AI Methods & Sub-Methods

AI Methods

AI Method	AI Method Definition	AI Method Sub-Categories
Integrated Sensing	Merges data from different hardware (e.g., Radar + Camera + LiDAR) to create an integrated data source base for other AI methods.	Raw Data Fusion (Low-level merging) Feature-Level Fusion (Object-level merging) Temporal Fusion (Time-based tracking) Cross-Datal Validation (Sensor cross-checking)
Perception & Recognition	Ability of AI to translate raw sensor data (video, LiDAR, photos) into meaningful information. Identifies what and where things are	Image Labeling (Image Classification) Locator Detection (Object Detection & Localization) Pixel-Level Mapping (Semantic Segmentation) Attribute Identification (Feature Extraction)
Predictive Analytics	Uses historical and real-time data to forecast future states. Estimates when or where something will happen	Time Series Analysis (Trend forecasting) Regression Modeling (Relationship estimation) Category Prediction (Classification) Similarity Grouping (Clustering)
Optimization & Control	Focuses on finding the "best" solution among millions of possibilities Determines the most efficient way to run a system.	Targeted Logic (Heuristics/Metaheuristics) Reinforcement Learning Scenario Analysis (Stochastic Optimization) Collective Logic (Swarm Intelligence)
Language & Knowledge Intelligence	Deals with unstructured human data (reports, emails, manuals). Understands, summarizes, or generates human-readable content.	Data Translation (Sequence-to-Sequence Modeling) Information Extraction (Named Entity Recognition) Concept-Based Search (Vector Embeddings) Sentence Structure Analysis (Semantic Parsing)
Automated Decision Support	Logic layer that takes the outputs from the other AI methods and suggests (or takes) an action. Decides what to do based on the data	Autonomous Agents (Agentic AI) Rule-Based/Expert Systems Benefit-Cost Modeling (Utility/Decision Analytics) Risk Based Logic Uncertainty Modeling (Probabilistic Reasoning)

AI Sub-Methods

AI Method Category	AI Sub-Method Type	AI Sub-Method Definition
Integrated Sensing	Raw Data Fusion (Low-level merging)	Combining unprocessed sensor signals (e.g., pixels and LiDAR points) into a single unified dataset before any individual analysis occurs.
	Feature-Level Fusion (Object-level merging)	Identifying distinct characteristics or objects in separate sensors and then mathematically merging them to form a single, high-confidence "truth."
	Temporal Fusion (Time-based tracking)	Integrating data points collected over a sequence of time to establish continuity, monitor movement patterns, and predict future positions.
	Cross-Data Validation (Sensor cross-checking)	Using independent data sources or different sensor modalities (e.g., Radar vs. Camera) to verify findings and eliminate "false positives" or sensor errors.
Perception & Recognition	Image Labeling (Image Classification)	Process of assigning a single definitive category or "tag" to an entire image or data frame based on its overall content.
	Locator Detection (Object Detection & Localization)	Identifying specific coordinates of multiple individual entities within a frame and drawing boundaries (bounding boxes) around each.
	Pixel-Level Mapping (Semantic Segmentation)	Categorizing every individual pixel in an image to define exact boundaries and shapes of various surfaces or objects.
	Attribute Identification (Feature Extraction)	Isolating and identifying specific detailed characteristics or patterns within an object, such as text, color, or structural anomalies.
Predictive Analytics	Time Series Analysis (Trend forecasting)	Analyzing data points collected at consistent time intervals to identify seasonal patterns, cycles, and long-term trends for future projection.
	Regression Modeling (Relationship estimation)	Mathematically estimating the strength and nature of the relationship between a dependent variable and one or more independent factors.
	Category Prediction (Classification)	Predicting which discrete bucket or group a new data point belongs based on historical training data (e.g., High-Risk vs. Low-Risk).
	Similarity Grouping (Clustering)	Using algorithms to discover natural groupings or patterns in data without pre-defined labels or categories.
Optimization & Control	Targeted Logic (Heuristics/Metaheuristics)	Utilizing advanced "rules of thumb" to efficiently navigate massive search spaces and find high-quality solutions without checking every single possibility.
	Reinforcement Learning	Training an AI agent to achieve a goal through a trial-and-error process where "correct" decisions are reinforced with mathematical rewards.
	Scenario Analysis (Stochastic Optimization)	Finding the most robust solution by accounting for random variables and uncertainty across multiple "what-if" potential future states.
	Collective Logic (Swarm Intelligence)	Decentralized problem-solving where a group of simple independent agents interact to find an optimal global solution.
Language & Knowledge	Data Translation (Sequence-to-Sequence)	Converting one sequence of structured or unstructured data into another (e.g., Speech-to-Text or Language-to-Language).
	Information Extraction (Named Entity Recognition)	Automatically identifying and pulling out specific, structured data points (names, dates, IDs) from unstructured text documents.
	Concept-Based Search (Vector Embeddings)	Converting text into mathematical coordinates (vectors) so the system can find information based on conceptual meaning rather than exact word matches.
	Sentence Structure Analysis (Semantic Parsing)	Breaking down the grammatical and logical structure of a sentence to determine the precise relationship between participants and actions.
Automated Decision Support	Autonomous Agents (Agentic AI)	AI systems capable of independent reasoning, using tools, and performing multi-step tasks to achieve a high-level objective without constant human prompting.
	Rule-Based/Expert Systems	A logic framework that applies a pre-defined library of human-coded "If-Then" rules and agency policies to data to reach a conclusion.
	Benefit-Cost Modeling (Utility/Decision Analytics)	Mathematically weighing the expected outcomes of various choices against their costs to determine the most "valuable" path forward.
	Risk-Based Logic	Decision-making frameworks that prioritize actions based on the probability and potential severity of negative impacts.
	Uncertainty Modeling (Probabilistic Reasoning)	Using probability math to reach decisions when data is incomplete, "noisy," or conflicting, providing a confidence score for each option.