Based on an evaluation of complementary strengths we believe there is a rapid path not only to better reading comprehension and global translation, but also being able to summarize and translate reliably, from handwritten notes (which will be important for Augmented Reality applications). HUMAN Protocol already produces more labeled image data than the whole of Stanford’s ImageNet on a daily basis.
Our goals are:
- 1.Read, summarize and translate from many forms of communication, accurately
- 2.Reduce bias & racism in algorithms.
- 3.Improve communication between humans and AI.
We are an open-source, cloud-first organization aimed at creating shared, public global resources for machine learning. We believe time is of the essence and are community and collaboration-focused.