OpenAI introduced a new version of DALL-E, its AI model that generates photorealistic images based on text descriptions. The second-generation text-to-image generator turns words into art pieces. More: - OpenAI first introduced DALL-E, a portmanteau of Salvador Dalí and WALL-E, in January 2021.
- The neural network generates photorealistic images based on simple written descriptions.
- For example, OpenAI employee Aris Konstantinidis used the tool to generate pictures of bandana-wearing pandas riding motorcycles in the desert.
- DALL-E 2 is faster and produces higher resolution images than its predecessor. While not released to the public, the latest version does allow researchers to sign up for a preview.
- OpenAI plans to release it at a later date to third-party developers through an API. It could be particularly useful for graphic designers and app developers.
About OpenAI: - The San Francisco-based AI research lab was formed in late 2015 to pursue safe and helpful AI projects for the public.
- Its original founders were Sam Altman, Elon Musk, Greg Brockman, Ilya Sutskever, Wojciech Zaremba, and John Schulman, who initially pledged a combined $1B+ toward the venture.
- Musk later resigned from the board to avoid conflicts of interest with Tesla, among other reasons.
- In 2019, the non-profit lab transitioned to a for-profit company and announced a $1B investment package from Microsoft.
- In 2020, OpenAI released its popular language model GPT-3, which has formed the basis for applications that generate code and copy for websites, ads, and social media.
- DALL-E 2 is built on OpenAI's image recognition/computer vision system CLIP, which itself is based on GPT-3.
TECHRADAR 
| |
Amazon and Johns Hopkins University announced a new AI institute that will focus on advances in machine learning, natural language understanding, speech processing, and computer vision. The center is called the JHU + Amazon Initiative for Interactive AI (AI2AI). More: - AI2AI will be housed at Johns Hopkins' Whiting School of Engineering.
- Its founding director is Sanjeev Khudanpur, an associate professor in the Department of Electrical and Computer Engineering.
- Amazon has invested an undisclosed amount into AI2AI, which will span five years.
- The funding will support research projects led by faculty, doctoral fellowships, gift funding, and community projects and events.
- Many of these workshops and events are "aimed at making AI activities more accessible to the general public in the Baltimore-Washington, D.C. region," according to a news release.
JOHNS HOPKINS UNIVERSITY
| |
 A message from MALWAREBYTES Top trends in remote work and collaboration. Remote work has changed our use of tools across cloud collaboration, cybersecurity, backup, and other solutions. We asked people who look after security* for their organizations what is changing in their environment: - 78% report an increase in the use of cloud storage platforms
- 76% saw in increase in the use of video conferencing
- 73% noted an increase in the use of identity management and cloud access tools
- 72% reported an increase in the use of instant messaging
- 71% noted an increase in the use of cybersecurity and antivirus tools
There has never been a better time to get your team’s data, devices and servers protected. Scan, clean and protect all your devices and servers in one easy-to-use console. During our Spring Sale, save 25% for a limited time. Compare Savings *Malwarebytes 2022 Still Enduring From Home report
| |
U.K. researchers developed an algorithm that they say can detect depression based on a person's Twitter profile. The algorithm was trained on databases containing the mental health information and Twitter history of thousands of people. More: - The algorithm weighs 38 factors, including use of positive and negative words, emoji use, and number of followers.
- It is "platform independent," so it could be used on other social media platforms like Facebook or WhatsApp.
- In testing, it was 88.4% accurate using the Tsinghua Twitter Depression Dataset and 70.7% accurate using Johns Hopkins University's CLPsych 2015 dataset.
- Accuracy above 90% is "considered excellent in machine learning," according to one of its creators, Professor Abdul Hamid Sadka, director of the Institute of Digital Futures at Brunel University London.
- Its developers hail from Brunel University London, the University of Leicester, Queen Mary University of London, University of Kent, and AstraZeneca.
- Their research paper was published in IEEE Transactions on Affective Computing.
TECHNOLOGY NETWORKS
| |
Google Maps is using AI and its Duplex conversational technology to update the hours of businesses that are out of date. Google says it's on track to update the hours for more than 20 million businesses globally in six months. The automatic updates can help "keep up with the pace of change" since pandemic-related restrictions took hold. More: - Google explained how the system updates business hours in a blog post.
- Researchers built a machine learning model that first identifies if a business' hours are likely wrong.
- It considers factors like when their profile was last updated and their "popular times" information, which could conflict with their posted hours.
- Once a business is flagged, the algorithms attempt to predict the most accurate hours and update them.
- The AI model looks at the hours of nearby shops, website information, Street View images, and the Google Maps community to predict and verify.
- In some countries, Duplex, Google's AI technology that interacts over the phone, can call businesses directly and ask for hours, the company said.
THE VERGE
| |
 A message from MASTERWORKS WINTER IS COMING J Pow is raising rates. It’s been 18 years since the last cycle of rate hikes. If you don’t recall the days of 2004 this is what typically happens: - Borrowing money becomes more expensive
- Stock earnings decrease
- Less growth
So what can you do? Almost every financial advisor on the planet would tell you to diversify, In fact, more experts are advocating for an unexpected asset class, blue-chip art. Here's Their Thesis: - Art has a .01 correlation to stocks according to Citi
- Real assets like art appreciate well inflation is high
- Blue-chip art prices outpaced the S&P 500 by 164% from 1995-2021
You can access this exciting asset with Masterworks. This app valued at over $1 billion makes investing in multimillion-dollar art easy. And for a limited time, they’re giving Inside Subscribers priority access *See important Reg A disclosures
| |
Viz.ai, an AI health care and medical imaging startup, has raised $100M at a $1.2B valuation. More than 1,000 hospitals in the U.S. and EMEA now use the Israeli's startup's AI-powered disease detection and care coordination platform. More: - The Series D round was led by Tiger Global and Insight Partners, with participation from early backers, such as Scale Ventures, Kleiner Perkins, Threshold, GV (formerly Google Ventures), Sozo Ventures, CRV, and Susa.
- Viz.ai's platform leverages AI and ML technology to identify acute diseases.
- It uses AI to detect ischemic and hemorrhagic stroke, cerebral perfusion, pulmonary embolisms, and aortic disease.
- In February, the FDA cleared Viz.ai's AI algorithm for detecting brain aneurysms. The system can analyze CT signs to look for signs of weakness in cerebral blood vessels.
- Founded in 2016 by Mansi and CTO/AI specialist David Golan, Viz.ai saw its revenue reach $12M in 2021.
- Viz has raised a total of $250M to date.
THE WALL STREET JOURNAL
| |
Quick Hits - Autonomous construction equipment manufacturer Built Robotics received $64M in Series C financing.
- More than half of the databases used in 7,000 clinical AI papers came from the U.S. and China, according to a recent review.
- Crypto VC firm Hashed led a $2M extension round for LOVO, a creator of AI-generated voice NFTs, bringing its total pre-Series A funding to $6.5M.
- Data from PitchBook and the National Venture Capital Association show that U.S. VCs raised $73.8B in fresh funds last quarter. That figure sets a new record for VC fundraising in any three-month period to date.
| |
Upcoming Events: - April 12 - Venturing in VC: Episode #15 with Katelin Holloway (Seven Seven Six) (Register Here)
- April 13 - Achieve 1M+ Record/Second Kafka Ingest without Sacrificing Query Latency. (Register Here) *
- April 14 - Shopping In The Metaverse (Register Here)
- April 26 - Need To Know NoCoder - Gautam Nimmagadda (Quixy) (Register Here)
- April 28 - Alternative Investments 2: Crypto Safety Standards (Register Here)
- May 12 - Making Metaverse Dating a Reality (Register Here)
- May 19 - Understanding Fractionalization (Register Here)
- May 24 - Meet Our Fund 3 (Register Here)
- May 31 - Ethereum Merge with Anthony Donofrio (Ethereum) (Register Here)
- June 01 - Selling Consumer Goods in Web3 (Register Here)
*This is a sponsored listing.
| |
 |  | Beth is a writer and analyst at Inside.com covering artificial intelligence and daily tech news. Since starting work at Inside, she has covered various topics including virtual reality, drones, and e-commerce. Prior to that she was a public policy and investigative reporter for The Arizona Republic, where she won a Pulitzer Prize nomination and First Amendment Award for reporting on the rising costs of pensions. Reach her at Beth.Duckett@inside.com. |   | | Editor | Aaron Crutchfield is based in the high desert of California. Over the last two decades, he has spent time writing and editing at various local newspapers and defense contractors in California. When he's not working, he can often be found looking at the latest memes with his kids or working on his 1962 Ford. | |
| |
Did you know contemporary art outpaced the S&P 500 by 164% from 1995-2021? Now you do. Get access | |
Comments
Post a Comment