Master’s Thesis at KSRI and KIT: Techniques for Overcoming Data Abundance

August 1, 2020

The Challenge: Machine Learning on Small & Distributed Data Sets

In his master’s thesis with the Karlsruhe Service Research Institute (KSRI), prenode team member Michael Jahns explores techniques for overcoming data abundance. He researches how companies can implement machine learning for use with small and sensitive data sets. In the project, he evaluates different ML techniques with regard to prediction performance and computational effort. In his conclusion, he recommends users to implement Sequential Transfer Learning or Federated Transfer Learning. Both of these approaches show high potential for learning across distributed data sets while increasing data privacy and security.

Read more in the abstract below.

Abstract

“For Machine Learning (ML) technologies based on Neural Networks, a higher amount of data for training generally leads to better results. When ML is used in a production environment, companies often cannot exploit the full potential out of it as only small data sets are available to them. However, if many companies train similar models with the same target, they could benefit from collaborating by exchanging and aggregating their data to enlarge the amount of training data. Such consolidation may be prohibited by law or contracts for data confidentiality.

A structured literature review is conducted to select suitable techniques that enable this cooperation. To this end, we identify Federated Learning (FL), Federated Transfer Learning (FTL), Sequential Transfer Learning (STL) and data generation with Generative Adversarial Networks (GANs) as potential solutions. These techniques are implemented in five different use cases where data is distributed, but similar models need to be trained. Each use case consists of multiple separated data pools and the data is preprocessed identically for each technique. The trained models differ between the use cases, but not between the techniques, thus allowing a direct comparison of the prediction performance. The results are evaluated and compared for prediction performance and computational complexity across the five use cases.

Based on the prediction performance, STL proves to be the most promising technique closely followed by FTL. FL and the data generation technique cannot achieve such improvements. Inspecting the computational complexity, the STL technique performs worse than the other techniques. The result of this thesis is a suggestion for users as to what technique they should implement based on their data set properties to increase the prediction performance on a distributed data pool.” (Jahns, 2020)

Interested in writing your thesis at prenode? Contact us

Video: Microsoft Intelligent Manufacturing Award 2023 | Meet the winners

Consent required! Click here to view the video and accept all necessary cookies.

Share this article

Linkedin IconTwitter IconFacebook Icon

Read more

September 3, 2024

prenode honored as top AI startup for the fourth time

In a comprehensive analysis of over 1000 AI startups, carried out by appliedAI Institute for Europe gGmbH in collaboration with Deutsche Telekom, NVIDIA, Intel, UnternehmerTUM and eleven renowned venture capital companies, prenode was named one of the 687 most promising AI startups in Germany.

Read more

April 18, 2024

Factory-X: The digital ecosystem for manufacturing — and prenode is on board!

Since February 2024, Factory-X has been driving forward the digitization of German industry. Based as part of the Manufacturing-X Initiative, which is funded by industry, politics and science under the umbrella of the Federal Ministry of Economics and Climate Protection (BMWK), Factory-X aims at the secure and efficient exchange of data throughout the production and supply chain.

Read more

August 4, 2023

prenode continues to shine in the German AI Startup Landscape for the third consecutive year

We proudly announce that it has again secured a spot in The German AI Startup Landscape for the third time. This recognition places prenode as one of the 508 most promising AI startups in Germany for the current year.

Read more

May 19, 2023

prenode and TRUMPF Win the "SCALE!" Category of 2023 Microsoft Intelligent Manufacturing Award (MIMA)

prenode, in partnership with TRUMPF Werkzeugmaschinen SE & Co. KG, has unveiled an innovative technology that empowers live monitoring and remote troubleshooting for Equipment-as-a-Service (EaaS) in the sheet metal industry.

Read more

April 11, 2022

prenode at the AI German Startup Landscape 2022

We are very happy to announce that prenode has been listed for the second time on the AI Startup Landscape 2022 (in the section platform) as one of the 304 most promising AI startups in Germany.

Read more

February 8, 2022

prenode is Member of the German AI Association

The German AI Association connects the most innovative AI and deep tech companies with industry partners and politics. It is the largest AI network in Germany with more than 350 members.

Read more

December 29, 2021

prenode is on the Industrie 4.0 Map

We are honored to be part of the Industry 4.0 map once again in the area of "Analytics/Efficiency".

Read more

December 13, 2021

Robin Hirt guest at Podcast of the Technologiefabrik Karlsruhe

In episode #25 of the Tech + Founder Podcast of the Technologiefabrik Karlsruhe, prenode CEO Dr.-Ing. Robin Hirt talks about AI in mechanical engineering.

Read more

October 14, 2021

Robin Hirt presents prenode at VDMA STARTUP THE FUTURE 2021

Our CEO Robin Hirt participated in the digital event START-UP THE FUTURE of the VDMA.

Read more

July 26, 2021

prenode receives Award as AI Champion Baden-Württemberg 2021

We are very proud to be recognized as an AI Champion in the category of small and medium-sized enterprises by the Ministry of Economics of Baden-Württemberg!

Read more

May 1, 2021

Federated Learning on Medical Data: Collaborative Research Project with ZEISS

In a joint effort with researchers from the ZEISS Group and the Carl Zeiss Meditec AG, we applied our decentralized approach to carry out anomaly detection in OCT images taken of eyes.

Read more

February 3, 2021

Academic Activities at KIT 2021: Lectures on Digital Services and Artificial Intelligence

prenode is actively involved in academic lectures at the Karlsruhe Institute of Technology (KIT). Among others, prenode co-CEOs Robin Hirt and Ronny Schüritz are contributing know-how and practical insights to lectures at the Institute of Information Systems and Marketing (IISM) and the Karlsruhe Service Research Institute (KSRI).

Read more

February 1, 2021

Partnership: Cybus & prenode

With the support of prenode, manufacturing companies can use services based on artificial intelligence (AI) without compromising data protection. Thus predictive maintenance, automatic anomaly detection, or automatic quality control become feasible.

Read more

January 10, 2021

Winner of AI Competition BW: Predictive Maintenance for WEISSER Precision Turning Machines

We are proud to announce that our proposal was accepted as part of the innovation contest by the Ministry of Economics BW. Together with WEISSER Group and the Hahn-Schickard Gesellschaft, we are developing a predictive maintenance solution for WEISSER precision turning machines.

Read more

January 5, 2021

Publication at HICSS 2021: Hirt et al. on Sequential Transfer Learning in Networks

In the recently published Proceedings of the 54th Hawaii International Conference on System Sciences (HICSS), prenode’s co-CEO Robin Hirt explored the possibilities of reusing neural nets on distributed data sets.

Read more

October 16, 2020

Robin Hirt gives Keynote at the InnovationFestival @ karlsruhe.digital

Our CEO Robin Hirt gave a keynote speech at the Innovation Festival Karlsruhe on the topic of "Data protection and artificial intelligence: Is that even possible?"

Read more

October 6, 2020

prenode wins the Cyber Chapions Award 2020

We are proud to have won first place at the CyberChampions Award 2020!Our idea convinced the jury in the final so that we succeeded in winning the INIT Innovation Award and a prize of €5000.

Read more

June 1, 2020

Institute Cooperation: Research Projects with Critical Information Infrastructures Group

We believe that new approaches and technologies explored in academia enable new opportunities in the field of AI and machine learning.

Read more

January 1, 2020

prenode as Research Advisor: Activities in Advisory Councils of Cross-Domain Projects

prenode actively contributes know-how on machine learning and AI development in various research projects. Through our position on the advisory boards of two collaborative projects, we can exchange cross-domain knowledge. This enables us to provide expertise on tech topics and learn about the needs and requirements of the involved partners.

Read more

October 1, 2019

StartUpSecure Initiative: Research Project Supporting the National IT-Strategy

prenode is proud to be among the selected projects of the StartUpSecure program. The initiative aims to support research teams that focus on cybersecurity topics. StartUpSecure was launched as part of the national IT security strategy and is led by the Federal Ministry of Education and Research.

Read more