Bifrost AI & NTT Data: A Case Study

Building bridges between AI startups and Japanese corporations is incredibly rewarding. The collaboration between Bifrost AI and NTT Data is an example.

In my article for Yomiuri Shimbun’s The Japan News, I share how a dinner at a Tokyo izakaya led to a collaboration that’s tackling real-world AI challenges. 

Read the full background and story in my article here.

Project Impact & Results

NTT Data, one of the world's largest technology companies and a recognized leader in AI, published this article detailing how they use the Bifrost AI solution. This collaboration tackled a core challenge we hear all the time: the scarcity of high-quality data for AI training and its high cost.

The Goal:
As stated by NTT DATA, “The goal of the PoC was not just to assess potential improvements in model performance and robustness, but also to determine if it is possible to reduce the reliance on real images—often limited and costly—while still obtaining high-performing models.”

The Results:
- Bifrost AI’s synthetic data increased mean average precision (mAP) by almost 10 percentage points for the object detection use case.
- NTT DATA also highlighted the ease of use of the Bifrost AI platform.

Read the full article here.