Skip to content

DARPA – Information Innovation Office – Environment -driven Conceptual Learning (ECOLE) BAA – HR001122S0052

Home / Funding Opportunity / DARPA – Information Innovation Office – Environment -driven Conceptual Learning (ECOLE) BAA – HR001122S0052

DARPA – Information Innovation Office – Environment -driven Conceptual Learning (ECOLE) BAA – HR001122S0052

twitterlinkedin

DARPA is soliciting proposals in the areas of human language technology, computer vision, AI, reasoning and human-computer interaction as part of the ECOLE program. ECOLE aims to create AI agents capable of continually learning from linguistic and visual input to enable human-machine collaborative analysis of image, video and multi-media documents during time-sensitive, mission-critical analytic tasks, where reliability and robustness are essential.
The program is focuses on two technical area, both of which should be addressed in proposals:
  1. Distributed Curriculum Learning – utilization of multimedia data to construct an automated curriculum for incremental learning from observance of objects to acquire knowledge of their properties and affordances.
  2. Human-Machine Collaborative Analysis – construction of a HMI to iteratively improve ECOLE representations while simultaneously expediting the analysis of multi-modal data, including multi-media and social media. The core function of the HMI will be to permit a human analyst to interact directly with ECOL’s learning representations as they are applied to the understanding of a particular set of data instances.
Abstracts due 29 September 2022; full proposals due 14 November 2022

You may be also interested in ...