RagaAI Secures Funding to Develop Tool for Diagnosing and Fixing AI System Flaws

RagaAI Secures Funding to Develop Tool for Diagnosing and Fixing AI System Flaws

RagaAI zabezpiecza finansowanie by rozwijać narzędzie do diagnozowania i naprawy wad w systemach sztucznej inteligencji

RagaAI, a startup founded by Gaurav Agarwala, an alumnus of companies like Nvidia Corp. and Texas Instruments Inc., has received $4.7 million in funding for the development of a tool that aims to diagnose and fix flaws in artificial intelligence (AI) systems. The seed round investment of $4.7 million was led by pi Ventures, with participation from Anorak Ventures and TenOneTen Ventures.

RagaAI, based in Silicon Valley, uses foundational models to detect and eliminate AI-related issues such as hallucinations and the failure to consider real-time critical information. The founder, 44-year-old Gaurav Agarwal, shared his “near-death experience” while testing a semi-autonomous vehicle on a California highway. The AI system failed to detect obstacles on the road, forcing him to manually brake and avoid an accident. Agarwal emphasized the urgent need to prevent such situations, especially in advanced AI applications such as cancer detection, aircraft maintenance, and AI-based recruitment tools.

RagaAI is another startup aiming to address the needs arising from the global development of artificial intelligence. Currently, most efforts are focused on building models and applications, with less attention given to diagnosing and testing errors. The RagaAI platform offers over 300 tests that help businesses identify problems and find their root causes. The tool is capable of detecting issues such as flawed data labeling and biased influence. It also aids in detecting hallucinations – false and misleading information presented as fact – and deliberate attacks aimed at inducing model malfunctions.

The funding secured by RagaAI will be used for research and expanding the team of engineers, currently primarily based in Bangalore. The company is already working with several major clients in the e-commerce, aerospace, and medical imaging industries, although specific names have not been disclosed. While the platform has already reduced failures by 90%, no further details have been provided on this matter.

Frequently Asked Questions (FAQ)

The source of the article is from the blog combopop.com.br