Tuesday September 25, 2018
Home Science & Technology A bot that ca...

A bot that can sketch like human? Microsoft is developing one!

Each image contains details that are absent from the text descriptions, indicating that this AI contains an artificial imagination

0
//
68
Microsoft acquired the start-up PlayFab. Pixabay
Microsoft acquired the start-up PlayFab. Pixabay
Republish
Reprint
  • Microsoft is developing a bot which can draw what you want through Artificial Intelligence technology
  • These pictures will be created by the computer from scratch, pixel by pixel
  • Currently, the technology is imperfect but the researchers are looking forward to develop a model through which help humans and bots to interact with each other

Microsoft is developing a bot that can draw what you want it to by leveraging Artificial Intelligence (AI) technology — programmed to pay close attention to individual words when generating images from caption-like text descriptions.

The technology, which the researchers simply call the drawing bot, can generate images of everything from ordinary pastoral scenes — such as grazing livestock — to the absurd and a floating double-decker bus.

ALSO READ: Microsoft to offer cloud services to Indian start-ups

Each image contains details that are absent from the text descriptions, indicating that this AI contains an artificial imagination.

For now, the technology is imperfect. Pixabay
For now, the technology is imperfect. Pixabay

“If you go to Bing and you search for a bird, you get a bird picture. But here, the pictures are created by the computer, pixel by pixel, from scratch. These birds may not exist in the real world — they are just an aspect of our computer’s imagination of birds,” Xiaodong He from Microsoft’s research lab in a blog post late on Thursday.

According to results on an industry standard test, reported in a research paper posted on arXiv.org, the bot produced a nearly three-fold boost in image quality compared to the previous state-of-the-art technique for text-to-image generation.

The core of this bot is a technology known as a "Generative Adversarial Network" or GAN. Pixabay
The core of this bot is a technology known as a “Generative Adversarial Network” or GAN. Pixabay

The network consists of two Machine Learning models — one that generates images from text descriptions and another, known as a discriminator, that uses text descriptions to judge the authenticity of generated images.

ALSO READ: Microsoft slashes 7,800 jobs, mostly in phones unit

The researchers said that text-to-image generation technology could find practical applications acting as a sort of sketch assistant to painters and interior designers or as a tool for voice-activated photo refinement.

“For AI and humans to live in the same world, they have to have a way to interact with each other. The language and vision are the two most important modalities for humans and machines to interact with each other,” The blog post explained. (IANS)

Click here for reuse options!
Copyright 2018 NewsGram

Next Story

AI Helps Find Source Of Radio Bursts 3 Billion Light Years Away From Earth

The researchers developed the new, powerful machine-learning algorithm and reanalysed the 2017 data, finding an additional 72 bursts not detected originally.

0
Space, radio
'AI helps track down mysterious cosmic signals', Pixabay

Scientists say they have used artificial intelligence (AI) to discover 72 new fast radio bursts from a mysterious source about three billion light years away from Earth.

The initiative may advance the search to find signs of intelligent life in the universe, said researchers from the University of California, Berkeley in the US.

Fast radio bursts are bright pulses of radio emission mere milliseconds in duration, thought to originate from distant galaxies.

However, the source of these emissions is still unclear, according to the research published in The Astrophysical Journal.

Theories range from highly magnetised neutron stars blasted by gas streams from a nearby supermassive black hole, to suggestions that the burst properties are consistent with signatures of technology developed by an advanced civilization.

 

earth, radio
While most fast radio bursts are one-offs, the source here, FRB 121102, is unique in emitting repeated bursts. Wikimedia Commons

 

“This work is exciting not just because it helps us understand the dynamic behaviour of fast radio bursts in more detail, but also because of the promise it shows for using machine learning to detect signals missed by classical algorithms,” said Andrew Siemion from the University of California – Berkele.

 

Researchers are also applying the successful machine-learning algorithm to find new kinds of signals that could be coming from extraterrestrial civilisations.

While most fast radio bursts are one-offs, the source here, FRB 121102, is unique in emitting repeated bursts.

This behaviour has drawn the attention of many astronomers hoping to pin down the cause and the extreme physics involved in fast radio bursts.

The AI algorithms dredged up the radio signals from data were recorded over a five-hour period in 2017, by the Green Bank Telescope in West Virginia in the US.

Radio
The researchers developed the new, powerful machine-learning algorithm and reanalysed the 2017 data, finding an additional 72 bursts not detected originally. (IANS)

An earlier analysis of the 400 terabytes of data employed standard computer algorithms to identify 21 bursts during that period.

“All were seen within one hour, suggesting that the source alternates between periods of quiescence and frenzied activity,” said Berkeley postdoctoral researcher Vishal Gajjar.

Also Read: HCL Launches AI Based ‘HCL Turbo’

The researchers developed the new, powerful machine-learning algorithm and reanalysed the 2017 data, finding an additional 72 bursts not detected originally.

This brings the total number of detected bursts from FRB 121102 to around 300 since it was discovered in 2012, researchers said. (IANS)