into powerful tools that not only find information but also understand and reason with it:
Download this video clip.Video player: Video 3
Transcript: Video 3
Voice-over:
Most of us are familiar with search engines like Bing.
We type in a query and instantly we are presented with a list of web pages that might have the answer we are looking for.
But there is also another type of engine that works differently than a search engine, called the reasoning engine.
A reasoning engine is a system that applies logic and inference to draw conclusions, make decisions, summarise information, or solve problems based on data and knowledge.
Reasoning engines offer more elaborate responses compared to traditional search engines.
They can assist with tasks like writing parts of a grant proposal, providing ideas for a project, researching a topic, or analysing data in a file.
They can even generate images from text descriptions.
So how does a reasoning engine work? A reasoning engine starts with a foundation of knowledge.
It is fed a variety of data, from text to code, and it might even absorb information across the web.
This is considered the pre training phase, where the engine learns language patterns, grammar and facts.
During the next phase, the engine relies on the guidance of human supervisors to steer its learning journey.
These supervisors guide the engine through the complexities of language, ensuring it understands not just the what, but the why behind the words. When you ask a question, the reasoning engine springs to life.
It sifts through its learned knowledge, weighing words and considering context.
It is not just about finding an answer, but it is about creating an answer that is informative, relevant and engaging.
The journey also does not have to end with one query or response.
With every interaction, the reasoning engine learns and adapts to answer your question based on the prompt you give it.
It is a continuous cycle of learning and evolving, striving to understand and communicate more effectively.
As we continue to explore and innovate, reasoning engine experiences like Microsoft Copilot are opening new possibilities for user creativity in generating text, images and other media.
They are not just changing the way we search for information, but also how we interact, learn and create in the digital world.
Video 3
Interactive feature not available in single page view (see it in standard view).
It’s no longer just about finding an answer; it’s about creating responses that are informative, relevant, and engaging. Reasoning engines don’t just provide answers. They understand the context and nuances of your queries to deliver more personalised responses.