NAVSEA Open Topic for Model-Based Systems Engineering (MBSE) Acceleration via Artificial Intelligence or Machine Learning Data Ingestion

Q&A section updated

# Q&A Changes Summary **New Question Added:** - Q1: Clarifies input document formats for Phase II prototype—raster scans of paper/microfilm (including handwritten redlines), born-digital PDFs, or native CAD files (DWG, DGN, Visio). Asks whether paper originals will be furnished as scanned images or if scanning capability is required. **Structural Change:** - Original Q1 (multi-part technical question on GFI, Cameo licenses, SysML/Cameo versions, and modeling standards) renumbered to Q2 with identical content. - Original Q2–Q19 shifted down one number accordingly. **Net Result:** One substantive new question focused on **input data formats and scanning requirements** for Phase II deliverables; all other questions retained without content updates.

# Q&A Changes Summary **New consolidated Q1** combines five critical technical requirements previously scattered across multiple questions: - Government-furnished documentation availability, classification level, and timeline - Cameo/CATIA Magic license provision or commercial pricing expectations - SysML version baseline (v1.x vs. v2 transition timing) - Required modeling methods, profiles, or program-specific architecture constraints - Model naming/identifier conventions **Key consolidation effect:** Government reorganized responses to the most frequently asked technical specifications (software versions, GFI data, licensing, modeling standards) into a single upfront Q&A rather than addressing them piecemeal. This suggests the government prioritized clarity on infrastructure and baseline assumptions for all performers. **No new substantive answers added.** The updated Q&A appears to reorganize and repackage previous pending responses rather than release new government position statements. Questions 2–19 remain substantively unchanged from the previous iteration, indicating responses to core feasibility, accuracy metrics, human-in-the-loop clarification, and data availability remain pending or were deferred into the consolidated Q1 response.

**Changes to Q&A Section:** Added 7 new clarifying questions (Q1-Q7) covering: facility clearance timing, unclassified data availability for Phase I, automation vs. human modification criteria, unclassified Phase II path options, Cameo integration timing, and acceptable SysML delivery formats (XMI import vs. native files). Reorganized previous 10 questions into Q8-Q23, consolidating duplicates. Key clarifications now address: number of anticipated awards, evaluation metrics prioritization, synthetic data acceptability, feasibility demonstration methods, human-in-the-loop thresholds, AEGIS baseline/artifact emphasis, and reference model availability for ground truth comparison.

Status changed from Pre-Release to Open

# Q&A Changes Summary **Structure Change:** Q&A was reorganized and consolidated. Original Q1–Q3 were replaced with numbered Q1–Q11, with significant consolidation and duplication. **Key Additions:** - Q1: Specific deliverable format requirements (SysML version, Cameo version, model exchange format) - Q2: Government dataset availability for Phase I - Q3–Q4: Merged duplicate questions on model scope, accuracy metrics, automation level, government data, and Cameo/SysML versions - Q5: AI environment deployment (on-premises vs. approved external service) - Q6: Emphasis on hardest ingestion challenges (cable sheets, network config data) - Q7: SysML v1 vs. v2 transition timeline alignment with Feb 2026 OUSW guidance - Q8: Specific tool versions (CATIA Magic/Cameo Systems Modeler version, Teamwork Cloud usage) - Q9–Q11: Reposted original Q1–Q3 verbatim **Critical Clarifications Pending:** No answers provided to new or reposted questions; all remain unanswered. Original response-pending status appears unchanged. **Notable Pattern:** Significant question duplication suggests possible administrative consolidation effort without government responses yet issued.

Added new Q1 with 3 questions on AI model usage (commercial vs. Navy-approved), Phase II deployment environment (on-site vs. external Government service), and Navy test data provision. Retained all 5 questions from original Q2/Q3 covering Cameo integration architecture, SysML version targeting, synthetic data acceptability, model accuracy metrics, and human involvement clarification.

The original Q1 was expanded into 5 sub-questions clarifying: (1) flexibility in Cameo integration architecture, (2) required SysML/Cameo versions, (3) use of synthetic data for Phase I, (4) model validation metrics, and (5) reconciliation between "no human input" automation vs. post-generation modifications.

This Q&A clarifies that applicants need to confirm whether their conversion solution must operate natively within Cameo MagicDraw or can use external tools for model generation with subsequent import, and should identify specific tool/SysML version compatibility requirements to propose against.

New opportunity: NAVSEA Open Topic for Model-Based Systems Engineering (MBSE) Acceleration via Artificial Intelligence or Machine Learning Data Ingestion