Google has demonstrated a new artificial intelligence system that can create images based on text input. The Imagen publication model, created by the Brain team at Google Research, provides an “unprecedented degree of image realism and a profound level of language understanding.”
This isn’t the first time we’ve seen AI models like this. OpenAI’s DALL E (and its successor) did a similar magic, turning text into images. However, Google’s version is trying to create more realistic images. The researchers created a benchmark and asked humans to rate each image from an array of AI systems. Google said they “prefer Imagen over other models in side-by-side comparisons, both in terms of sample quality and image text alignment.”
It’s not publicly available, and there are reasons for that. “Data sets of this nature often reflect social stereotypes, oppressive viewpoints, and degrading, or otherwise harmful, associations with marginalized identity groups,” the researchers wrote. Imagen has inherited “social biases and the limitations of large linguistic models” and has depicted “stereotypes and representations.” The team said the AI encodes social biases, including the tendency to create images of people with lighter skin and place them in certain gender stereotyped roles. The system can be used to create hateful images to intentionally cause abuse.
The team may eventually allow the rest of us to play with the model to generate our own images, but the researchers need to think about the framework first — a challenge in itself.
– Matt Smith
The biggest stories you may have missed
But it costs more than the standard Starlink service.
Starlink Internet service for RV drivers and “vanlife” types direct. While applying for Starlink Platter and the service will put customers on a waiting list until 2023, Starlink for RVs is immediately available and will be shipped to buyers right now. However, network resources for RV connections are always prioritized, and the service costs $135, which is $25 more than a regular Starlink connection.
Uses NVIDIA’s Reflex technology to reduce system latency.
ASUS has unveiled what it calls its “world’s first” 500Hz G-Sync, 1080p ROG Swift 500Hz gaming monitor. Designed for competitive gaming, it uses a proprietary panel and incorporates NVIDIA’s G-Sync Esports technology for increased motion clarity. It also uses NVIDIA’s Reflex Analyzer technology, which provides real-time statistics to help you reduce end-to-end latency if you’re using a mirror-optimized mouse and NVIDIA GPU.
Its main highlight remains the 500Hz refresh rate, which pulls eight times faster than typical 60Hz displays.
Good price, fast but lacks lit/stacked sensors.
Canon launched its first EOS R APS-C crop sensor cameras, the 32MP EOS R7 and the 24MP EOS R10. New models align Canon’s APS-C series of cameras and the full-frame RF series, so you can finally Use the lenses interchangeably. More importantly, it carries impressive specs, such as mechanical shutter shooting speeds of 15fps and 4K video of up to 60fps.
The EOS R10 costs $980 for the body only, $1,100 for the RF-S18-45mm lens, and $1,380 for the RF-S18-150mm lens. Meanwhile, the R7 will sell for $1,500 for the chassis only and $1,900 for the S18-150mm lens. Both should arrive later this year.
Direct participation is a major development of remote work.
Microsoft’s new Live Share feature should make it easier for Teams apps to enable real-time collaboration. If that sounds familiar, it’s because Microsoft announced plans to make Teams the go-to option for collaborative apps last year. Direct Sharing is based on the Fluid Framework, Microsoft’s attempt to break up traditional document components and make them collaborative. Microsoft says several partners, including Accenture, Frame.io, and Hexagon, are already building Live Share experiences into Teams projects.
But they are for data centers.
Later this year, NVIDIA will begin selling a liquid-cooled version of its A100 GPU to data centers. The GPU maker is positioning the video card as a way for cloud computing companies to make their facilities more energy efficient. NVIDIA claims that a facility equipped with water-cooled A100 GPUs runs the same workload as an air-cooled data center while using 30 percent less power.
All products recommended by Engadget are handpicked by our editorial team, independently of the parent company. Some of our stories include affiliate links. If you buy something through one of these links, we may earn an affiliate commission.