Bard Enhancements: Google Sheets export, enhanced logic and reasoning skills

Google has unveiled significant updates to its artificial intelligence (AI) tool, Bard, aiming to enhance its mathematical capabilities, coding proficiency, and string manipulation skills through a new “structured, logic-driven” system.

Following the recent introduction of Gmail and Docs export functionalities, Bard now allows users to export generated responses, including table-based content, directly into Google Sheets. This new feature empowers users to swiftly transform generated data, such as creating tables for volunteer sign-ups at an animal shelter, into a format that is easily manageable and shareable within Google Sheets.

However, the most noteworthy update to Bard lies in its improved logic and reasoning skills. In March, Google implemented enhancements to enable Bard to answer math and logic-based queries more accurately. The company is committed to continuously refining and improving the quality of the responses generated by the AI system. As a result, Bard now incorporates a structured, logic-driven framework that enables it to recognize and effectively address computational prompts, such as calculating future population growth or determining prime factors. For example:

  • “The present population of the town is 3,888. What would be the expected population for the next two years and five years if we assume an annual growth rate of 20% for two years, followed by 30% for five years?”
  • “What are the prime factors of 15,683,615?”
  • “Reverse the word ‘Lollipop’ for me.”

In response to such computational queries, Bard leverages a combination of natural language generation and traditional code execution. By employing this more conventional approach to computation, Bard is capable of generating and executing code that solves the given problem or answers the question. The generated code is even included in the response and can be copied for further analysis or utilization. This methodology, although formulaic and somewhat rigid, enables Bard to deliver impressive results, such as solutions for long-division problems.

Google underscores the advantage of this approach over relying solely on large language models (LLMs) that predict the most likely subsequent words. In the realm of math problems, simply relying on LLMs often yields suboptimal results, as they tend to provide initial answers without performing the necessary arithmetic calculations.

The newly introduced system in Bard combines the power of LLMs with traditional code execution to enhance the accuracy of computation-based word problems and math queries by approximately 30%, as determined through rigorous testing on internal challenge datasets. Google, however, urges caution, as Bard may occasionally fail to generate code to assist with prompt responses, the generated code may contain errors, or the executed code may not be included in the response.

Google’s latest advancements in Bard aim to make the AI tool a more reliable and effective resource for users seeking assistance with mathematical tasks, coding queries, and string manipulations. With the addition of the Google Sheets export functionality and improved logic and reasoning skills, Bard continues to evolve as a valuable tool in various domains, facilitating problem-solving and enhancing productivity for its users.


FOLLOW US ON – Telegram, Twitter, Facebook and Google News

Leave a Comment