Qualcomm says its latest smartphone system-on-a-chip, the Snapdragon 8 Gen 2, was engineered with "groundbreaking AI." The chip is expected to power most Android phones and tablets before the end of 2022 and into 2023. More: - Qualcomm announced the system-on-chip during this week's Snapdragon Tech Summit, saying it is "purpose-built for AI."
- All of the chip's processing elements are used for AI processing in what Qualcomm calls its "AI engine."
- The SoC's AI-related features include Snapdragon Smart, which increases the chip's AI performance; Snapdragon Sight, which improves object recognition; and Snapdragon Connect for AI-powered connectivity.
- Snapdragon Smart, for example, enables faster natural language processing, with gains of up to 4.35x its predecessor.
- Brands that signed up to adopt the platform into their mobile devices include ASUS, Honor, iQOO, Motorola, nubia, OnePlus, OPPO, Redmagic, Redmi, Sharp, Sony, vivo, Xiaomi, XINGJI/MEIZU, and ZTE.
HOT HARDWARE 
| |
Emirati artificial intelligence firm G42 participated in a $42M funding round for The Applied AI Co., according to Bloomberg. G42 is chaired by the UAE's national security adviser, Sheikh Tahnoon bin Zayed Al Nahyan, who is brother to UAE President Sheikh Mohamed bin Zayed and is a UAE government and Abu Dhabi ruling family member. More: - The Applied AI Co. is a developer of AI-powered software solutions
- It was founded last year by Goldman Sachs banker Arya H. Bolurfrushan and a team of "Silicon Valley engineers, applied AI experts, and investment professionals."
- According to Bloomberg, G42, an AI and cloud computing company, invested in The Applied AI Co. along with Dubai's Al Maktoum family.
- The funding is expected to help Applied AI Co. expand and build AI products for the insurance, healthcare, government, and pharma sectors.
BLOOMBERG
| |
 A message from SYNTHESIS AI What is synthetic data, and what does it mean for computer vision and generative AI? Synthetic data is computer-generated information that models the real world, but where does it apply? What are the benefits and tradeoffs, and how does it compare to real data for developing ML models? We answered these and other questions in our Synthetic Data Guide for Computer Vision. Our guide will help users better understand: - What synthetic data is in the context of computer vision
- Benefits, tradeoffs, use cases, and applications for synthetic data in machine learning
- How synthetic data is transforming the way we develop software
Learn More
| |
A new benchmark approach can help evaluate and shed more light on fast-growing AI language models. The Stanford AI Center for Research on Foundation Models (CRFM) announced the project, called Holistic Evaluation of Language Models (HELM), to "serve as a map for the world of language models" as they grow in popularity. More: - CFRM was born out of the Stanford Institute for Human-Centered Artificial Intelligence (HAI).
- So far, there hasn't been a unified standard to compare and evaluate language models and take into account their fairness, robustness, and other issues.
- HELM is intended to provide this transparency, according to CFRM, which envisions a "collaboration with the broader AI community."
- The core tenets of HELM, which will be updated regularly, include evaluating models for accuracy, calibration, robustness, fairness, bias, toxicity, and efficiency.
Zoom-out: - A team led by CRFM director Percy Liang evaluated different models on various scenarios. The models were either open-source, private, or offered via commercial APIs.
- The team found that fine-tuning the models from human feedback "is highly effective in terms of accuracy, robustness, and fairness" and allows smaller models to compete with those 10x their size.
- Open models, such as Meta's OPT and BigScience's BLOOM, still underperform non-open models such as OpenAI's InstructGPT davinci v2, though the team noted that open models "have improved dramatically over the last year."
VENTUREBEAT
| |
Ubisoft and Riot Games announced a research project to develop AI systems that are better at detecting toxic gaming behaviors online. Called Zero Harm In Comms, the project will involve the sharing of machine learning data to better identify online harassment. More: - The project will "initiate cross-industry alliances to accelerate research on harm detection," according to Yves Jacquier, executive director of Ubisoft La Forge.
- It will involve using AI to track and teach AI bots to detect toxicity during interactions in multiplayer games.
- Jacquier, who pioneered the project with Riot's Wesley Kerr, said it's a "very complex problem to be solved" to develop the best algorithm for identifying toxic content and ensuring that data shared between the companies via a framework can do that while preserving players' privacy.
- The plan is to expand Zero Harm In Comms into an industry-wide collaboration.
TECHRADAR
| |
 A message from AIRSLATE airSlate is the best way to handle document workflows. Edit PDFs, create web forms, eSign documents, or automate workflows. Build, organize, complete, and automate everything that has to do with documents. Reclaim up to 40% of the hours spent completing tasks manually. - No-code setup: Time-consuming coding is a thing of the past. Automate your document workflows with airSlate’s no-code Bots in just a few clicks.
- All-in-one solution: Create simple document processes, automate your most complex workflows, edit PDFs, and eSign your documents.
- Natively multi-cloud: Integrate with any cloud service via no-code Bots. Automatically update spreadsheets, store documents, and pre-fill agreements.
Revolutionize how your business operates. Discover what airSlate document workflow automation can do for you. Start Free Trial
| |
AI security camera company Spot AI secured $40M in financing led by Scale Venture Partners. The company provides a cloud-based platform that interprets video footage from CCTV cameras for businesses. It's designed to improve workplace safety, security, and operations. More: - Spot's system can analyze the footage for changes in activities, such as the movements of assistants in a store or damage done to a vehicle at a car wash.
- Businesses can use their own cameras or receive them free from Spot.
- The system is based on AI and computer vision tech in the cloud and, in the case of Spot's own cameras, at the device edge.
- Businesses can choose where and what they want to monitor through a drag-and-drop interface.
- Its dashboard delivers insights about the video footage, such as episodes of unsafe behavior or operation bottlenecks.
- Spot, based in Burlingame, California, exited stealth mode last year and has raised $63M to date.
TECHCRUNCH
| |
M&A Roundup: - Stellantis signed a deal to acquire AImotive, which develops advanced AI and autonomous driving software, for an undisclosed amount.
- Anumana, a Mayo Clinic AI spinout, has picked up cardiac AI company NeuTrace. NeuTrace has focused on AI applications that assess electrical signals in the heart.
- Meta is acquiring the Cambridge-based startup Audio Analytic for an undisclosed amount. Audio Analytic trained AI to recognize everyday sounds like a dog barking or glass breaking. The startup will join Meta's Reality Labs division, which is focused on building the metaverse.
- Accenture said it will acquire Japanese data science company ALBERT after closing a tender offer, absorbing its team of data scientists to help strengthen its AI capabilities.
| |
QUICK HITS - Apis Cor holds the current Guinness World Record for the largest 3D-printed building on Earth. Invest in the future of construction.*
- Elon Musk has identified a possible successor who may eventually replace him as Tesla CEO, according to a company board member, who did not name the person.
- Private AI, a provider of ML models that remove and replace personally identifiable information in datasets, secured $8M in a Series A funding round led by BDC Capital.
- The U.S. Department of Defense awarded a five-year, multimillion-dollar software licensing contract to Accrete for its threat-detecting AI software. Financial terms were not disclosed.
- Ben Taylor, former chief AI evangelist at DataRobot and a co-founder of deep learning startup Zeff.ai, was hired as ML platform Dataiku's first chief AI strategist.
*This is a sponsored post.
| |
| |
 |  | Beth is a writer and editor at Inside.com covering artificial intelligence and daily technology news. You can 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 and 1972 Fords. | |
| |
| |
Comments
Post a Comment