SECOR Chatbot – SECOR – Exchange market for semiconductors

The SECOR CHATBOT uses AI to recommend which commodity semiconductors are best suited for a component.

Today, it is usually difficult for an employee who creates a specification sheet or develops a component to decide which semiconductor would be the right one for the component in question.

The SECOR CHATBOT provides AI-based support to the automotive industry in selecting the right commodity semiconductor for a specific component, thereby increasing RESILIENCE compared to single-source ASICs (application-specific integrated circuits).

In future, this employee will ask the SECOR CHATBOT a short question, e.g. “Which standard semiconductor should I select for the next ESP control unit for the Golf successor?”

The SECOR CHATBOT uses the SECOR PLATFORM and AI-based training to generate a list of commodity semiconductors from manufacturers’ current offerings. This is made possible by federated learning, an AI method. It was developed to process personal data, in particular disease data (e.g. tumours or skin cancer). It has achieved great success in the medical field, one example being Auicura medical.

The SECOR CHATBOT provides AI-based support to the automotive industry in selecting the right commodity semiconductor:

WHY SECOR chatbot?

KI Knowledge Exceeds the Knowledge of Most Employees

The CHATBOT uses the SECOR PLATFORM with all offered commodity semiconductors, including technical data sheets (currently automotive focussed), and the training BOMs (parts lists) to learn automatically. After each phase of the learning process, the CHATBOT becomes “smarter”. The more companies join in and provide data protection-compliant parts lists (BOMs) for components, the faster the CHATBOT will cover all components in the vehicle.

Faster Integration of New Employees in the Semiconductor Sector

The shortage of skilled labour in the semiconductor sector is present everywhere. New employees generally require long training periods in order to be able to support component-related semiconductor decisions. The SECOR CHATBOT usually has more knowledge than an average employee in this field. It therefore supports the day-to-day work in the specialist area.

The Main Features of SECOR Chatbot

The technical core of the SECOR CHATBOT is AI, in the rarely used form of federated analytics or federated learning. Successfully used for skin cancer detection, as it can process personal or secret information in a data protection-compliant manner within the framework of machine learning.

Parts lists from participating companies are analysed based on AI and the SECOR PLATFORM in compliance with data protection regulations. The result in the form of the CHATBOT is only available as a service to the participating companies. The AI uses federated learning to combine the BOMs of existing components with the data sheets of the commodity semiconductors currently on offer in a data protection-compliant manner.

In order to further optimise this concept, the TUM Technical University of Munich (www.tum.de) is interested in a new research project for a specific assembly / material group.

TUM is supporting the joint process from the scientific side. The research project is proceeding in phases. In the next phase, planned for summer 2024, the focus will be on all important components of the electric drive of a BEV vehicle.

Discover How Company Analyses Can Change the Way You Work

The SECOR CHATBOT is currently only available to companies that have participated in at least one CHATBOT training phase.

Together, the participating companies benefit from the fact that the SECOR CHATBOT becomes smarter and more efficient from phase to phase based on AI. Through federated learning, the company-specific parts lists are processed in a decentralised manner in order to train the CHATBOT in a decentralised manner without the data being exchanged in any form between the participating companies and SECOR.

Previously, these companies only had their own employees’ expertise and their own BOMs from the past as a source for decision-making. In future, these companies will benefit from the expertise of the other companies without having to disclose their business secrets.

Take the Next Step

Contact us, we are planning the next phase of the research project with the TUM School of Computation, Information and Technology (Technical University of Munich). We are looking for at least one OEM and Tier-1. Together we will define a specific assembly / material group. As part of the research project, we will then train using the federated learning method and then the improved version of the SECOR CHATBOT will be available to all participating companies (including predecessor companies). The project preparations will start in May 2024.