CSE research on IVA's 100 list

IVA has published this year' list of 100 "current research projects with the potential to create benefits, through commercialization, business and method development or societal impact". Four spinoff companies from the Department of Computer Science and Engineering are on the list.​

Dpella Alejandro Russo

Using and sharing data analytics based on personal data can create competitive advantages and lead to new business opportunities for companies and organizations. However, the possibilities are limited due to rules regulating the processing of personal data, such as the GDPR.​ Dpella has a business concept based on results from the SSF-funded project Octopi​, and was on the 100-list also last year. 
"During the past year, with the support of MobilityXlab and GU Ventures, we have, among other things, started an innovation project together with Ericsson to explore privacy-preserving analyzes" says  Alejandro Russo, professor in the  Computing Science division. 

EmbeDL Devdatt Dubhashi

EmbeDL is based on research performed in the EU-projects LeGaTO and VEDLIoT, as well as the Vinnova-funded project Sharpen​
"Our solution is an optimization engine that automatically tailors a state of the art neural network model to be deployed on a specific hardware so that it meets the customer’s requirements and specifications. We will bring the power of AI into the hands of people who would not have been able to even conceive of deploying the vast resources that the BigTech companies command" says Devdatt Dubhashi​, professor in the  Computing Science division. 

Repli5

Repli5​ works with the generation of digital twins and synthetic data. A digital twin is a digital replica of objects and environments in reality which can be used to very effectively test and verify AI systems without having to costly, slow and error-prone processes in the real world. Vinnova has funded development of the technique through the projects Sharpen and Realsim. 
"The potential is to replace most of the need for real-world data with synthetic data generated from digital twins. In this way, companies can make a big leap in efficiency, cost reduction and quality of solutions" says Devdatt Dubhashi.

ZeroPoint Technologies Per Stenström

ZeroPoint Technologies​ is a commercialization project linked to the European processorinitiative
Since 1995 professor Per Stenström and his colleagues in the Computer and Network Systems division have been working to develop new principles for how computers can work faster despite limitations in computer memory. One of the most striking principles is about memory compression, and will provide faster, cheaper and more climate-adapted IT systems in the future. It is this achievement that is further developed in the company ZeroPoint Technologies.​​​​​​​​​​



Page manager Published: Tue 10 May 2022.