Left Menu
Development News Edition

AI systems have severe limitations, can be fooled pretty easily: Study

Devdiscourse News Desk | Los Angeles | Updated: 08-01-2019 13:24 IST | Created: 08-01-2019 12:25 IST
AI systems have severe limitations, can be fooled pretty easily: Study
( Image Credit:Flickr)

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.

(With inputs from agencies.)

Download The Devdiscourse News App for Latest News.


TRENDING

OPINION / BLOG / INTERVIEW

Diya Jalao for COVID 19: How Modi put Power Grids and power warriors at risk?

While Prime Minister Narendra Modis staunch supporters are busy in search of hidden science behind the sudden announcement of 9minutes9pm campaign but his ignorance of the actual science has put the nation in another danger. The scientists ...

Energy markets post-COVID 19: Recovery may take time, transition to continue

As oil markets have crashed, the experts believe the oil prices will come down drastically and continue making the energy transition a financial burden for many economies for a couple of years.  However, the market alone was never...

'Diya jalao' for COVID 19: Modi needs to practice his teachings and control fake news on social media

In his video message to the nation on Friday the Prime Minister talks about peoples collective power to win the invincible. However, he has made no concreate efforts to win over the confidence and represent the collective power of the polit...

Tourism post-COVID 19: Lessons learned, out-of-box strategies to ensure the revival

In the pre-COVID 19 period the tourism industry was seen with much hope throughout the world. However, the COVID 19 pandemic has completely paralysed the tourism industry and its dependent sectors. The tourism industry in the post-COVID 19 ...

Videos

Latest News

OPEC meeting postponed until Thursday: Azerbaijan

A special meeting to discuss output cuts between OPEC and other major oil producers including Russia has been postponed until Thursday, energy-rich Azerbaijan saidThe meeting has been postponed until April 9, energy ministry spokeswoman Zam...

Delhi BJP chief Manoj Tiwari gets threatening message

Delhi BJP president Manoj Tiwari has received a threat on his cell phone from an anonymous person. The accused also sent the link to a Facebook video in which a man can be purportedly heard warning Tiwari against his comments on quarantined...

COVID-19 cases surge to 3,072: Union Health Ministry

The number of confirmed cases of coronavirus in the country has surged to 3,072 and 13 new deaths, according to the Ministry of Health and Family Welfare website.Increase of 525 COVID-19 cases in the last 24 hours, the largest spike in a da...

Will boycott Verka village for refusing cremation of Nirmal Singh, says Shiromani Ragi Sabha

A group of Hazoori Raagi on Saturday announced that it will boycott Verka village in Punjab after its residents did not allow the cremation of Padma Shri awardee Bhai Nirmal Singh Khalsa. Khalsa 62, former Hazoori Raagi at the Golden Temple...

Give Feedback