Trust in Autonomous Response in High Consequence Environments

The Seeker is looking for a method with the ability to assess the relative meaning and importance of new or changing information, and convey or explain the basis of recommendations in human terms such that individuals can comprehend and manage on their own.

In the increasing automated and abstracted environment, collaborative high consequence decision-making is an interactive activity though typically with individuals managing the first reaction. For example, decision-making within situations affecting safety of people, property and/or services involves humans who continuously monitor multiple information sources, analyze and assess conditions, recommend potential actions, and inform and advise decision-makers who must then choose the “right” solution. In the near future there will be an increased demand placed upon those first line individuals to react to developing situations and the demands for decision-making will increase.

The Seeker is looking for a framework for interaction between the automated system and front line individuals that builds trust and confidence through communication but does not compromise proficiency through burdensome exchanges.

This is a Reduction to Practice Challenge that contains Theoretical components that require written documentation only. A Demonstration of a solution with documentation may be accepted for the full award.

The Seeker is looking for a framework that will improve the trust of autonomous machine learning systems by the human operator. This should provide fairly quick feedback to help the operator understand/trust the process and decision making of the system.

This is a Reduction to Practice Challenge, but also has Theoretical Challenge components. Your submitted proposal should include the following:

1. Solver shall provide a detailed explanation of the idea that can meet the above Solution Requirements.
2. Solver shall provide the rationale laying out why the idea will work and how it can be tested. This rationale should address each of the Solution Requirements described above and could be supported with any relevant examples and data.
3. The Seeker may then elect to proceed and authorize a proof of concept demonstration for selected Solvers for the full award. Solvers should be clear in their submissions that they wish to participate in step 3.

Award Amount:           $ 15,000 for Delivery 1&2

                                    $ 30,000 for 1, 2 &3   

Receipt of a Challenge award is contingent upon theoretical evaluation and experimental validation of the submitted Solutions by the Seeker. If multiple proposals meet all the Solution Requirements, the Seeker reserves the right to award only the solution which Seeker believes best fits its needs. 

To receive an award, the Solvers will have to transfer to the Seeker their exclusive Intellectual Property (IP) rights to the solution. However, the Seeker will be willing to consider a licensing agreement for a partial award if exclusive IP cannot be transferred by the Solver. 

Submissions to this Challenge must be received by 11:59 PM (US Eastern Time) on February 15, 2019. Late submissions will not be considered.