利用人工智能增强作战人员的能力.pdf

1、Empowering the Warfighter with AI/MLTeneika AskewJune 2025zs8 2-K ZZ8 G834$1,000,000,000 24 hours Restrictions:No NecessitiesNo Debt PaymentsNo Savings or Wall St.InvestmentsNo Direct Donations to CharitiesNo Buying Gold or Precious MetalsNo Purchase of Personal Luxury ItemsUS Spending Onlyz z-5 Str

2、ategic Investments:Large-Scale Infrastructure ProjectsResearch and DevelopmentEducation InitiativesTechnology and InnovationCommunity DevelopmentPublic Health CampaignsCultural Investments Long Term Impact ROI Stakeholder Benefits M _Z 2 Impact of Machine LearningDe-obligation Use CaseRisk Based Sam

3、pling Use CaseAutonomous Auditing1234 G8 2 68 v GZ -87$190M+Cleared in Unmatched Transactions257KHours saved using AI/ML$211BULOs reviewed&validated by BSOs86%Reduction in ML Model Runtimes$400BEV NQE BSO Compute Cost$9.1MLabor Cost Savings as a result of AI/ML$2.2B+De-obligations from RBS BEV Sampl

4、ing$649MActive Fund De-obligations$306M+Adjustments Identified in 1st Quarter Expired Phase-2024$2,700RBS Yearly BSO Compute Cost$80%Less samples requiring review$1.25MDecrease in Job CostsKf 8 8+M Z E_ _f 28Zz8DatabricksModel Development WorkflowData AcquisitionBuild&Run PipelineValidateModel Deplo

5、yment WorkflowTest&DeployMonitorETL&Feature TransformationTrain ModelEvaluate ModelBuild&Run PipelineData ScientistPull/Push Pipeline CodeHostingModel RegistryEndpointData Scientists&Active UsersUnofficial Feature StoreGitLab z8-z892 10Goal:Catch funds BEFORE they expire at the latest,1st Quarter of

6、 Expiration!2 Using advanced analytics to identify systemic trends driving unexecuted/cancelled obligation11Goal:Build a data science model to identify documents at risk of expiring and in need of an adjustment.Provide a probability score,utilizing the first 5 months(at mid-year)of data and updates