Microsoft has been upgrading its artificial intelligence abilities by adding what was described as an "algorithm of thoughts." The ending result reportedly gives the AI the ability to reason like a human being.
Microsoft Adds Introduces its 'Algorithm of Thoughts' Tech Aimed to Upgrade AI with Human-Like Reasoning Capabilities
According to the story by Decrypt, Microsoft has just revealed a new training method called the "algorithm of thoughts," also known as the AoT, which is designed to help improve artificial intelligence by giving it the ability to make human-like reasoning.
This type of technology was designed for large language models like the popular ChatGPT, which the company is already using. A PDF was released explaining the "algorithm of thought" with the heading also noted that it was designed for "enhancing exploration of ideas."
The Company's Next Step is to Invest Heavily in AI Models Like Those of OpenAI Including ChatGPT and Dall-E
It was noted that this new approach is going to be the company's next step moving forward as it plans to invest heavily when it comes to AI, particularly with the creator of ChatGPT, OpenAI. The latter is also known as the creator of another powerful AI model, Dall-E.
So far, Microsoft has highlighted how its AoT technology has the potential to change the game by providing a more "streamlined problem-solving path." The approach would focus on "in-context learning," which gives the artificial intelligence to follow an organized manner to look for different solutions.
The Result of the Technology Gives AI Models Problem-Solving When It Comes to Less Resource-Intensive Approach
The main result of the new technology would make these artificial intelligence models capable of problem-solving in a less resource-intensive approach. The paper states that Microsoft's technique reportedly performs better compared to single-query methods and is getting closer to the approach of multi-query methods with large tree searches.
Microsoft also highlighted how their results suggest that giving a model instructions regarding the algorithm can result in performance better than its own algorithm. According to researchers, artificial intelligence will get better "intuition" from the technique when it comes to the AI's search process.
AoT Fixes Problems Compared to Limited In-Context Learning Techniques Compared to the 'Chain of Thought' Approach
As also reported by The Coin Republic, it was also noted that the AoT method fixes the problem when it comes to having limited in-context learning techniques which can be seen in the particular approach called the "chain of thought" or COT.
The CoT approach can come up with incorrect intermediate steps resulting in AoT assisting the model when it comes to using algorithmic examples in order to come up with better and more reliable answers.
Read Also: US Air Force Seeks $6 Billion AI-Powered Drone Fleet Amid Geopolitical Pressures
AoT is Looking to Give LLMs the Ability to have Dual Facets When It Comes to Augmenting Reasoning
This AoT approach comes from humans and machines and can be used to improve the performance of generative AI. Although humans are much better when it comes to intuitive cognition, algorithms follow a more organized path.
The research paper also reveals that AoT is looking to give LLMs the capability of having dual facets in their abilities when it comes to augmenting reasoning.
Related Article: ChatGPT Faces Major Outage Affecting Users, OpenAI Rushes to Restore AI Chatbot
