Friday, 09 December 2022

 

 

LATEST NEWS TPL : Bengaluru Spartans displace Finecab Hyderabad Strikers from the top on Day 2 BJP's win in Gujarat shows people's anger against dynasty politics : Narendra Modi BJP's lotus blooms wide in Gujarat, Congress washed out ISL 2022-23 : Hugo Boumous' late penalty helps ATK Mohun Bagan beat Jamshedpur FC PKL : Rakesh Narwal leads Haryana Steelers to win against Telugu Titans Shehnaaz Gill To Enter In The Bigg Boss Season 16 With Rap King MC Square PKL : Dabang Delhi KC strong second half performance earns them a tie, seal playoff spot Total 2,46,914 persons recruited by UPSC & SSC in last 5 years, vacant posts nearly 9.79 lakh Delegation of Gurjar Desh Charitable Trust calls on Lt Governor Manoj Sinha Lt Governor Manoj Sinha addresses Message of Humanity Conference organised by Dogra Sadar Sabha, J&K 2 day Workshop on "Prevention of sexual harassment of Women at workplaces/Institutions” concludes at GCW Udhampur District Level Aadhaar Monitoring Committee meet held at Udhampur Rural Sanitation department achieves milestone of 1000 ODF Plus villages under Swachh Bharat Mission (Grameen) DDC Rajouri reviews progress of Health Sector projects taken up under District Capex Budget, other schemes 2 Day Training Program on registration of Births & Deaths through CRS Portal concludes at Jammu JKRLM organises awareness Workshop to educate women SHGs about Menstrual Health & Hygiene DC Samba Anuradha Gupta chairs District Level Aadhaar Monitoring Committee Meeting Dr. Arun Kumar Mehta declares 33rd Sr National Tennis Ball Cricket Championship open DC Srinagar Mohammad Aijaz Asad assesses performance of Blocks under Jal Jeevan Survekshan DC Anantnag Dr Basharat Qayoom conducts late night surprise check at District Hospital VC DDC Kupwara chairs joint Awareness/ Enrolment drive at Potushai Lolab

 

Even smartest AI models don't match human visual processing: Study

Technology, Research, Study, Artificial Intelligence

Web Admin

Web Admin

5 Dariya News

Toronto , 18 Sep 2022

Deep convolutional neural networks (DCNNs) do not see objects the way humans do using configural shape perception and that could be dangerous in real-world artificial intelligence (Al) applications, say researchers.

DCNNs are the type most commonly used to identify patterns in images and video."Our results explain why deep AI models fail under certain conditions and point to the need to consider tasks beyond object recognition in order to understand visual processing in the brain," said researcher James Elder from York University in Toronto.

"These deep models tend to take 'shortcuts' when solving complex recognition tasks. While these shortcuts may work in many cases, they can be dangerous in some of the real-world AI applications we are currently working on with our industry and government partners," Elder added.

For the study, published in the journal iScience, the team employed novel visual stimuli called "Frankensteins" to explore how the human brain and DCNNs process holistic, configural object properties.

"Frankensteins are simply objects that have been taken apart and put back together the wrong way around. As a result, they have all the right local features, but in the wrong places," Elder said.

The investigators found that while the human visual system is confused by Frankensteins, DCNNs are not revealing an insensitivity to configural object properties.

According to the researchers, modifications to training and architecture aimed at making networks more brain-like did not lead to configural processing, and none of the networks were able to predict trial-by-trial human object judgements accurately.

"We speculate that to match human configural sensitivity, networks must be trained to solve a broader range of object tasks beyond category recognition," Elder noted.

 

Tags: Technology , Research , Study , Artificial Intelligence

 

 

related news

 

 

 

Photo Gallery

 

 

Video Gallery

 

 

5 Dariya News RNI Code: PUNMUL/2011/49000
© 2011-2022 | 5 Dariya News | All Rights Reserved
Powered by: CDS PVT LTD