AI and machine learning have the potential to help solve many issues resulting in climate change through efforts in key areas that leverage the power of AI. Such areas include the development of clean power, smart-transportation, cities and homes, sustainable production as well as sustainable land-use. These segments can contribute positively towards climate change. Here’s a
few of the things Microsoft is doing:
Smart Grid: ‘the internet brought to our electric system’
Power: AI and cloud technologies are being used to improve the electrical efficiency of buildings by centralising these operations on cloud
Irrigation & Farming: e.g. remotely monitoring, managing and analysing water usage
Lego Mindstorms have paved the way for many programmable toys. And Austrian startup Robo Wunderkind is building a new kind of Lego-like programmable kit. The startup first launched on the TechCrunch Disrupt stage and just raised $1.2 million (€1 million) from SOSV, Austrian Federal Promotional Bank and multiple business angels.
Robo Wunderkind also has special blocks to turn your dumb robot into a connected one. In addition to the usual sensors, such as proximity sensors, motion detectors and light sensors, the company also has some more sophisticated ones. You can put a tiny camera in your construction, use an IR blaster and receiver and program a tiny LED screen. But the best part is that Robo Wunderkind also sells Lego adapters so that you can put together a sophisticated robot that uses both Lego bricks and Robo Wunderkind modules
MIT researchers have developed a computer interface that can transcribe words that the user verbalizes internally but does not actually speak aloud.
The system consists of a wearable device and an associated computing system. Electrodes in the device pick up neuromuscular signals in the jaw and face that are triggered by internal verbalisation — saying words “in your head” — but are undetectable to the human eye. The signals are fed to a machine-learning system
that has been trained to correlate particular signals with particular words.
In the present work, researchers from DeepMind and UCL draw on recent advances in artificial intelligence to introduce a new theory of reward-based learning. Here, the dopamine system trains another part of the brain, the prefrontal cortex, to operate as its own free-standing learning system.
This new perspective accommodates the findings that motivated the standard model, but also deals gracefully with a wider range of observations, providing a fresh foundation for future research.
OpenAI launched a transfer learning contest that measures a reinforcement learning algorithms ability to generalise from previous experience. In typical RL research, algorithms are tested in the same environment where they were trained, which favours algorithms which are good at memorisation and have many hyperparameters.
Instead, their contest tests an algorithm on previously unseen video game levels. This contest uses Gym Retro, a new platform integrating classic games into Gym, starting with 30 SEGA Genesis
Stealthy startup 6d bytes came out of hiding yesterday to debut Blendid, a fully autonomous, smoothie-making robot station. Using a combination of machine learning, machine vision and robotics, Blendid promises to deliver customisable and consistently tasting smoothies to consumers whenever they want.
Blendid is an all-in-one robotic installation that can be set up on-site in food service operations. Inside the kiosk, the robot system blends, pours and serves up smoothies. No humans are needed, except for the site operator to manage inventory and address any issues that may arise.
In early April, California’s new rules that allow automakers, tech giants and just about anybody to test fully driverless cars on its roads finally took effect. But before those companies can realize their ride-hailing robot taxi ambitions, they have to wait for the state to adopt a proposal issued by the California Public Utilities Commission. The public utility regulator’s proposed rules would allow autonomous vehicles to give rides to the public as part of a pilot program — that is, so long as their creators meet a few conditions.
To start with, only cars with backup drivers can initially take passengers. That won’t be a problem, since only one (unnamed) company has applied for permission to test no-driver cars in California, thus far. However, the rules are expected to extend to fully driverless cars in the future. Companies must also provide the service for free — the CPUC plans to develop regulations for paid rides, but that’ll come after this initial set of rules get approved.