Identifying and analyzing changes in an environment (e.g., production processes, physical systems, components, battlefield, etc.) are essential to predicting their outcomes and addressing them quickly. AI/ML models can integrate sensor inputs from infrastructure and technology, to track changes over time, detect potential weaknesses or issues, and dynamically adjust operations.

Cycle 1 Ideathon Requested Datasets

Cybersecurity and Anomaly Detection

• High-fidelity simulation data generated with domain subject matter experts (SMEs).
• Data related to cybersecurity, IoT, and sensor-based datasets.
• Datasets that address anomalies, including rationale for labeling to understand data principles.

Energy and the Grid

• Weather Data: Needed for seasonal forecasting in energy management.
• Emissions Signatures: Data on vehicle emissions for preventative health monitoring.
• Degradation Data: Information on the degradation of fabricated solar cells exposed to UV radiation.

Manufacturing and Supply Chain Data

• Corrosion Data: Physical inspection data, including thermal inspection for assessing corrosion impact.
• Composite Materials Data: Data from sensors used to detect subsurface delamination in heated areas.
• Device Fabrication Data: Information on the processes and materials used in device manufacturing.
• Supply Chain Data: Data concerning materials used in supply chains.

Data Management and Automation

• Automation of Labeling and Annotation: Interest in automating the time-consuming process of labeling datasets for SMEs.

Human-AI Teaming

• Datasets on how humans interact with AI tools.
• Datasets on human decision making with AI tools.
• Datasets on how AI interacts with and learns from human users.

Physical security data

Transforming diverse data from design, production, and operations into insights is essential in supporting better decisions and more adaptive systems across a manufacturing lifecycle. AI/ML models can connect data streams, identify patterns, extract signals, and automate analyses across complex, multimodal data workflows.

Cycle 2 Ideathon Requested Datasets

Sensor & Time Series Data

• Industrial and manufacturing process monitoring: Manufacturing/SCADA time-series, automated testing logs, and experiment data for tracking operations, quality, and anomalies.
• Critical infrastructure and structural health monitoring: Sensor streams from bridges, pipelines, and other infrastructure, including strain, pressure, temperature, and load-response data.
• Energy and nuclear system monitoring: Nuclear cooling-system strain/temperature cycles and pipeline condition data for detecting leaks, corrosion, fatigue, or abnormal behavior.
• General environmental and sensor-stream analytics: Vibration, flow, humidity, agriculture-environment data, and other continuous sensor streams used for monitoring, prediction, and diagnostics.

Imaging & Visual Data

• Quality-control photographs (boards, solder bonds).
• X-ray inspection images.
• Electron-microscopy (TEM/SEM) microstructure image sets.
• In-situ AM machine images/time series.

Materials Testing & Qualification Data

• Non-destructive evaluation (NDE) measurements.
• Long-term creep and fatigue test records.
• Irradiation-behavior data for nuclear materials.
• Corrosion performance datasets in varied environments.

Simulation & Synthetic Data

• Physics-based synthetic surrogate datasets (modeled leaks, noise).
• Digital-twin simulation inputs (scaled-facility sensor streams).

Documentation & Knowledge Management

• Legacy scanned R&D, operational logs, regulatory documents.
• Fuels & materials properties handbooks (e.g. INL GIV materials properties handbook).
• Industry-specific standards (e.g. NQA-1).
• Domain ontologies and knowledge-graph definitions for specialized expertise.

Operations & Supply-Chain Data

• Supply-chain and raw-material traceability records.
• Power-distribution and microgrid resilience data.

Cybersecurity & Anomaly Detection Data for AM

• Cybersecurity logs, anomaly-detection datasets.
• Adversarial-ML perturbation datasets.

Model Validation & Benchmarking Data

• AI/ML benchmarking test results.
• Model audit-trail and validation records.

Data Formats, Ingestion Tools, & Best Practices

• Master data catalog or directory metadata (dataset/SME index).
• AI-ready formatted datasets (e.g. JSON, structured tables).
• Helper-function code snippets for data ingest.
• Agentic AI with tool-calling to process scanned data into useful formats.