We address the challenges of the industry via Artificial Intelligence
In this article promoted and published by Empresa XXI in its March 1st edition, Dr. Jorge Posada, Deputy Director of Vicomtech, delves into the real applications of artificial intelligence in industry. As a centre of reference in this field, we share with our community these reflections and examples that will undoubtedly be useful to understand what the role of AI is today, how companies can benefit from AI technologies and what are the challenges for the future.
The impact of Artificial Intelligence (AI) in different sectors and applications is an undeniable reality that affects many professional and personal areas of our society. In Industry, Health, Banking, Transportation, Administration ... It is not fiction science: it is a reality today, and its influence will increase more and more in the coming years. Our main scientific-technological lines specialise precisely in several of the key technologies of Artificial Intelligence, in the capacity of computer systems to present a behavior that we would classify as intelligent, and in its relationship with interaction, visualization and communication between people and systems.
Research applied to real challenges
The main lines of research in AI must be mainly directed to the industrial and social challenges that affect us. This is very important, in our perspective, the technological specialization in AI makes full sense in its application in real systems, which solve a real need. For us, the first thing is to understand the challenge of the application, and then study how AI (and other technologies) can solve it. We work in this sense in multiple sectors: in industry and advanced manufacturing we apply AI very directly, with computer vision and machine learning techniques that industrial systems use for the purpose of detecting dimensional or surface failure of parts, for example. We also take multidimensional data from industrial sensors and predict the behavior of machines or systems using data intelligence and visual analytics. Or we facilitate through AI the intelligent interaction of operators with the new manufacturing environments in Industry 4.0, with intelligent and flexible systems that enable their integration and learn from reality.
In Transportation, automated and connected driving (at different levels) allows us to move towards autonomous driving, in which vehicles can perceive the environment, decide on strategies and act in a similar, and even better, way than humans. We work with first-rate international partners, leading several European initiatives, in aspects of perception using cameras and other sensors, and in how to manage and learn through AI from the immense amount of videos (digital media) and other data generated in this context, applying deep learning and other AI techniques.
In Health, prognosis, diagnosis and treatment are changing in a very relevant way thanks to AI, which allows more and more personalization and the greater effectiveness of medicine in every way. Some relevant examples we work in include AI techniques such as clinical recommendation systems or automatic detection of vascular, ocular or tissue problems through image analysis and machine learning.
Another important area that has been improved thanks to this new generation of AI systems is Natural Language Processing: applications such as machine translation and dialogue systems (including chatbots) have acquired unthinkable levels of precision and naturalness a few years ago. We work intensively in this area with relevant companies and institutions both locally and internationally.
Evolution of AI in the future
Contrary to what it might seem, AI is a science that has come a long way in the last 70 years, it is not something new. What has allowed the recent explosion of applications is the coincidence of several circumstances in 6 dimensions: the availability of massive data, the exponential increase in processing capacity and the decrease in consumption, machine learning algorithms and techniques (such as deep learning, but many others too), the number of connected IoT devices (tens of billions), ubiquitous high-performance connectivity (the now coming 5G, for example) and cloud computing. All these factors together have a cost-benefit ratio very close to real needs and have allowed this explosion. The challenges and trends go precisely along this line: to continue to delve into these 6 axes to make AI even more applicable. However, there is a dimension that must be emphasized and that is not technological: the cultural, social and ethical impact of AI. It is essential to work on this. Likewise, standardization and access to data are other highly relevant aspects.
Some application examples
We have recently compiled at Vicomtech a catalog with the 100 most relevant projects for the use of AI in real applications in recent years. For example, the VI-DAS (Vision-inspired driver assistance systems) project led by Vicomtech, with Honda, Valeo, TomTom, etc. It was selected by the European Commission in 2019 (AI for Europe Factsheet -DSM-) as one of the 6 European reference projects in Artificial Intelligence. With the company IKOR we have developed the SIVI project financed by the Basque Government an automatic learning and augmented reality system to assist the operators of assembling electronic parts, capable of adapting to tens of thousands of components and with less than 3 minutes training time for the operator. We lead the European DESIREE project with Onkologikoa, Bilbomática and a relevant international consortium for IA applied to clinical decision systems in Breast Cancer. And recently, supporting the social mission of technology, we have promoted the BATUA initiative (www.batua.eus), a translator of Artificial Intelligence from Basque to Spanish for general use and free for all citizens, with the best translation quality, and with more than 25,000 translations daily.
Our main objective is to continue deepening the application of artificial intelligence, visualization, interaction and communication in the real problems of society, addressing quality research with tangible results for companies and institutions.