AFRL updates interactive learning BAA
On May 21, the Air Force Research Laboratory (AFRL) posted an updated broad agency announcement (BAA) for Interactive Learning for Mission Planning.
This announcement is for an Open, 2-Step BAA open and effective until 30 Sep 2023. Only white papers will be accepted as initial submissions; formal proposals will be accepted by invitation only. While white papers will be considered if received prior to 4 PM Eastern Standard Time (EST) on 30 Sep 2023, the following submission dates are suggested to best align with projected funding:
Offerors are requested to hold white papers until FY23.
FY23 by 31 Mar 2022
Seeking innovative applications of Interactive Learning techniques and technologies to Air Force (AF) planning problems, particularly in the realm of tactical route planning. Do note that this program is domain agnostic and white papers detailing work in other domains including strategic and tactical air, space, cyber, etc. will be entertained.
The Interactive Learning for Mission Planning (ILMP) program will test the merit of Interactive Learning (IL) when applied to AF planning problems, and will develop the software system required to prove the applicability of IL to this problem space. This program will explore the union of IL, current AF automated planning tools, such as tactical route planning tools, and human planning Subject Matter Experts (SME). The ILMP team is soliciting white papers that propose methods for leveraging IL in planning domains.
Proposed interactive algorithms will gather and learn from SME feedback to produce plans more amenable to SME preferences. Such approaches will gather feedback from SMEs efficiently and effectively; for example, limit the work to be done by SMEs, and gather sufficient feedback from SMEs to adequately reflect preferences in produced plans. Selected planners are to be mature and well supported, and will be minimally changed. Approaches will rely on SME feedback and machine learning to improve plan quality. Additionally, proposals will outline experiments to test the efficacy of learner-planner-SME combinations in sufficiently complex domains.
Full information is available here.