All animals are born with inherent behavioral capabilities, arising from neural circuits that are encoded in the genome. It is common knowledge that the brain plays a crucial role as it contains numerous neural connections required to control complex behavior. However, the genome has limited capacity for all that information.
For decades, scientists have wondered how the genome, with space for only some of all that information, could still encrypt information and pass it from one generation to another.
Researchers now believe that the rules which encrypt the formation of circuits have to pass through a ‘genomic bottleneck’ as they’re passed from one generation to another. This has led to the discovery that this bottleneck can also capture important features of the circuit, resulting in improved transfer learning to new datasets and tasks.
Now researchers have determined that a new AI algorithm inspired by DNA can perform tasks like video games and image recognition almost as effectively as properly trained artificial intelligence networks. The algorithm works by mimicking how genomes encrypt behaviors with limited space, calling to attention the evolutionary advantage of efficient compression of information.
This finding may aid in the development of advanced artificial intelligence systems that can run on smaller devices.
This research was led by Professors Alexei Koulakov and Anthony Zador at Cold Spring Harbor Laboratory, who were inspired by the ability of the genome to compress lots of information.
Zador posited that the limited capacity of the genome was the very thing that made us as we are, which formed the basic of the study’s ‘genomic bottleneck’ algorithm. For their study, the researchers developed a computer algorithm that folded a great deal of data into a small package.
Once this was done, they tested this algorithm against artificial intelligence networks which had undergone numerous training rounds. They discovered that the new and untrained algorithm could carry out tasks as effectively as advanced AI, even going as far as to play games like Space Invaders really well.
This, the researchers explained, showed that the algorithm may have innately understood how to play the video game.
In their report, the researchers stated that artificial intelligence still wasn’t able to replicate our natural abilities now. However, the algorithm allowing for levels of compression not yet seen in artificial intelligence is still an impressive feat. The researchers highlight in their conclusion that this feature could have many applications in tech.
The study’s findings were reported in PNAS. Other researchers involved included Sergey Shuvaev and Divyansha Lachi, both of whom are postdoctoral fellows at the laboratory.
More R&D work is being done by tech firms like International Business Machines Corp. (NYSE: IBM). As the months roll by, we should expect many more AI innovations aimed at easing problem-solving in various industries.
About AINewsWire
AINewsWire (“AINW”) is a specialized communications platform with a focus on the latest advancements in artificial intelligence (“AI”), including the technologies, trends and trailblazers driving innovation forward. It is one of 70+ brands within the Dynamic Brand Portfolio @ IBN that delivers: (1) access to a vast network of wire solutions via InvestorWire to efficiently and effectively reach a myriad of target markets, demographics and diverse industries; (2) article and editorial syndication to 5,000+ outlets; (3) enhanced press release enhancement to ensure maximum impact; (4) social media distribution via IBN to millions of social media followers; and (5) a full array of tailored corporate communications solutions. With broad reach and a seasoned team of contributing journalists and writers, AINW is uniquely positioned to best serve private and public companies that want to reach a wide audience of investors, influencers, consumers, journalists, and the general public. By cutting through the overload of information in today’s market, AINW brings its clients unparalleled recognition and brand awareness.
AINW is where breaking news, insightful content and actionable information converge.
To receive SMS alerts from AINewsWire, text “AI” to 888-902-4192 (U.S. Mobile Phones Only)
For more information, please visit www.AINewsWire.com
Please see full terms of use and disclaimers on the AINewsWire website applicable to all content provided by AINW, wherever published or re-published: https://www.AINewsWire.com/Disclaimer
AINewsWire
Los Angeles, CA
www.AINewsWire.com
310.299.1717 Office
Editor@AINewsWire.com
AINewsWire is powered by IBN