A new form of artificial intelligence (AI) must be able to adapt flexibly to different situations, to think logically and to have a deeper insight into the world than Chatgpt and Co. In the Excellence cluster “Bilateral Artificial Intelligence”, for which the starting shot is given on Wednesday evening, Project Manager and Ki-pioneer Sepp Hochreiter of the University of Linz (JKU) stated.
This would succeed due to the combination of two important research strands in the field of AI. The “Sub-symbolic AI” is about machine learning, ie learning based on data, such as neuronal networks or chat. In the “Symbolic AI” the focus is on dealing with artificial systems with knowledge and conclusions. On the one hand, large amounts of data must be processed and learning methods are used on the other hand, which are reproduced in the human brain.
Domestic expertise is bundled
The aim is to develop the basic principles for such a relatively wide AI, which has a higher problem -solving skills and can deal with abstractions. By bundling the domestic expertise in this area, the AI must adapt flexibly to various challenges and environments. It can implement complex plans, taking into account the connections, for example organizing a trip or helping to renovate a house as cheap and ecologically as possible.
Hochreiter explained that other actors and countries also rely on a combination of sub-symbolic and symbolic AI. In this country, however, there has been years of preparatory work to bring the groups that investigate here, which should protect a certain lead. In particular scientists from Technical University (TU) in Vienna and Graz, the Institute of Science and Technology Austria (ISTA) in Klosterneuburg (Lower Austria), the University of Klagenfurt and the Vienna University of Economic Affairs.
Chatbots have no “world knowledge”
Current AI models, for example, specialize in texting, driving or image processing in medical diagnostics, according to the head of the Institute for Machine Learning in the run-up to the kick-off event, which is taking place today at the University of Linz. They are very suitable for this, but also useless. AI Supported chatbots based on so-called large language models (Large Language Model, LLM), according to Hochreiter, cannot think logically and have no “world knowledge”: “You have no feeling that a tree also actually exists and not just as a letter.”
A self -drying car is overwhelmed by the decision whether it can drive on an opposite plastic bag or whether that is an obstacle. “To do this, she must know that something is not as massive if it can be blown away by the wind. This understanding of the world is still missing, “said Hochreiter. A “broad AI”, on the other hand, must have sufficient knowledge to adapt to new situations. She cannot connect to the skills of people, for which an “artificial general intelligence” is needed, from which someone is still a good distance.
Deepseek devoured a lot of money
The Chinese start-up Deepseek recently caused a stir. It presented a powerful AI model that was reportedly created against a fraction of the costs of the American competitive models. The company even invested a lot of money in advance to be able to present this cheap model at the end, according to the expert: “We would have had the technology, but never the team or computer sources to think and test a model like this.” There will soon be a few groups, probably also in Europe, “who imitate”.
After the high -scaling phase, in which larger models were built and more data were used, Hochreiter now sees a transition to the industrialization of the AI: “Now smaller models can now also score that specialize in certain areas of responsibility. They don’t have to know everything about Goethe or about the Renaissance. As soon as AI finds its way in the production processes and logistics of the companies, there will be many opportunities for Europeans to play here. “If you build small models for the industry, I will see opportunities again. And that will come, “Hochreiter is convinced.
In the course of the “Excellent = Austria” initiative, Nine Cluster of Excellence has so far been set up by the Science Fund FWF and the Ministry of Education – to a wide range of research topics. “Bilateral AI” is created for five years and is financed with 33 million euros. The FWF contributes 60 percent to the institutions involved. After a positive evaluation, the term can be extended to ten years by a total of 70 million euros in financing.
Source: Krone
I am an experienced and passionate journalist with a strong track record in news website reporting. I specialize in technology coverage, breaking stories on the latest developments and trends from around the world. Working for Today Times Live has given me the opportunity to write thought-provoking pieces that have caught the attention of many readers.
