Google DeepMind created an algorithm to speed up the process of training artificial intelligence for complex games.
The algorithm, called AlphaDev, is capable of creating and running variations of existing algorithms, reducing the time required for training by up to 100 times.
Researchers hope AlphaDev will be applied in various sectors of artificial intelligence.
- Researchers created an algorithm called AlphaDev that trains AI for complex games.
- AlphaDev can create and run variations of existing algorithms which significantly reduces training time
- A 100 times reduction in training time is possible with AlphaDev
- AlphaDev will be used to improve AI in various sectors
- Enables game developers to train an agent on sample gameplay faster.
- It’s amazing how a small improvement at the base machine code level can have a planet wide impact.
- In this case, improving baseline encryption 30% faster – that happens on every device on the planet every second of every day…
- Wow – what other under-the-hood improvements are just waiting to be discovered like that.
The AI-generated algorithms are already being used by millions of developers.