Marketable innovations emerge when research results extend beyond the laboratory

The development of self-driving cars is perhaps the best-known example today of how research and practical application can be closely linked. These systems are based on complex artificial intelligence, sensors, data processing and real-time decision-making — scientific advances that have moved from the theoretical to the practical, being directly integrated into everyday technology.

A self-driving vehicle can detect hazards, including an unexpected pedestrian, much more accurately by relying not only on its own sensors, but also by accessing images from nearby CCTV cameras. This approach can also be applied in the energy sector, where the technology can optimise energy distribution and predict demand, significantly reducing losses.

These developments also highlight the need for a change in approach within the research community. Discovery research remains essential, but the practical exploitation of scientific results and their translation into market value are now key drivers of the Hungarian economy.

"The creation of HUNRENTECH also serves this purpose. It aims to identify promising intellectual creations, support their patenting and exploitation, and find suitable partners for their development," Roland Jakab, CEO of HUN-REN, told Portfolio.

Hungarian research and innovation is going global

HUN-REN employs over 3,500 researchers who contribute to science every day, generating ideas that could deliver significant social and economic benefits. However, the research network has not yet established a technology transfer company to exploit research results, and researchers have lacked the business acumen to turn innovations into marketable products or services. Consequently,

scientific innovations often gained value and were produced in other countries rather than in Hungary.

'We need to think in terms of value chains. When we discover something, we must immediately start considering how it could be useful for the country. It's crucial that we have people who can identify potential applications and sales opportunities, obtain patent protection, and negotiate licence agreements," stressed the HUN-REN CEO.

The service centre is built around the researcher. It offers training to help them understand how to bring an innovation to market and develop their project management and team-based research skills. The service centre also assists with international tenders, thereby strengthening the effectiveness and international presence of domestic research.

Roland Jakab pointed out that launching the technology transfer business within the Hungarian Research Network will fill a gap spanning several decades. The operation of one of the world's most successful technology transfer offices will serve as a model for this endeavour. In addition to sharing their methodology and best practice, Cambridge Enterprise has offered to appoint delegates to the organisation's board of directors.

"Our goal is to adopt the world's most successful models, which is why we are collaborating with the Massachusetts Institute of Technology (MIT) to organise the AI Symposium 2025 event with Nanyang Technological University (NTU), the world's leading AI university, here in Budapest," confirmed Jakab.

Another forward-looking outcome of the AI Symposium held in May is the strategic cooperation that is now being implemented, including the preparation of joint doctoral programmes, the launch of exchange activities, and participation in international research consortia.

"Through such international contacts, we gain knowledge and inspiration, and create opportunities for Hungarian researchers, young talent, and innovative businesses in both the domestic and international arenas."

The AI Hackathon, organised as part of the Hungarian Research Network's Artificial Intelligence Month, is an example of this international approach and the promotion of innovation. The winning idea sought to solve scientific problems by looking for analogies from completely different fields of science.

Almost fifty people attended the two-day event, during which the teams worked on disruptive yet feasible ideas. The researchers were supported by national and international mentors, including two OpenAI solution architects. OpenAI also provided technological support for the hackathon.

More and more research is based on background information prepared by AI

There are already concrete examples of how modern AI tools can be used for research purposes. One such example is the agent framework, which has been available since October 2023. One of its well-known applications is Google AI Co-Scientist.

These new models are capable of representing PhD-level knowledge in a given field.

This is also a major improvement, as researchers have traditionally collaborated with teams of PhD students and senior researchers to define and research topics together. However, if AI can deliver PhD-level knowledge, a researcher could deploy up to 10,000 virtual 'PhD agents' to solve the same problem.

These agents formulate hypotheses, design experiments and conduct research, and then publish their findings. These results generate new hypotheses, and the cycle continues. This type of large-scale, parallel research can dramatically accelerate the pace of scientific discovery.

And that's not even taking innovation into account. The agent framework can also be helpful when developing a research result or patent into a marketable product. Targeted, domain-specific virtual PhD agents can analyse the idea, determining what it is really about, where it can be exploited and what technology or business area it can be applied to.

These agents can recognise how ideas can be turned into products or services.

If the first agent determines that the idea can be incorporated into a product in the automotive industry, for example, the next agent will identify which companies are involved in this area. Another agent will then analyse which departments might be involved in product development. The system will also search for relevant employees on LinkedIn and analyse their profiles, activities and interests.

Based on this information, it creates a personalised marketing message that you just have to send. This process can be automated and completed in a matter of hours, whereas previously it would have taken months or even years. This radically speeds up the innovation process, enabling new ideas to reach the market much faster.

Roland Jakab says that it is particularly exciting to think about the opportunities that will arise if these tools become commonplace in research and innovation.

Integrating advanced models and simulations into the scientific ecosystem could help us to find solutions to every known disease in the foreseeable future.

"It is therefore important to understand how these tools work, and how we can integrate them into our daily lives and research. If we don't start using them and continue with traditional methods, our research could quickly become uncompetitive. However, those who take advantage of these opportunities will make discoveries much faster and more efficiently."

SPONSORDED CONTENT --- The publication of this article was sponsored by HUN-REN

Cover image (for illustration purposes only): Getty Images