What is IJCAI-PRICAI?

IJCAI is a top conference in the field of the Japanese Society for Artificial Intelligence, where researchers in various fields of artificial intelligence (AI) gather, and PRICAI is a conference launched to promote the use of AI in the Pacific Rim countries.
In 2020, the 29th IJCAI and the 17th PRICAI were jointly held as IJCAI-PRICAI2020.

<Overview>
Name: IJCAI-PRICAI2020
Date: Tuesday, January 7, 2021 to Friday, January 15, 2021
Venue: Online
Number of participants: 1,972

IJCAI-PRICAI2020 held online!

Initially, it was scheduled to be held in Yokohama, but due to the impact of COVID-19, it was held online.
It was a virtual venue that could be accessed online, so I was able to look around the venue without going to the venue, which is an advantage of holding it online♪

A view of the main venue. The scenery of Yokohama, which was scheduled to be held, was spreading!

Macnica will also be exhibiting at the sponsor booth!

First lecture at the industrial session!

IJCAI-PRICAI2020 was held over nine days.
At the Industrial Session on Thursday, January 14, 2021, Macnica 's AI expert group  Members of the AI Research & Innovation Hub (ARIH) gave speeches.

Theme: Tool Degradation Estimation with Ensemble Learning
Speaker: AI Research & Innovation Hub
Takahiro Kusunoki (Principal), Mary Grace Malana (Data Scientist), Jerelyn Co (Data Scientist)

The theme is "Estimation of Tool Deterioration by Ensemble Learning"

Various tools that support the manufacturing process.
Estimating these deterioration states is very effective in reducing "maintenance costs" and "production costs."

In this joint research between Mitsubishi Heavy Industries and Macnica, we worked on estimating tool deterioration based on vibration data collected during machining.
However, when we tried learning, we found that the amount of data was small and we could not obtain the desired estimates using only a single learning device.

That's why we focused on ensemble learning.
By selecting features using signal processing and strategically combining the results of multiple learning devices, we succeeded in performing accurate degradation estimation even with a small amount of training data!

In the actual lecture, we will tell you the details of the above research contents.
*A subscription is required to view the video (English). Click here to apply ↓

in conclusion

IJCAI-PRICAI is usually held in each country, but I think that more people participated because it was held online.
We participate in various academic conferences every year and publish research columns of papers. If you are interested, please take a look at the other columns ♪