By 
On October 16, the U.S. Army posted an Industry Day Announcement for the 3rd Annual DoD Artificial Intelligence (AI) Industry Day. The event will take place on November 13, and capability statements from attendees interested in 1:1 meetings are due no later than November 5.
The Army Research Laboratory (ARL), together with the Algorithmic Warfare Cross-Functional Team (AWCFT or “Project Maven”) and the Joint Artificial Intelligence Center (JAIC), are hosting the 3rd Annual DoD AI Industry Day on November 13, 2019 to discuss AI software prototyping activities and future industry participation. All technical businesses engaged in AI/ML are invited to submit information on their capabilities (see website listed below for more information).
The ARL Broad Agency Announcement (BAA) for Basic and Applied Scientific Research added the “Artificial Intelligence and Machine Learning” (AI/ML) topic in 2017 under the Information Sciences Campaign. The Artificial Intelligence and Machine Learning topic can be found by searching announcement “W911NF-17-S-0003” at Grants.gov or FBO.gov.
Over the course of the program, Project Maven and the JAIC have developed a broad range of nascent algorithms using a variety of approaches that have processed tremendous volumes of DoD data to generate a new repository of algorithmic output and metadata. In addition to pursuing innovative ideas to process the raw DoD data, Project Maven and the JAIC are also seeking innovative AI approaches on how to analyze the newly generated AI output and library of AI algorithms.
In support of Project Maven and the JAIC AI prototyping, the DoD is holding an AI Industry Day focusing on the following areas:
1. Training Data: Standardizing, cleansing, preparing, and managing data for AI algorithm training, including new labeling techniques and tools for analytics and metrics of training data.
2. Data and Inference Platforms: New or improved methods for aggregating joint all-domain data sources across all warfighting functions in scalable development and production cloud platforms.
3. Algorithms: Frameworks and tools for creating AI/ML algorithms, and the algorithms themselves, for AI prototyping against raw data as well as against AI generated metadata.
4. User Interface (UI)/User Experience (UX): Interfaces to combine traditional, foundational intelligence data and AI; run the AI in inference against raw intelligence data; and provide display, search, and analysis capabilities across all combined data at scale.
5. Integration at the Edge: New or improved processes and methods for integrating AI/ML within systems at the edge to bring AI/ML to constrained, or RF-denied, computational environments; to include multi-sensor/sensor fusion innovations incorporating AI for on-platform understanding and platform automation.
6. Testing: New methods to test, evaluate, and determine the effectiveness of AI/ML approaches, including comparisons between approaches.
Event and Registration Details: The one-day Industry Day will consist of presentations in the morning followed by an opportunity for networking and 1:1 panel sessions in the afternoon. Attendees are limited to two (2) representatives per company. Attendees can request an optional 1:1 panel session during registration. For 1:1 panel sessions, attendees must submit a capability brief/presentation via e-mail to AI@arl.army.mil no later than Tuesday, 5 November 2019. Changes made to the capability brief/presentation after 5 Nov 2019 can be explained in person or with hardcopies during the 1:1 panel session. For those wanting to submit capability briefs/presentations but are unable to participate in a 1:1 panel session or are unable to attend AI Industry Day, please submit via email to AI@arl.army.mil no later than 14 November 2019. Refer to the following website for Event and Registration Details as well as 1:1 panel session construct and capability brief/presentation format: http://www.cvent.com/d/3hqmfr. All submissions must be in accordance with the guidance provided at this website.
Full information is available here.
Source: FedBizOpps
