Topic

TECHNOLOGY AREAS: Trusted AI and Autonomy; Advanced Computing and Software; Hypersonics; Integrated Network System-of-Systems; Quantum Science; Advanced Materials; Advanced OBJECTIVE: The objective of this Phase I STTR project is to establish a foundational framework for a Supply Chain Intelligence Platform tailored to the Defense Industrial Base (DIB). In Phase I, the focus will be on demonstrating the technical feasibility of two core components: (1) automated data extraction from open sources, and (2) enhanced entity disambiguation within a knowledge graph. These components will set the stage for scalable and real-time data processing capabilities that are critical for subsequent phases of development. This project aims to lay the groundwork for a scalable system that can later incorporate advanced features such as predictive analytics and knowledge graph optimization in Phase II. A successful Phase I effort will demonstrate that AI-driven data collection and entity resolution techniques are not only technically feasible but also crucial for enhancing the Air Force’s ability to make informed, timely decisions about its supply chain logistics, starting with use cases such as the KC-46 platform. Beyond Phase I, the longer-range end state of this project is a fully operational Supply Chain Intelligence Platform integrated within the Earth 616 framework. This platform will autonomously collect and analyze data, enhancing the DIB’s data integrity and decision-making capabilities. The project will deliver AI intelligence assistant capability that provides analysts with comprehensive, accurate, and actionable insights, ultimately enhancing decision-making, risk analysis, and response times. The main objective is to automate and scale the collection and integration of diverse data sources into a supply chain intelligence platform. Without this capability, knowledge graphs used on a large scale, such as the DIB supply chain, become unwieldy in query performance and suffer from duplicative data entities. This is essential to address the operational gaps in supply chain management for the DoD. DESCRIPTION: The STTR project aims to enhance the Department of Defense’s (DoD) supply chain visibility and decision-making capabilities through a comprehensive integration of advanced data analytics technologies. This project represents a significant innovation by introducing automated entity disambiguation within complex datasets, which has not been fully realized in Earth 616’s current capabilities. Additionally, the real-time optimization of knowledge graphs represents a novel technological leap in how the Air Force can manage its supply chain data at scale. By reducing the need for manual interventions and enhancing the scalability of data analysis, this STTR project pushes beyond the current state of the art in supply chain intelligence. Here is a detailed summary of the work proposed to meet these objectives: Automated Data Collection: The project will develop and deploy AI-driven agents capable of autonomously collecting data from diverse sources such as APIs, web scraping, and existing databases. These agents will leverage advanced Large Language Models (LLMs) to accurately identify, collect, and classify critical supply chain information, ensuring a continuous flow of data. This capability will significantly reduce the reliance on manual data entry and increase the speed and accuracy of data available for analysis. Entity Disambiguation: Utilizing machine learning algorithms, the project aims to implement sophisticated models that can accurately disambiguate and categorize entities across different data sources. This will prevent data duplication and enhance the reliability of the information within the DoD’s supply chain knowledge graph. Accurately categorized data is crucial for effective risk management and strategic planning. Knowledge Graph Refactoring: The project will develop algorithms designed to continuously analyze and optimize the performance of the knowledge graph. These algorithms will recommend restructuring steps to maintain the graph’s efficiency and relevance as more data becomes available. This will support scalable and dynamic supply chain analysis, enabling the DoD to adapt to changing conditions and requirements quickly. Integration and Testing: All developed components—data collection agents, disambiguation models, and refactoring algorithms—will be integrated into a single operational system. This system will be rigorously tested to ensure high stability, reliability, and compatibility with existing DoD data infrastructures. Deployment and Evaluation: The final stage of the project involves the deployment of the Supply Chain Intelligence Platform within selected DoD environments. This platform will be specifically integrated into the Earth 616 initiative, enhancing its data processing and analysis capabilities. Post-deployment, the system will undergo thorough evaluation to ensure it meets all predefined performance benchmarks and effectively enhances the DIB’s data integrity and decision-making capabilities. Documentation and Training: Comprehensive documentation will be prepared to detail the operational procedures, maintenance, and troubleshooting of the deployed system. Additionally, tailored training programs will be developed to ensure that DoD personnel can effectively utilize the new platform. Collaborative Research and Development: Incorporate cutting-edge research in AI, machine learning, and data analytics into the platform to ensure the project remains at the forefront of technological advancements and is equipped with the latest methodologies and tools. The KC-46 is the initial use case for this project, providing a focused, real-world application to demonstrate how the solution will improve logistics management, readiness, and operational decision-making for the Air Force Materiel Command (AFMC). The platform will automate data ingestion and analysis, leading to improved supply chain visibility, faster maintenance, and reduced delays in resource allocation for critical systems. By focusing first on the KC-46, the project provides a clear and specific demonstration of the technology’s potential to improve logistical efficiency across larger Air Force platforms. PHASE I: The primary focus of the Phase I effort will be on developing and demonstrating the feasibility of automated data extraction and entity disambiguation within the knowledge graph, which are essential building blocks for the larger Supply Chain Intelligence Platform. The project will focus on three main tasks: Automated Data Extraction: The project will develop an AI-driven mechanism to autonomously extract relevant data from open sources, including government databases, public procurement platforms, and commercial sources. By leveraging advanced language models and natural language processing (NLP) techniques, this system will be able to parse complex documents such as compliance reports and procurement records to extract critical supply chain information. The system’s effectiveness will be measured by its ability to collect data accurately and efficiently, significantly reducing the time and effort required for manual data entry. The primary deliverable for this objective is a working prototype capable of demonstrating real-time data extraction from multiple open sources. Entity Disambiguation within the Knowledge Graph: The project will implement machine learning algorithms that refine the process of entity disambiguation, ensuring that entities (e.g., suppliers, products, and contracts) from diverse data sources are accurately categorized and integrated into a unified knowledge graph. This is critical for avoiding data redundancy, improving query performance, and maintaining the integrity of the supply chain database. The Phase I deliverable will include a prototype model that demonstrates improved entity resolution across at least two types of datasets, such as vendor details and manufacturing records, while ensuring that the knowledge graph remains efficient and relevant. Integration and Proof of Concept: All developed components—data extraction mechanisms and entity disambiguation models—will be integrated into a proof-of-concept system. This prototype will be designed to interface with existing DoD infrastructure (such as Earth 616), focusing on scalability and ease of integration. The system will be tested in a controlled environment using real-world data from relevant sources like KC-46 suppliers, ensuring that the AI-driven solutions can function effectively within the complexities of the DIB. Success criteria for this Phase I effort will include achieving an acceptable level of accuracy in data extraction and entity resolution and demonstrating the potential for scalability in a future Phase II effort. The expected outcome for Phase I is the development of a proof-of-concept system that demonstrates the ability to autonomously extract data from open sources and accurately categorize entities within a knowledge graph. The feasibility of this technical component will be the key to justifying the follow-on phases. The goal is to establish the system’s ability to scale, automate data categorization, and demonstrate the platform’s potential for integration into broader AFMC operations. The technical foundation laid in Phase I will be crucial for developing the more complex predictive models and knowledge graph refactoring algorithms in Phase II. Phase I will span three months and include the following milestones: Month 1: Complete the design and initial testing of the automated data extraction prototype from open sources. Month 2: Develop and refine entity disambiguation models, with initial testing conducted on sample datasets relevant to KC-46 suppliers. Month 3: Integrate both systems into a proof-of-concept platform and conduct end-to-end testing within a controlled environment, ensuring feasibility for scalability in Phase II. These objectives will lay the foundation for Phase II, which will expand the platform to include predictive analytics and knowledge graph refactoring for scale. The Phase I effort will focus on the development of a proof-of-concept system capable of collecting data autonomously and categorizing it efficiently. This limited scope ensures feasibility within the three-month timeframe and aligns with the resources available in Phase I PHASE II: Building upon the successful demonstration of data extraction and entity disambiguation in Phase I, Phase II will expand the platform’s capabilities. While Phase I focuses on developing the fundamental data extraction and entity disambiguation components, Phase II will expand these capabilities into a fully scalable system. Specifically, Phase II will introduce real-time data processing with low latency and the development of dynamic knowledge graph refactoring algorithms, ensuring that the system can handle expanding datasets and more complex relationships across the Defense Industrial Base (DIB). These algorithms will be essential for maintaining query performance and data integrity as the platform scales. Objective 1: Develop an Automated Data Collection System. Performers should design an automated data collection system that is highly efficient in ingesting data from a variety of sources such as DoD systems, supplier databases, and open web sources. The system must incorporate advanced AI technologies capable of extracting and preprocessing large volumes of data using a content-based hashcode similar to blockchain mechanisms. This should ensure that each piece of data is mapped accurately to an Earth 616-like ontology, facilitating seamless integration into the knowledge graph (KG) database. The objective is to achieve real-time data processing with high accuracy and minimal latency. Objective 2: Enhance Entity Disambiguation Models. This project should focus on developing machine learning models that robustly disambiguate and categorize entities within complex datasets. The models must be capable of leveraging distributed credential and attribute-based policy frameworks to effectively manage and permission graph queries. This ensures that data integrity and security are maintained while providing tailored data access based on user credentials, enhancing both the security and usability of the supply chain data. Objective 3: Implement Scalable Knowledge Graph Refactoring. The project should include the creation of dynamic algorithms for refactoring and optimizing knowledge graphs to handle expanding datasets effectively. The solution should utilize a high-performance distributed cluster architecture, akin to Neo4j’s capabilities, which allows for connecting datasets across billions of nodes and trillions of relationships with millisecond response times. This ensures that the knowledge graph can scale dynamically and maintain performance even under the stress of large-scale operations. System Integration and Prototyping. Integration of all developed components into a cohesive system that conforms to military standards for software and hardware performance is crucial. The system must be designed to integrate seamlessly with existing Air Force IT infrastructures, ensuring compatibility and interoperability. Prototyping should include simulation environments that mirror real-world operational conditions to validate system performance across a range of scenarios. Deployment and Operational Testing. The deployment phase should focus on operational environments that will benefit significantly from enhanced data analytics capabilities. This includes conducting extensive field tests to evaluate the system’s impact on operational efficiency and decision-making accuracy in real-time conditions. Metrics for success should be clearly defined, with benchmarks set for data processing speeds, accuracy of analytics outputs, and user satisfaction. User Feedback and Iterative Improvement. Establishing a continuous feedback loop with end-users is essential for refining the system. This involves systematic collection and analysis of user interactions and satisfaction to guide ongoing improvements. The feedback process should be structured to allow quick iterations of the software to adapt to new requirements or address any issues arising during operational deployment. A successful Phase II effort will deliver the following: ● A robust automated data collection system capable of integrating data from multiple, diverse sources into a unified, actionable format. ● Advanced entity disambiguation models that ensure high levels of data integrity and usability. ● A dynamically refactored and scalable knowledge graph that supports comprehensive and timely supply chain analysis. ● An integrated prototype tested and proven in operational Air Force environments, demonstrating significant improvements in data-driven decision-making and operational efficiency. A successful Phase II effort under this topic will deliver an integrated suite of advanced data collection, analysis, and knowledge management tools that will significantly enhance the predictive analytics and operational decision-making capabilities of the Air Force. These technologies will provide the Air Force with unprecedented capabilities to manage and utilize data effectively, ensuring readiness and strategic advantage in various operational scenarios PHASE III DUAL USE APPLICATIONS: Phase III will take the mature capabilities developed in Phase II—including real-time data processing, advanced entity disambiguation, and scalable knowledge graph refactoring—and integrate them into operational environments like Earth 616. By this stage, the technology will be at TRL 6, and Phase III will involve rigorous testing and validation under real-world conditions to advance the system to TRL 9, making it fully operational for deployment across Air Force platforms. A roadmap for the expected Phase III efforts is as follows: Technology Maturation and Validation: At the entry of Phase III, the technology will be at TRL 6, indicating that a prototype has been demonstrated in a relevant environment. Phase III will focus on further refining these prototypes and validating them in operational environments to address specific Air Force requirements, particularly within logistics systems subject to adversarial threats, as identified in the “Conducting Logistics While Under Attack” strategy. The goal is to advance the technology to TRL 9, signifying that the system is fully operational and can be deployed across various platforms. This includes extensive testing under real-world conditions to ensure the technology can withstand operational demands without substantial modifications. Integration with Earth 616 STRATFI Effort: A critical component of Phase III will be the integration of this technology into the existing Earth 616 STRATFI initiative, which aims to bolster the defense industrial base and optimize the operational readiness of Air Force platforms such as the KC-46. The advancements made in automated data collection and entity disambiguation during Phase I will be leveraged to enhance Earth 616’s capabilities, providing predictive analytics to anticipate logistical challenges and optimize resource allocation. The incorporation of advanced data integrity and supply chain analytics will help secure the operational integrity of vital systems under conditions of contested logistics. Transition to AF Futures Command Integrated Capabilities Command (ICC): Upon achieving TRL 9, the project will transition to the Air Force Futures Command’s Integrated Capabilities Command (ICC). This transition will mark a key milestone in the technology’s full-scale deployment and integration across the Air Force Materiel Command (AFMC) and potentially other DAF platforms. The ICC will oversee the scaling of the technology across multiple domains, ensuring operational readiness and resilience across different Air Force logistics environments. This step is crucial for leveraging the full capabilities of the technology to support DAF’s “Readiness to Deploy and Fight” imperative. Operational Deployment and Scaling: The deployment strategy in Phase III will focus on ensuring the technology’s scalability and adaptability across multiple Air Force bases and operational platforms. The autonomous data collection capabilities, entity resolution algorithms, and knowledge graph optimization tools developed during the earlier phases will be scaled to meet the needs of a broader range of Air Force platforms, enhancing decision-making and risk management capabilities. The system’s adaptability will be a key factor in achieving uniform operational readiness and rapid deployment capabilities across geographically dispersed bases and units. Government Approvals and Regulations: Throughout Phase III, the project will navigate necessary government approvals and regulatory requirements to ensure compliance with all federal and military standards. This includes obtaining certifications for cybersecurity, secure data handling, and privacy. As part of the transition planning, the project team will work closely with government bodies to ensure that the system meets all regulatory requirements and is approved for operational deployment within military networks. Key milestones will include certifications that validate the system’s security posture, ensuring its resilience in contested environments. Additional DAF Customer Opportunities: Beyond the initial application within the Earth 616 framework and KC-46 SPO, Phase III will focus on extending the technology’s application across other platforms within the Air Force Materiel Command and beyond. The system’s advanced data management and predictive analytics capabilities will enhance logistical support and operational planning for other critical platforms, ensuring the Air Force’s ability to sustain operations in increasingly complex and contested logistical environments. Opportunities will also be explored to integrate the technology into other DoD service branches and allied systems, ensuring a broader impact across national defense supply chains. Commercialization Potential: Although this is a defense-focused project, the potential for commercialization in sectors such as commercial aviation and critical infrastructure is evident. AI-driven supply chain intelligence could be used to impro REFERENCES: 1. H. S. Magableh and M. A. Al-Naymat, “Integrating Blockchain with ERP for a Transparent Supply Chain,” Information, vol. 11, no. 2, p. 87, 2020; 2. K. Baryannis, D. Validas, and S. Dani, “Predictive analytics for complex supply chain networks using machine learning and graph theory,” Computers & Industrial Engineering, vol. 136, pp. 358-369, 2019; 3. C. Bode and S. M. Wagner, “Structural drivers of upstream supply chain complexity and the frequency of supply chain disruptions,” Journal of Operations Management, vol. 36, pp. 215-228, 2015 KEYWORDS: Supply Chain Intelligence, Automated Data Collection, Entity Disambiguation, Knowledge Graph Refactoring, Advanced Computing, Machine Learning, AI-Driven Analysis, DoD Supply Chain Management, Predictive Modeling, Data Integrity