With world-class research institutes in artificial intelligence (AI), Sweden keeps up with all the current latest ideas or even steps out ahead. But blue-sky research doesnt always result in practical solutions which you can use by industry. Thats where AI Sweden plays an integral role.
A significant first rung on the ladder in the development of AI Sweden came when Mikael Ljungblom was working as a political advisor to the Swedish minister for digital development. While going to see how many other countries were doing, Ljungblom and his colleagues saw that countries such as for example Japan and China were buying AI. Given the significance of the technology for competitiveness and societal development, they sensed a have to develop an AI centre in Sweden. That has been in 2017.
I began to talk with different actors in the ecosystem and I met Mats Nordlund, who’s now our colleague, Ljungblom tells Computer Weekly. He was employed in the automotive industry in Gothenburg and he previously exactly the same mission when i did. We locked arms and tried to start this together. We’d great support from Vinnova, that is Swedens innovation agency.
Vinnova provided the bottom funding for that which was to become AI Sweden. Johanna Bergman joined your time and effort in 2018. She had previously been focusing on pre-hospital e-health at the Lindholme Science Park in Gothenburg.
We got this assignment to create an AI centre in Sweden, says Bergman. We started by performing a feasibility study by soliciting ideas from all actors we’re able to think about. We talked to Swedish industry, with authorities, municipalities, regions, academia plus some international actors, attempting to understand what this will be. It had been clear we have to focus on applied AI, not fundamental research. Our assignment was to accelerate applied AI in Sweden.
In February 2019, we inaugurated the centre, which we called AI Sweden. Partners of most sorts were involved from startups to corporations, academia to public sector. It had been vital that you have this spread right from the start.
Ljungblom is currently head of communications and public affairs at AI Sweden. Johanna Bergman is currently head of strategic initiatives. She was recently recognised on her behalf achievements when you are listed in the Forbes 30 Under 30 for Science & Healthcare in Europe.
Following the inauguration, we immediately started up collaborative AI projects cross-sector and involving different types of organisations to focus on real AI problems, Bergman tells Computer Weekly. Which has evolved into what we do today. Lindholme Science Park helped us a whole lot initially and is today among our host organisations.
Benefits of a horizontal approach
The truth that AI Sweden operates across sectors helps it be not the same as most AI research centres far away. Instead of going for a vertical approach by concentrating on a couple of industries, AI Sweden has identified several areas which are important across several sectors.
One particular area is edge learning, that is a new paradigm in AI where learning occurs nearer to the foundation of working out data. The models which are built from the training process on the edge are then aggregated in a central location where they’re shared. The info itself is never used in the central location, which includes the desirable side-effect of assisting to ensure data privacy.
Following the inauguration, we immediately started up collaborative AI projects to focus on real AI problems. Which has evolved into what we do today
Johanna Bergman, AI Sweden
That is not the same as traditional machine learning (ML), where in fact the training is centralised and runs on the centralised dataset. AI Sweden has create an Edge Learning Lab, understanding that partners from different sectors such as for example automotive, healthcare and the general public sector have to work with the brand new paradigm.
The Edge Learning Lab carries a test platform, alongside shared resources for developers, data scientists, students and researchers for more information about various areas of edge learning. AI Sweden and partners Zenseact, HPE, Volvo Cars and Ericsson launched the lab in Gothenburg on 27 April 2022. The lab can help differing people and organisations quickly create a work place for experiments.
Another horizontal section of focus may be the Swedish language. None of the global tech actors value developing models for a language like Swedish, so AI Sweden made a decision to take that up itself.
What we see now are major advancements in neuro-scientific language technology and what we call foundational models huge models that understand natural language and that may generate language, says Bergman. Nearly every week, some organisation somewhere on the planet releases a fresh language model they are big language models that may generate text from images and images from text, for instance. While we see major advancements with English models, very little work has been done for smaller languages like Swedish.
To help to fill this gap, AI Sweden setup several projects addressing language models for Swedish. Not merely may be the language lab helping Swedish companies, nonetheless it serves for example for other cultures whose languages aren’t yet represented in AI models.
The purpose of the various projects would be to create a national knowledge hub on natural language understanding (NLU) to greatly help stimulate NLU focus on Swedish. AI Sweden addresses language model training, evaluation and deployment. In addition, it works on ethical and security aspects in a number of ongoing projects around NLU. Swedish authorities and public sector organisations will be the main stakeholders and you will be able to utilize the models to summarise texts, and for translation and writing support, among other activities.
Along with edge learning and Swedish natural language processing, AI Sweden also works on additional horizontal topics, but have not yet setup programmes for them. Included in these are privacy-preserving AI, complex systems, and management and logistics planning. The theory is to discover the areas which are relevant for most different sectors.
Networking and collaboration to synergies
Networking and collaboration are a significant section of what we do at AI Sweden, Ljungblom tells Computer Weekly. We make an effort to connect individuals who may have common interests without knowing it. Our horizontal approach we can develop new and sometimes surprising collaborations.
One of the most surprising synergies is that some hospitals use models produced by the automotive industry for self-driving cars. They use these same models and ways to predict the way the hospital will function.
Another unexpected collaboration is between your automotive industry and university researchers studying bird behaviour. The automotive industry is utilizing a dataset produced by researchers to categorise behaviour of Baltic Sea birds. This dataset pays to for the automotive industry as a proxy dataset for collaborative projects on self-driving.
Networking and collaboration are a significant section of what we do at AI Sweden. We make an effort to connect individuals who may have common interests without knowing it. Our horizontal approach we can develop new and sometimes surprising collaborations
Mikael Ljungblom, AI Sweden
The approach of AI Sweden would be to either take what’s already happening in pure research and apply it to industry or take what already works in a single domain and apply it to some other. Not merely is this good for society, but it addittionally helps Swedish organisations remain competitive on a worldwide level. By helping Swedish organisations overcome a few of the common challenges, AI Sweden accelerates the usage of artificial intelligence and helps companies get prior to the curve on AI adoption.
Talent development and networking
When AI Sweden was founded, it wasnt obvious that it might are likely involved in assisting develop the proper AI talent in Sweden. But since it began networking with actors in various industries, it became clear that AI Sweden was uniquely positioned to identify what talents were needed also to develop consolidated programmes that could help develop those skills.
Just this past year, AI Sweden started a couple of talent programmes. These aim not merely to build up specific talents, but additionally to greatly help develop the proper company cultures for AI.
AI Sweden also tries for connecting the dots and obtain the different elements of the ecosystem to interact. Among its aims would be to make certain government investment has been spent in the perfect way. It can this by staying up to date with who is focusing on what and finding synergies to help make the the majority of the funds.
Another way AI Sweden is helping accelerate adoption of AI is by addressing societal challenges. AI changes just how people work and can probably transform industries.
To greatly help prepare organisations for the big changes, AI Sweden works together with partners to ensure they do something towards becoming more AI mature. As individual organisations are more mature in applying AI, the complete ecosystem benefits.
Weve found a structured method of achieving this, says Ljungblom. We use an AI maturity assessment tool that people adapted from the Initiative for Applied Artificial Intelligence in Germany. Its an excellent tool that applies a structured approach for benchmarking improvements and identifying what areas to spotlight whenever using AI. This enables organisations to take the task they’re doing to another level.
National benefits of Sweden
Finland was quite early to make AI a strategic priority, says Bergman. We are able to gain a whole lot from collaborating in the Nordics. For instance, several Nordic countries have virtually identical languages, when it involves language applications, there’s great prospect of working together. Thats needs to happen.
Ljungblom highlights an interesting thing about Sweden is its decentralised government structure, with autonomous regions and municipalities. Hence, we have been forced to get effective means of working with plenty of different public entities, he says. If we compare the EU to the united states and China, we have been a lot more diverse and decentralised, with 27 different member states. In ways, this resembles the autonomous structure of Sweden, so we may have something to provide there.
Two areas of Swedish culture give it a competitive advantage in AI. One is that it is definitely a technologically friendly country, with a population that’s quick to become digital also to use digital tools. In addition, it includes a good entrepreneurial system with an excellent ecosystem of startups. Simultaneously, there are a great number of large Swedish companies, such as for example Ericsson, Saab and Ikea, and these businesses tend to match technological advances.
Another advantage is that Swedish folks are used to collaborating. Its section of their culture. In technology, they collaborate across sectors and between industry and academia. This mindset supports AI, where an ecosystem is necessary and where most of the building blocks could be shared across several application domains.
AI Sweden depends on this pervasive attitude of cooperation to greatly help connect the dots.