Australian AI experts head to New Orleans for the latest in computer vision
Thousands of the world鈥檚 top artificial intelligence (AI) experts will travel to New Orleans, La., next week for the leading conference on computer vision; among them, researchers from the Australian Institute for Machine Learning (AIML) presenting 23 papers at the forefront of new research in the field.
Known as , the Conference on Computer Vision and Pattern Recognition is held annually in North America and is broadly considered the most important conference in AI; producing the world鈥檚 fourth-most impactful scientific papers in a ranking by , beaten only by three long-established science 鈥榖ibles鈥 (Nature, The New England Journal of Medicine, and Science).
Computer vision is a field of artificial intelligence (AI) that deals with how computers can gain meaningful understanding of information from visual input such as digital images and video. Applications are broad, and include everything from autonomous vehicles and robots, medical imaging, space satellites, and even popular social media apps like TikTok.
CVPR paper submissions are highly competitive, and subject to a rigorous blind peer review process. This year, the conference has a (2,067 papers from more than 8,000 full submissions). CVPR will be held from 19 - 24 June 2022 at the Ernest N. Morial Convention Center in New Orleans, La., the first in-person conference after the 2020 and 2021 conferences were held online due to Covid-19.
In addition to the main conference program, it will also include a major trade expo and more than 130 specialist workshops. More than 9,000 machine learning students, researchers, and industry professionals are expected to attend.
AIML鈥檚 Director, Professor Simon Lucey, an author of four papers accepted to CVPR 2022, is looking forward to this year鈥檚 conference. He says that CVPR鈥檚 rising prominence speaks to the importance of computer vision research in fueling innovation, not just in AI and tech fields, but across industry and business more broadly.
鈥淚t鈥檚 remarkable for any type of engineering or mathematical discipline. The reason that鈥檚 the case is this subset of AI, computer vision, has really been driving the field,鈥 Professor Lucey says.
鈥淭he big innovations in AI in the last 10 years have really stemmed from computer vision. That鈥檚 why it鈥檚 so way out in front.鈥
鈥淭he value of computer vision isn鈥檛 just that it鈥檚 amazing technology, it鈥檚 where it鈥檚 applied and what it鈥檚 doing for us that鈥檚 important. We鈥檙e seeing significant developments in diagnostic health care, environmental and agricultural management, as well as autonomous vehicles and defence technology. It鈥檚 bringing about efficiencies across a whole range of really big industries.鈥
鈥淭hat鈥檚 why it鈥檚 so competitive. It鈥檚 incredibly valuable.鈥
While other academic fields typically publish in scholarly journals, with longer publication lead times, the machine learning research community (and many other areas of computer science) often publishes in conference proceedings. This is primarily due to the faster turnaround time from paper submission, acceptance, to publication. The accelerated process favours the AI industry, where new generations of software capability are increasingly measured in months, not years.
An institute of the 成人大片, AIML was established in 2018 at Lot Fourteen, South Australia鈥檚 innovation precinct, in partnership with the Government of South Australia. With more than 160 members, AIML is Australia鈥檚 largest and best-ranked site for AI research capability[1]. In the speciality of computer vision research, it typically ranks among the global top five[2].