Sumitomo Rubber, NEC advance AI-based tire, rubber R&D

Kobe, Japan – Sumitomo Rubber Industries (SRI) and NEC Corp. are accelerating their efforts to "build a globally competitive research and development platform," as outlined in their strategic partnership announced in July. (ERJ report)

Specifically, the two companies said they had identified key themes to be addressed by 2030, using NEC's artificial intelligence (AI) technologies and SRI’s R&D capabilities.

According to a 26 Nov release, SRI and NEC have been long-time partners working on “co-creation activities”, most recently advancing a tire development AI platform from 2022 to support tire design.

Under the new strategic deal, the two will further expand on the past co-creation activities, developing “advanced AI agents that incorporate the knowledge of SRI researchers and engineers.”

In its earlier stage, the partnership conducted pilot trials in tire material compounding prediction using “pseudo-quantum annealing technology.”

This, SRI said, involved analysing and extracting data about material properties from the company’s past experiments.

Next, based on these trends, NEC's pseudo-quantum annealing technology, which SRI said can “quickly derive optimal solutions from a vast number of complex combinations,” was used to search for candidate material types and blends that would satisfy the target properties.

The proposed compounds by the system were then compared with a rubber compound for premium tires previously developed by SRI.

The process, SRI said, confirmed that it was possible to “derive a rubber compounding proposal that satisfied more than 90% of the target properties.”

Furthermore, the time required to derive an equivalent compounding proposal could be reduced by 95% compared to the time required by an unskilled person.

SRI said it expects that the use of this technology will enable the efficient development of rubber compounds for all tire categories, from standard to premium tires, regardless of the user's level of expertise.

Another research theme included discovering new materials using AI agents and a materials-exploration solution, for which SRI said the companies achieved “significant results”.

The Japanese group explained that the search for advanced new materials requires “researching vast amounts of literature across multiple fields.”

Furthermore, "discovering" beneficial materials and their combinations from across different fields is “extremely difficult.”

In their demonstration, NEC’s AI experts collaborated with SRI’s material developers to explore new materials for premium tires, which require advanced performance.

They extracted the thought processes and implicit knowledge involved in material exploration and built an AI agent that was trained on these knowledge, SRI explained.

The AI agent then utilised a materials exploration solution that combined generative AI and graph-based AI to narrow down candidate materials.

Specifically, verification was conducted using as a model the rubber material for premium tires whose surface softens upon contact with water – something that SRI said was difficult to develop using conventional knowledge alone.

As a result of the demonstration, the AI agent aggregated and analysed materials-development knowledge from publicly available literature in various languages, said the group.

It also "autonomously" collected deeper related information, thereby expanding the search range for candidate materials.

According to SRI, compared to manual work – which often required rework due to failure to reflect important requirements or limited search scope – the time required for search was reduced by 60-70%.

“This confirmed that the system can improve both accuracy and comprehensiveness, enabling the efficient discovery of new material candidates that would have been difficult to identify using conventional methods,” SRI added.

As part of the strategic partnership, the two companies will expand their activities, including these two pilot projects and other technologies, with the aim of establishing an advanced AI-driven R&D platform by 2030.

The partners also aim to create new businesses and innovations to address social issues.