The California PUC granted Waymo a permit to operate 24 hours/day in San Francisco taking select members of the public for rides with no safety driver in the vehicle. Waymo says it will begin this shortly. This comes on the heels of them expanding such service in Phoenix, as reported in my article on how the death of self-driving cars has been greatly exaggerated earlier this week.

This service will be with “trusted testers” rather than members of the broad public that can ride in Chandler and Phoenix, Arizona.

Waymo gets a permit to take the public and will start running with trusted testers. They also start in LA and show off their new robotaxi. Cruise also ups its game. Zoox got into an accident but it was in manual. Aurora reminds us they do more than trucks.

Recent progress in generative models has paved the way to a manifold of tasks that some years ago were only imaginable. With the help of large-scale image-text datasets, generative models can learn powerful representations exploited in fields such as text-to-image or image-to-text translation.

The recent release of Stable Diffusion and the DALL-E API led to great excitement around text-to-image generative models capable of generating complex and stunning novel images from an input descriptive text, similar to performing a search on the internet.

With the rising interest in the reverse task, i.e., image-to-text translation, several studies tried to generate captions from input images. These methods often presume a one-to-one correspondence between pictures and their captions. However, multiple images can be connected to and paired with a long text narrative, such as photos in a news article. Therefore, the need for illustrative correspondences (e.g., “travel” or “vacation”) rather than literal one-to-one captions (e.g., “airplane flying”).