Inference Engines
Specialized Reasoning for Digital Phenotype and Functional-State Intelligence
ProXplor develops specialized inference engines for distinct forms of health and medical intelligence.
InferPheno™ interprets digital phenotype information. InferFunc™ interprets functional-state information. They use different data structures, analytical methods and inference pathways while operating on the shared foundation of ProXplor Nexus™.
Together, they translate the scientific methodology of the ProXplor Research Framework™ into deployable intelligence for standardized products and customized enterprise solutions.
Why Specialized Inference Matters
Health intelligence is not a single computational problem.
Phenotype-related images, functional assessments, bioelectric signals, connected device data, behavioral information and longitudinal records represent different aspects of health. They cannot be interpreted responsibly through an identical analytical process.
Each data type requires:
appropriate acquisition and quality controls;
clearly defined variables;
domain-specific knowledge structures;
purpose-built inference logic;
evidence and confidence boundaries;
product-specific output rules;
appropriate governance and regulatory controls.
ProXplor therefore uses specialized inference engines rather than treating all health-related data as interchangeable inputs to a single general-purpose model.
Shared Foundation. Distinct Reasoning Architectures.
InferPheno and InferFunc operate within ProXplor Nexus, ProXplor’s Health and Medical Intelligence Foundation Model & Core Technology Platform.
ProXplor Nexus provides the shared environment for:
health and medical knowledge representation;
multimodal data structuring;
model and inference orchestration;
contextual and longitudinal state representation;
output governance;
explainability and traceability;
product integration;
enterprise deployment.
Each engine applies its own domain-specific processing and reasoning architecture within this shared environment.
This separation enables ProXplor to preserve methodological precision while maintaining consistency across its broader technology ecosystem.
InferPheno™
Digital Phenotype Inference Engine
InferPheno™ is ProXplor’s Digital Phenotype Inference Engine.
It structures and interprets phenotype-related images and authorized contextual information to generate appropriately scoped digital phenotype intelligence.
Within the ProXplor technology architecture, digital phenotype refers to observable characteristics captured through standardized digital inputs and interpreted in relation to biological, behavioral, environmental and temporal context.
InferPheno does not equate observable phenotype-related features with a diagnosis. It analyzes structured phenotype-related features, identifies relevant patterns and applies defined inference pathways according to the purpose and boundaries of each product.
Phenotype-Related Inputs
Depending on the product configuration, authorized inputs may include:
standardized frontal facial images;
standardized facial profile images;
standardized tongue images;
demographic and contextual information;
self-reported lifestyle information;
previous assessments and longitudinal records.
Input requirements, data retention policies and permitted uses are defined separately for each product and deployment.
InferPheno is designed for phenotype inference, not identity recognition or identity verification.
From Image to Digital Phenotype Intelligence
InferPheno applies a structured inference process.
Input Quality Assessment
The engine evaluates whether submitted images meet the technical and product-specific requirements necessary for analysis.
Phenotype Feature Structuring
Relevant observable features are converted into structured variables suitable for computational interpretation.
Contextualization
Phenotype-related variables are interpreted in relation to authorized contextual information rather than being treated as isolated visual characteristics.
Phenotype State Representation
The engine organizes relevant variables into a multidimensional representation of the observed phenotype state.
Inference
Defined knowledge structures, models and inference logic are applied to identify relationships and generate appropriately bounded interpretations.
Output Generation
Results are translated into product-specific reports, insights or decision-support outputs according to the intended use of the application.
InferPheno™ Outputs
Depending on the product and intended use, InferPheno can support:
structured representations of phenotype-related features;
interpretation of multidimensional phenotype patterns;
phenotype-informed health and lifestyle insights;
comparison across repeated assessments;
longitudinal tracking of relevant changes;
appropriately scoped recommendations or next steps;
integration with enterprise workflows and applications.
Within ProXplor products, source data, direct observations and model-generated inferences are maintained as distinguishable information layers.
Products and Applications
InferPheno powers the Phenotype Series™, ProXplor’s standardized digital phenotype intelligence product family.
Its capabilities can also be integrated into:
the Holmony App;
authorized professional applications;
enterprise health platforms;
institutional programs;
customized applications developed through ProXplor Nexus™ Solutions.
The specific inputs, outputs and claims available in each application depend on its intended use, configuration and applicable regulatory requirements.
InferFunc™
Functional-State Inference Engine
InferFunc™ is ProXplor’s Functional-State Inference Engine.
It structures and interprets functional assessment, bioelectric, connected device and contextual data to generate functional-state intelligence and support health navigation.
Functional state describes how relevant physiological and behavioral functions are operating and interacting at a given time. It is not defined by a single measurement. It is represented through relationships among multiple variables, systems and contextual factors.
InferFunc is designed to move from isolated measurements toward a structured understanding of functional organization, system relationships and change over time.
Functional-State Inputs
Depending on the product configuration, authorized inputs may include:
functional assessment data;
bioelectric signal acquisition data;
connected health device data;
relevant physiological measurements;
self-reported health and lifestyle information;
behavioral and environmental context;
previous assessments and longitudinal records.
The availability and interpretation of each input depend on the acquisition method, product configuration, intended use and applicable regulatory requirements.
From Functional Data to Health Navigation
InferFunc applies a distinct functional-state inference process.
Data Acquisition and Quality Control
Data are captured through product-specific acquisition systems and evaluated for technical quality and usability.
Variable Standardization
Relevant measurements and observations are converted into consistently defined functional variables.
Functional-State Representation
Variables are organized into structured representations of relevant functions, relationships and system-level patterns.
Contextual and Systemic Analysis
The engine evaluates functional information in relation to individual context, behavioral factors and other authorized data.
Functional-State Inference
Defined models, knowledge structures and inference logic are applied to generate appropriately scoped interpretations.
Navigation and Decision Support
Outputs are translated into health navigation, lifestyle-oriented guidance or clinical decision support according to the intended use and regulatory status of the product.
Reassessment
Repeated assessments can be used to compare states, identify relevant changes and refine longitudinal understanding.
InferFunc™ Outputs
Depending on the product and intended use, InferFunc can support:
structured functional-state representations;
functional-state indices;
interpretation of system-level relationships;
identification of relevant patterns and trends;
longitudinal comparison and reassessment;
general wellness and health optimization guidance;
clinical decision support within appropriately validated and authorized products;
integration with professional and institutional workflows.
InferFunc outputs are governed by the intended use of the product through which they are delivered.
Products and Applications
InferFunc powers the HNS Series™, ProXplor’s health navigation product family.
AIHNS™
AIHNS™—Artificial Intelligence Health Navigation System—is designed for general wellness and health optimization.
It translates functional-state intelligence into non-diagnostic health navigation and lifestyle-oriented guidance.
CIHNS™
CIHNS™—Clinical Intelligence Health Navigation System—is designed for clinical use.
Its deployment is subject to product-specific validation and applicable regulatory requirements in each jurisdiction.
InferFunc capabilities can also be integrated into authorized enterprise platforms, professional workflows and customized applications developed through ProXplor Nexus™ Solutions.
Two Engines with Different Responsibilities
InferPheno and InferFunc are complementary, but they are not interchangeable.
Different Primary Inputs
InferPheno primarily interprets standardized phenotype-related images and contextual information.
InferFunc primarily interprets functional assessment, bioelectric, connected device and contextual data.
Different Representations
InferPheno constructs digital phenotype representations.
InferFunc constructs functional-state representations.
Different Inference Logic
Each engine uses knowledge structures, variables and inference pathways appropriate to its analytical domain.
Different Product Families
InferPheno powers the Phenotype Series™.
InferFunc powers the HNS Series™, including AIHNS™ and CIHNS™.
Shared Technology Foundation
Both engines operate on ProXplor Nexus and follow the scientific and methodological principles of the ProXplor Research Framework.
How the Engines Work Across Applications
InferPheno and InferFunc can be deployed independently or incorporated into broader applications through ProXplor Nexus.
The Holmony App serves as a client-facing access layer and can present eligible reports, insights and services supported by InferPheno, InferFunc or other ProXplor Nexus capabilities according to product configuration and user authorization.
Enterprise and institutional applications can use one engine, both engines or additional Nexus capabilities according to their objectives, available data, workflows and regulatory requirements.
When multiple capabilities are used, their outputs remain methodologically distinguishable. Information from different engines is not combined without defined integration logic, appropriate evidence and product-specific governance.
Engines Are Technology, Not End-User Products
InferPheno and InferFunc are core inference technologies.
They generate structured intelligence but do not independently define the complete product experience.
The products and applications built on them determine:
the intended user;
the intended use;
the permitted data inputs;
the report or interface;
the scope of recommendations;
the required professional oversight;
the applicable validation pathway;
the regulatory classification;
the privacy and security controls;
the deployment environment.
This separation allows ProXplor to apply the same core engine capabilities across different products while preserving clear functional, commercial and regulatory boundaries.
Responsible Inference by Design
ProXplor’s inference engines are developed around principles of responsible health and medical intelligence.
These principles include:
purpose-specific data use;
separation of observation and inference;
traceable analytical pathways;
explicit confidence and uncertainty boundaries;
product-specific output controls;
protection of personal and health-related data;
human oversight where professional judgment is required;
model and knowledge-base version control;
ongoing evaluation and refinement;
compliance with applicable regulatory requirements.
The objective is not to generate conclusions beyond what the available data and intended use can support. It is to transform complex information into structured, explainable and appropriately bounded intelligence.
Two Specialized Engines. One Intelligence Foundation.
InferPheno and InferFunc represent two distinct pathways for understanding health-related information.
One begins with digital phenotype. The other begins with functional state.
Both are grounded in systems science, supported by ProXplor Nexus and translated into products designed for clearly defined users and purposes.