AI lab DeepMind has developed AI software capable of controlling a key system in nuclear fusion, potentially paving the way for more stable and efficient reactions. The neural network was created by computer scientists at Alphabet subsidiary DeepMind and physicists from the Swiss Federal Institute of Technology in Lausanne. Background: - Nuclear fusion occurs from fusing together two atoms at very high temperatures, releasing energy without emitting greenhouse gases.
- The future of the technology as a viable power source currently relies on a "tokamak" machine, a donut-shaped experimental reactor that contains and re-creates the fusion process that powers the Sun and other stars.
- Inside a tokamak, hydrogen is superheated into a plasma state to enable the nuclei of hydrogen atoms to be fused, generating huge amounts of energy and heat to create steam (and power a turbine for electricity).
- The plasma is so hot that it has to be contained and suspended inside the tokamak using magnetic fields.
DeepMind's AI: - This is where the new algorithm comes into play. The software was trained to learn on its own how to best control the magnetic coils through reinforcement learning, a type of AI-based reward system.
- The neural network can maintain and manipulate the plasma into new configurations. When used on the Swiss Plasma Center's real-life tokamak, the AI system was able to function just like it had in a virtual reactor and control the plasma's shape for roughly two seconds.
- Some of the plasma shapes were close to the system's limits, something that scientists would not normally risk "without the confidence of AI," said Swiss Plasma Center scientist Ambrogio Fasoli.
- This could lead to better and new designs for tokamaks and controllers, as well as other "industrial and scientific" control mechanisms.
- The findings were published today in Nature.
MIT TECHNOLOGY REVIEW 
| |
Former Google CEO Eric Schmidt announced plans to invest $125M into AI research. The philanthropic fund, called AI2050, will support research projects that could help tackle AI bias, harms, geopolitical conflicts, and other "hard" challenges. More: - Payouts will be made to individual academic researchers over the next five years.
- Schmidt will chair the fund alongside Google's new head of technology, James Manyika.
- The funding comes after Schmidt and his wife, Wendy, donated $150M toward a research center based at the Broad Institute, which is focused on fighting diseases through data, machine learning, and more.
Zoom-out: - Schmidt was Google's CEO from 2001-2011 and served as executive chairman from 2015-2018.
- He formerly chaired the U.S. government's National Security Commission on Artificial Intelligence, which released a report saying that the U.S. is unprepared to compete in the new age of AI and risks falling behind China and other superpowers.
- Schmidt also co-wrote the book, "The Age of AI: And our Human Future," with Henry Kissinger and Dan Huttenlocher.
THE FINANCIAL TIMES
| |
 A message from ACTIVECAMPAIGN 10 pre-built email automations that get you started with automation fast. Don’t you wish there were more hours in the day? More hours in an hour even? These 10 email automation templates can’t add more hours to the clock — but they can save you some hours in your day with automation recipes, our term of pre-built automation templates. This email automation starter pack includes 0 automation recipes that cover a variety of tasks and touchpoints that encompass every step of your customer’s lifecycle. What you get: - 10 pre-built automation templates for marketing and sales to get your business running while you sleep
- Detailed import instructions while setting up the automations to ensure their success.
- A 14 day trial of ActiveCampaign to put these automations to the test.
Try now for free
| |
 Facial recognition is a key topic in the debate over Europe's proposed AI rules, which could take until 2024 or later to formally become law. The European Commission, the EU's executive branch, released its proposed rules to govern AI systems last year. They've gone through many months of discussions and revisions with the European Parliament, which could agree on a "common position" this November, predicts lawmaker Dragos Tudorache. More: - After that, the commission and lawmakers will have to hold talks with EU countries, which could take about a year and a half, Tudorache told Reuters.
- That pushes the timeline for the rules to be adopted and implemented across member states into 2024.
- Meanwhile, Tudorache said facial recognition remains a hotbed topic in the rules as lawmakers and others argue how much of the technology should be banned.
- Commission members want to allow some use of facial recognition in cases of serious crimes, but advocates who are concerned about mass surveillance and more prefer stricter rules.
- In October, European Parliament members did give their support for some of the draft rules. Members passed a nonbinding resolution that would outlaw police use of facial recognition in public places, as well as the creation of private facial recognition databases.
REUTERS
| |
 Soul Machines, a startup using AI to create digital lifelike people that could populate the metaverse, raised a $70M round by SoftBank Vision Fund 2. The Series B1 round included participation from Cleveland Avenue, Liberty City Ventures, and Solasta Ventures. Existing investors Temasek, Salesforce Ventures, and Horizons Ventures also participated. Background: - In 2016, tech entrepreneur Greg Cross founded Soul Machines with Mark Sagar, who won Academy Awards for his facial motion solutions for digital characters in films such as "Avatar" and "King Kong."
- That science went into the technology behind Auckland-based Soul Machines, which specializes in creating AI "virtual humans" that can interact with people online and through in-store screens.
- Its B2B clientele spans a range of industries and includes consumer brands like Proctor and Gamble and the entertainment industry, where it's been producing hyper-realistic digital twins of celebrities.
- The company believes its tech could usher in a new digital workforce that can act as salespeople and consultants for companies, as well as interact with users in future metaverse experiences.
More: - Soul Machines said it will use the funding to enhance its Digital Brain technology that uses "cognitive modeling" to recreate human-like qualities for its autonomous animated characters.
- The company also plans to expand its sales teams, mostly in Phoenix, Arizona, where it employs 120 staffers at a second research center.
- The company has raised $135M to date.
VOICEBOT.AI
| |
 A message from WEGIFT Digital incentives that deliver the business outcomes you need Breakthrough the noise, acquire new users, and close more deals with the market-leading digital gifting platform. Whether you are on the frontline driving customer acquisition or one of the heros supporting growth, generate new business and save lots of time. You can instantly send digital gift cards using our simple and powerful API, or you can just log in and send any reward from over 1,200 brands - including Amazon, Starbucks, and Target - at the press of a button. Digital gift cards are a proven way to increase new customer acquisition, retain customers, or incentivize users in referral campaigns, typically increasing conversion rates by more than 25%. It’s quick and cost-effective. WeGift can help you: - Improve conversion rates
- Reduce your CPA
- Save significant time
Companies all over the world use WeGift to drive growth, from signature start-ups like Seated and Fluz to established brands like Samsung, Vodafone and Energy Earth. Click the button below to see how WeGift works. Get a demo 
| |
 Data analytics company App Annie has changed its name to Data.ai, reflecting its plans to sell analytics products to companies using data and AI solutions. The company also announced a partnership with Similarweb, a web analytics firm, as it seeks to combine mobile and web analytics and unify datasets. More: - The company has historically gathered data and sold statistics about mobile-app performance and usage.
- Data.ai now calls itself the world's first "Unified Data AI company." One of its main goals moving forward is to sell "full-stack analytic products to enterprise clients."
- It plans to combine consumer and market data and apply AI to generate new digital insights.
- The update comes after Data.ai released an AI-driven app a year ago that enables companies to access mobile market data.
Zoom-out: - In 2018, former App Annie CEO Bertrand Schmitt was replaced by Theodore Krantz amid a massive overhaul to change how the company generates data estimates and maintains confidentiality. The company recently agreed to settle SEC fraud charges for $10M.
- Moving forward, Krantz said he sees "a blue ocean opportunity" to assemble more digital datasets and activate AI "so that enterprises can compete at a higher level."
VENTUREBEAT
| |
🍋 FRESH FUNDING 🍋 - Leaf Logistics (New York, N.Y.), machine learning-based transportation platform: $37M B led by Sozo Ventures, participation from Madrona Venture Group, Playground Global, Floodgate, Schematic Ventures, et al.
- Odaia (Toronto, Canada), AI-powered commercial insights SaaS platform for pharmaceutical companies: $13.8M A led by Flint Capital, participation from Innospark Ventures, Alumni Ventures, Graphite Ventures, et al.
- hOS (Boston, Mass.), stealth mode AI tech: $12.8M Seed led by NEA, participation from B5 Capital, Cortical Ventures, IA Ventures, Sequoia.
- Neo.Tax (Mountain View, Calif.), ML-based tax processing platform: $10M A led by Infinity Ventures, participation from Google Ventures, Acrew Capital, Fin Venture Capital, et al.
- Altis (Fort Lauderdale, Fla.), AI-based home exercise platform: $7M Seed led by PentaLab, participation from Jerome "Junkyard Dog" Williams.
- Alaffia Health (New York, N.Y.), AI platform to eliminate overpayments on medical claims: $5M led by Anthemis, participation from 1984 Ventures, Aperture Venture Capital, Tau Ventures, Twine Ventures, et al.
For the latest on AI startup funding and more in the VC community, subscribe to our newsletter Inside Venture Capital.
| |
QUICK HITS - Ready to empower your teams to quickly explore, analyze, visualize, and collaborate on all of your cloud data? It's time for Sigma.*
- So-called neuromorphic chips that mimic the brain's biology could be useful in other areas beyond AI, such as finance and X-ray analysis, according to a study led by Sandia researchers.
- Facial recognition startup Clearview AI is collecting 1.5 billion facial photos a month and is expected to have 100 billion images in its database within a year, according to a company pitch deck first reported by The Washington Post.
- Sway AI launched its no-code AI platform for enterprises to build and deploy AI workflows without prior AI knowledge or coding experience.
- The University of Florida has joined a National Humanities Center initiative to create college-level courses on ethical and equitable AI.
*This is sponsored content.
| |
Upcoming Events: - February 22 - The Future of Tax & Accounting for Crypto, DeFi, & NFTs is here. Are you ready? (Register Here) *
- February 22 - Venturing in VC - Episode #7 with Roy Bahat (Bloomberg Beta) (Register Here)
- February 23 - Need to Know NoCoder: Derya Sousa (Kianda Technologies) (Register Here)
- February 24 - Alternative Investments (Register Here)
- February 24 - Conversational AI in eCommerce with Flipkart (Register Here) *
- February 28 - Inside Investment DAOs with Global Coin Research (Register Here)
- March 10 - NoCode's Impact on the Gig Economy (Register Here)
- March 23 - Need to Know NoCoder: Nichole Bestman (Bubble) (Register Here)
- March 31 - How To Build In Web 3.0 (Register Here)
- April 07 - Blockchain and the Law (Register Here)
- April 14 - Shopping In The Metaverse (Register Here)
*This is a sponsored listing.
| |
 |  | Beth is an Inside Analyst and a former public policy/investigative reporter. She has covered AI, VR, technology, and e-commerce for Inside. When she's not busy writing, she's hiking in Arizona's desert, playing pickleball, or trying out new Mexican food recipes. |   | | 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. | |
| |
Acquire and engage customers at scale using eGift Cards from 800+ global brands with WeGift. | |
Comments
Post a Comment