Diffeo Labs seeks to go beyond the current state of the art in collaborative machine intelligence. What will collaboration look like 5 or 10 years from now? Can AI transform the way we work, both with technology and with each other? How will reductionism bridge the gap between machine learning tools and the emergence of intelligence? Our team uniquely brings together machine learning, cognitive science, and physics in pursuit of the next generation of machine intelligence for our world.
Diffeo started as a computer science research project combining natural language understanding and information retrieval in the Text Retrieval Conference, where Diffeo Labs organized and ran the Knowledge Base Acceleration evaluation from 2012 to 2014. Subsequently, we contributed to DARPA’s Memex program with a suite of open source tools and modeling insights, including the collaborative highlighting tool used in Diffeo’s applications. As part of Memex, we helped launch the TREC Dynamic Domain evaluation, which is ongoing. Many of our internal experiments use truth data from these efforts to simulate human-machine collaboration.