The NBA's regular season ends in just a couple of weeks, but Google just surprise-dropped a major overhaul of its popular Pixel Arena feature. The original Pixel Arena was advertised as a virtual space for fans to fully experience the NBA, but there wasn't much to do other than play trivia and watch 3D shot recaps. The new Pixel Arena, however, is chock-full of activities for the discerning basketball fan, according to an official blog post. The biggest change is that users can now roam freely around the space, so the emphasis on menu-diving has been reduced in favor of natural exploration. As you roam about the virtual arena, you want to look your best for onlookers, so the avatar creation process has also been significantly overhauled.

Google has upgraded its Teachable Machine, an in-browser web experiment it originally released in 2017 to teach people some basic principles of machine learning. The original version of the app was really just a fun little experiment that could be used to classify visual data from your webcam. But Google has added new modes to the system along with the option to export trained models, making Teachable Machine 2.0 a more functional system for building actual AI tools. Along with image data, Teachable Machine now works with audio and body pose input. Users can upload their own pre-collected datasets, sort data into more than three categories, and download and deploy their models locally or host them in the cloud.

Google has upgraded its Teachable Machine, an in-browser web experiment it originally released in 2017 to teach people some basic principles of machine learning. The original version of the app was really just a fun little experiment that could be used to classify visual data from your webcam. But Google has added new modes to the system along with the option to export trained models, making Teachable Machine 2.0 a more functional system for building actual AI tools. Along with image data, Teachable Machine now works with audio and body pose input. Users can upload their own pre-collected datasets, sort data into more than three categories, and download and deploy their models locally or host them in the cloud.

A month after Google announced breakthroughs in Text-to-Speech generation technologies stemming from the Magenta project, the company followed through with a major upgrade of its Speech-to-Text API cloud service. The updated service leverages deep-learning models for speech transcription that are tailored to specific use-cases: short voice commands, phone calls and video, with a default model in all other contexts. The upgraded service now handles 120 languages and variants with different model availability and feature levels. Business applications range from over-the-phone meetings, to call-centers and video transcription. Transcription accuracy is improved in the presence of multiple speakers and significant background noise.