Long way before AI systems take over humans jobs: Study

Researchers at University of California, Los Angeles (UCLA) in the US conducted various experiments which showed that it is easy to fool the deep learning neural networks.

Press Trust of India
Los Angeles, Publish Date: Jan 8 2019 1:16PM | Updated Date: Jan 8 2019 1:16PM
Long way before AI systems take over humans jobs: StudyRepresentational Image

Artificial intelligence (AI) systems have a long way to go before they can take over tasks and jobs traditionally performed by people, say scientists who highlighted the severe limitations of deep learning computer networks.

Researchers at University of California, Los Angeles (UCLA) in the US conducted various experiments which showed that it is easy to fool the deep learning neural networks.

"The machines have severe limitations that we need to understand," said Philip Kellman, a UCLA professor and senior author of the study published in the journal PLOS Computational Biology

According to Kellman, machine vision has drawbacks. 

In the first experiment, researchers showed colour images of animals and objects to one of the best deep learning networks, called VGG-19. 

However, the images had been altered. For example, the surface of a golf ball was displayed on a teapot; zebra stripes were placed on a camel; and the pattern of a blue and red argyle sock was shown on an elephant. 

VGG-19 ranked its top choices and chose the correct item as its first choice for only five of 40 objects.

"We can fool these artificial systems pretty easily. Their learning mechanisms are much less sophisticated than the human mind," said Hongjing Lu, a UCLA professor.

In the second experiment, the psychologists showed images of glass figurines to VGG-19 and to a second deep learning network, called AlexNet. VGG-19 performed better on all the experiments in which both networks were tested.

Both networks were trained to recognise objects using an image database called ImageNet.

However, both networks did poorly, unable to identify the glass figurines. Neither VGG-19 nor AlexNet correctly identified the figurines as their first choices.

Most of the top responses were puzzling to the researchers, such as VGG-19's choice of "website" for "goose" and "can opener" for "polar bear." On average, AlexNet ranked the correct answer 328th out of 1,000 choices.

"The machines make very different errors from humans," Lu said.

In the third experiment, the researchers showed 40 drawings outlined in black, with images in white, to both VGG-19 and AlexNet. These first three experiments were meant to discover whether the devices identified objects by their shape.

The networks again did a poor job of identifying such items as a butterfly, an airplane and a banana.

Researchers concluded that while humans see the entire object, the artificial intelligence networks identify fragments of the object.

"This study shows these systems get the right answer in the images they were trained on without considering shape," Kellman said. 

"For humans, overall shape is primary for object recognition, and identifying images by overall shape doesn't seem to be in these deep learning systems at all," he said.

There are dozens of deep learning machines, and the researchers think their findings apply broadly to these devices.

Latest News

  1. Forces launch searches in Nathipora Sopore
  2. Forces launch searches in Nathipora Sopore
  1. ‘Sit in temples like offering Namaz,’ says UP CM Yogi ahead of Priyanka’s Ayodhya tour
  2. ‘Sit in temples like offering Namaz,’ says UP CM Yogi ahead of Priyanka’s Ayodhya tour
  1. BSP state general secretary, 16 others join Congress in Jammu
  2. BSP state general secretary, 16 others join Congress in Jammu
  1. General Election 2019: ECI provides postal ballot facility to Kashmiri migrants
  2. General Election 2019: ECI provides postal ballot facility to Kashmiri migrants
  1. Cong releases video of 'BJP leader' exchanging money after note ban, Jaitley says fake sting
  2. Cong releases video of 'BJP leader' exchanging money after note ban, Jaitley says fake sting
  1. Omar Abdullah visits Hajin, says slain minor’s family struggling
  2. Omar Abdullah visits Hajin, says slain minor’s family struggling
  1. NC working hard for return of peace, 'our workers real mujahideen': Sagar
  2. NC working hard for return of peace, 'our workers real mujahideen': Sagar
  1. PDP offers legal assistance to JKLF, JeI to fight ban
  2. PDP offers legal assistance to JKLF, JeI to fight ban
  1. Pak places two Hindu girls in protection after alleged forced conversion
  2. Pak places two Hindu girls in protection after alleged forced conversion
  1. China destroys maps showing Arunachal as India's
  2. China destroys maps showing Arunachal as India's
  1. Modi creating two 'Hindustans' in India: Rahul
  2. Modi creating two 'Hindustans' in India: Rahul
  1. Imran Khan says Indo-Pak relations to remain tense till elections in India
  2. Imran Khan says Indo-Pak relations to remain tense till elections in India
  1. PDP’s Agha Mohsin files nomination for Srinagar Lok Sabha seat
  2. PDP’s Agha Mohsin files nomination for Srinagar Lok Sabha seat
  1. Forces launch searches in north Kashmir's Sopore
  2. Forces launch searches in north Kashmir's Sopore
  1. Bodies of drowned siblings fished out from nallah in north Kashmir
  2. Bodies of drowned siblings fished out from nallah in north Kashmir
  1. Strong banking sector indispensable for maintaining momentum of economic growth: J&K Bank chairman
  2. Strong banking sector indispensable for maintaining momentum of economic growth: J&K Bank chairman
  1. Peoples Conference leader Irfan Ansari files nomination papers from Srinagar
  2. Peoples Conference leader Irfan Ansari files nomination papers from Srinagar
  1. Former Joint Commissioner Kaneez Fatima joins NC
  2. Former Joint Commissioner Kaneez Fatima joins NC
  1. J&K Tourism, J&K Hoteliers Club organise Kashmir conclave in Mumbai to woo tourists
  2. J&K Tourism, J&K Hoteliers Club organise Kashmir conclave in Mumbai to woo tourists
  1. India, Pakistan armies trade heavy fire along LoC in Rajouri
  2. India, Pakistan armies trade heavy fire along LoC in Rajouri
  1. UP govt suspends eight school teachers for social media posts over Pulwama attack, Balakot strike
  2. UP govt suspends eight school teachers for social media posts over Pulwama attack, Balakot strike
  1. Opening Sharda Peeth welcome step, hope Pak opens more routes: Omar Abdullah
  2. Opening Sharda Peeth welcome step, hope Pak opens more routes: Omar Abdullah
  1. Several youth detained during nocturnal raids in Sopore
  2. Several youth detained during nocturnal raids in Sopore
x
This site uses cookies to deliver our services and to show you relevant news and ads. By using our site, you acknowledge that you have read and understand our Cookie Policy, Privacy Policy, and our Terms of Service.That's Fine