A naukri.com initiative
Mit
2M
307
Image Credit: Mit
A faster, better way to train general-purpose robots
- MIT researchers developed a versatile technique that combines a huge amount of heterogeneous data from many of sources into one system that can teach any robot a wide range of tasks.
- Their method involves aligning data from varied domains, like simulations and real robots, and multiple modalities, including vision sensors and robotic arm position encoders, into a shared “language” that a generative AI model can process.
- This approach can be used to train a robot to perform a variety of tasks without the need to start training it from scratch each time.
- This method could be faster and less expensive than traditional techniques.
- Their architecture called Heterogeneous Pretrained Transformers (HPT) that unifies data from these varied modalities and domains.
- The researchers align data from vision and proprioception into the same type of input, called a token, which the transformer can process.
- The larger the transformer becomes, the better it will perform.
- When they tested HPT, it improved robot performance by more than 20 percent on simulation and real-world tasks, compared with training from scratch each time.
- This approach enables robot learning methods to significantly scale up the size of datasets that they can train on.
- The researchers are working towards creating a universal robot brain that can be downloaded and used for robots without any training.
Read Full Article
18 Likes
For uninterrupted reading, download the app