DENSO TEN Limited.
AI visual inspection to improve soldering quality and safety of electronic circuit boards.
- Use Case
- Manufacturing
(Left to right) HACARUS CTO Naoki Kitora with Mr. Takuya Kondo, Yanmar Holdings Co., Ltd. R&D Center
Our impressions of working with HACARUS is of a company that provides excellent data science expertise through each step of the process – from data processing to analysis of data. The project also provided an opportunity for our R&D institute personnel to get acquainted with Sparse Modeling through the in-house study sessions – knowledge that we will be using in our work moving forward
Background
The Yanmar Group, which manufactures and sells internal combustion engines, agricultural machinery, construction machinery and small vessels among other goods, established its Research & Development Center in February 2000, as the core research and development center for sustainable future with a special focus on food production and energy conversion.
In the research of AI for agricultural machinery, while aiming to improve the accuracy of its conventional machine learning based tools, Yanmar had reached a limit where further enhancements couldn’t be achieved with modifications and parameter adjustments – to explore new methods, Yanmar decided to try HACARUS Sparse Modeling based approach to increase performance.
Why HACARUS
HACARUS, which specializes in building AI that captures core features using Sparse Modeling based technology, was thought to be a suitable solution provider to present new ways to enhance results. The consulting model provided by HACARUS, which is flexible in nature also suited Yanmar’s needs.
Service Content and Results
After Yanmar initially shared the characteristics of the data, and the background of the research, HACARUS provided guidance on how to perform data analysis.
Due to the confidential nature of the subject matter, rather than directly performing the analysis of the data, HACARUS provided hands-on support for the work of the research institute.
The result? – Through analysis from various angles that weren’t considered before, Yanmar were able to discover new features that contributed to accuracy improvements, and obtained knowledge about features that have a high impact on the results.
This not only dramatically improved the accuracy of the machine learning models, but also provided new perspectives for future data acquisition and analysis. In addition, through the consulting, learning from HACARUS experienced data scientists, the research institute staff were able to improve their skills and broaden their perspective.