Britain Courts Anthropic Amid US Defense Department Dispute 
Apple Turns 50: From Garage Startup to AI Crossroads 
Microsoft's $10 Billion Japan Investment: AI Infrastructure and Data Sovereignty Push 
SpaceX Eyes Historic IPO at $1.75 Trillion Valuation 
Nanya Technology Shares Surge 10% After $2.5 Billion Private Placement from Sandisk and Cisco 
NASA Artemis II: First Crewed Moon Mission Since Apollo Takes Four Astronauts on 10-Day Lunar Journey 
Annie Altman Amends Sexual Abuse Lawsuit Against OpenAI CEO Sam Altman 
OpenAI Pulls the Plug on Sora, Ending $1 Billion Disney Partnership 
Australia's Social Media Ban for Under-16s Sparks Global Movement 
SpaceX IPO Filing Expected This Week as Valuation Could Surpass $75 Billion 
NASA's Artemis II Crew Arrives in Florida for Historic Moon Mission 
Meta and Google just lost a landmark social media addiction case. A tech law expert explains the fallout 
SMIC Allegedly Supplies Chipmaking Tools to Iran's Military, U.S. Officials Warn 
California's AI Executive Order Pushes Responsible Tech Use in State Contracts 
TSMC Japan's Second Fab to Produce 3nm Chips by 2028 
Elon Musk Ties SpaceX IPO Access to Mandatory Grok AI Subscriptions 
Makemation: a Nollywood movie that shows AI in action in Africa 
Although several examples of artificial intelligence have gotten extremely adept at recognizing speech and voice patterns such as Siri and Cortana, recognizing other sounds has been a significant hurdle for machines. However, it would seem that MIT engineers finally managed to overcome this hurdle by teaching an AI how to recognize specific sounds and even identify the environments featured.
Some examples of sounds that AIs have had trouble dealing with in the past include crashing waves, a cheering crowd, indoor chatter, and even crying children, Phys.org reports. By allowing the AI to learn how to recognize which sounds come from what source via hours of watching videos, however, researchers from the Computer Science and Artificial Intelligence Laboratory at MIT created a sound-recognition system that is able to overcome this obstacle.
Basically, the researchers allowed the AI to watch scenes and settings in order to categorize them. They then used a second set of machines to discover the corresponding sounds to the videos. As Carl Vondrick, one of the authors of the study said, advancements in computer vision led to this breakthrough.
"Computer vision has gotten so good that we can transfer it to other domains," Vondrick said. "We're capitalizing on the natural synchronization between vision and sound. We scale up with tons of unlabeled video to learn to understand sound."
Another example of AI being able to recognize natural sounds is the work of Manny Tan, Kyle McDonald, and Jessie Barry. Both Tan and McDonald are programmers, while Barry is an ornithologist. By combining their expertise, they were able to teach an AI to catalog the different sounds that bird species make, and even generate visual elements in the process, Futurism reports.
The algorithm used basically produced the specific fingerprint of each specific bird based on the sounds they made using the “t-distributed stochastic neighbor embedding” machine learning technique. The results of the collaboration essentially opened the floodgates to new applications in various fields for detection and manipulation.
