(Subscribe to our Today's Cache newsletter for a quick snapshot of top 5 tech stories. Click here to subscribe for free.)
Toyota's newly-developed robot can take selfies and record how it does tasks at home.
These new capabilities are designed to amplify, not replace, human abilities, says Max Bajracharya, VP, Robotics at Toyota Research Institute (TRI).
The bots can operate in situations that are confusing to most other robots. For example, TRI notes that transparent or reflective items found at home confuse robots. They get fooled by a glass table, shiny toaster or transparent cup as most robots are programmed to react to the objects in front of them without considering the context.
To overcome this, TRI is training it's robots to perceive the scene in 3D while also enabling it to detect objects and surfaces. Researchers are using large amounts of synthetic data to train the system to learn quickly from past failures.
Also Read | Samsung's new robot can serve you a drink
Using synthetic data also diminishes the need for time consuming, expensive, or impractical data collection and labeling, researchers reckon.