The global household robots market size was valued at 10.3bn ( 7.7bn) in 2023 and is anticipated to hit 24.5bn by 2028, meaning such devices are an increasingly common sight in people's homes. Anyone who has watched a robot vacuum cleaner in action may argue these ideas are a little far-fetched, given that current machines sometimes struggle with the challenges presented by rugs and shoelaces while carrying out their core function. However, scientists from the University of Bath and the University of Calgary in Canada, have set out to prove that cleaners - and similar devices, such as lawnmowers - could be reprogrammed and modified relatively easily. Their study identified 100 functions the robots could possibly perform with simple adjustments. Other proposed tasks suggested by the scientists include a reprogrammed robot that carried the groceries from the car to the kitchen.

Ultra-realistic images created by AI show what your dinner could look like in 30 years' time as we're forced to eat'sustainability'. Experts have used AI tool Midjourney to bring to life the menu of 2054, which features bizarre dishes such as cricket salad and lab-grown steaks. There's even green spaghetti and'meat' balls made out of an aquatic plant, which look straight from the kitchen of another galaxy. These unusual creations could replace family favourite dishes such as the traditional Sunday roast or fish and chips, the scientists believe. They have lower carbon footprints than such classics, which means they could help in the battle against climate change – but would you eat them?

When planning with an inaccurate dynamics model, a practical strategy is to restrict planning to regions of state-action space where the model is accurate: also known as a model precondition. Empirical real-world trajectory data is valuable for defining data-driven model preconditions regardless of the model form (analytical, simulator, learned, etc...). However, real-world data is often expensive and dangerous to collect. In order to achieve data efficiency, this paper presents an algorithm for actively selecting trajectories to learn a model precondition for an inaccurate pre-specified dynamics model. Our proposed techniques address challenges arising from the sequential nature of trajectories, and potential benefit of prioritizing task-relevant data. The experimental analysis shows how algorithmic properties affect performance in three planning scenarios: icy gridworld, simulated plant watering, and real-world plant watering. Results demonstrate an improvement of approximately 80% after only four real-world trajectories when using our proposed techniques.