ClarusC64/clinical-quad-ventilation-stress-buffer-lag-coupling-ventilator-deterioration-v0.6

What this repo does

This repository contains a Clarus v0.6 intervention pathway dataset focused on ventilator deterioration dynamics.

The dataset evaluates whether a model can determine if a proposed intervention meaningfully stabilizes a deteriorating ventilated system.

The task requires reasoning from:

• system state
• trajectory toward instability
• boundary geometry
• recovery geometry
• intervention vector
• projected trajectory consequence

The model cannot read the answer directly.

It must infer stabilization from the structure of the case.

This shifts the benchmark from simple ventilator deterioration detection to intervention reasoning.

Core quad

The ventilator deterioration system is represented using four normalized variables.

• ventilation_stress
• physiological_buffer
• intervention_lag
• systemic_coupling

These variables capture the core structural drivers of ventilatory cascade progression.

Clinical variable mapping

The normalized quad variables correspond to measurable clinical signals.

Quad Variable	Clinical Measurements	Typical Indicators
ventilation_stress	Peak airway pressure Plateau pressure Hypoxemia burden Rising CO2 load	Pplat rising PaO2/FiO2 falling Hypercapnia worsening
physiological_buffer	Pulmonary reserve Hemodynamic tolerance Acid-base reserve Baseline organ resilience	Low lung compliance Shock vulnerability Poor reserve
intervention_lag	Delay to ventilator adjustment Delay to suctioning or secretion control Delay to proning Delay to sedation or paralysis optimization	Late vent changes Delayed proning
systemic_coupling	Lung-hemodynamic interaction Ventilation-perfusion mismatch propagation Inflammatory spillover Multi-organ coupling under respiratory strain	Hypoxia with hypotension ARDS plus circulatory stress

These measurements illustrate how normalized values in the dataset map to real ventilator physiology.

Prediction target

The target column is:

label_ventilator_stabilization

This label indicates whether the intervention pathway produces genuine stabilization.

Label logic

Default benchmark rule

A row is labeled positive only when both conditions hold:

stabilization_success = 1

and

trajectory_shift < -0.10

This rule filters out marginal corrections and ensures that positive examples represent meaningful stabilization.

Optional relaxed rule

Positive labels may trigger when:

stabilization_success = 1

This relaxed rule may be used for exploratory builds but is not the default benchmark configuration.

Row structure

Each row contains:

• core ventilatory state
• trajectory geometry
• perturbation geometry
• recovery geometry
• intervention vector
• projected trajectory consequence

Train rows include:

• stabilization_success
• label_ventilator_stabilization

Tester rows exclude these fields.

Why tester rows exclude stabilization_success

The tester file withholds:

• stabilization_success
• label_ventilator_stabilization

This prevents answer leakage.

The model must infer stabilization using:

• the starting ventilatory state
• drift toward the failure boundary
• recovery basin geometry
• the intervention vector
• the predicted trajectory consequence

This structure forces real intervention reasoning.

Minimal intervention path

minimal_intervention_path encodes the shortest viable stabilization pathway.

Example interpretation:

• 0 = no viable rescue path
• 1 = direct stabilization pathway
• 2 = multi-step stabilization sequence
• 3 = complex high-risk rescue pathway

The field remains visible because the benchmark evaluates whether the model can combine intervention structure with trajectory consequence to determine stabilization.

Files

• data/train.csv — labeled training dataset
• data/tester.csv — unlabeled benchmark dataset with withheld stabilization signal
• scorer.py — evaluation metrics and confusion matrix computation
• cli.py — command-line evaluation wrapper used for benchmark scoring
• README.md — dataset card and schema documentation

Evaluation

The scorer reports:

• accuracy
• precision
• recall_successful_stabilization
• failed_rescue_rate
• f1
• confusion_matrix

Primary metric:

recall_successful_stabilization

Secondary metric:

failed_rescue_rate

Interpretation:

recall_successful_stabilization measures how reliably the model detects interventions that genuinely stabilize the ventilated system.

failed_rescue_rate measures how often the model fails to recognize a viable stabilization pathway.

These metrics prioritize intervention reasoning rather than generic classification accuracy.

Schema

train.csv columns

• scenario_id
• ventilation_stress
• physiological_buffer
• intervention_lag
• systemic_coupling
• drift_gradient
• drift_velocity
• drift_acceleration
• boundary_distance
• perturbation_radius
• collapse_trigger
• recovery_distance
• recovery_gradient
• return_feasibility
• delta_ventilation_stress
• delta_physiological_buffer
• delta_intervention_lag
• delta_systemic_coupling
• trajectory_shift
• minimal_intervention_path
• stabilization_success
• label_ventilator_stabilization

tester.csv columns

• scenario_id
• ventilation_stress
• physiological_buffer
• intervention_lag
• systemic_coupling
• drift_gradient
• drift_velocity
• drift_acceleration
• boundary_distance
• perturbation_radius
• collapse_trigger
• recovery_distance
• recovery_gradient
• return_feasibility
• delta_ventilation_stress
• delta_physiological_buffer
• delta_intervention_lag
• delta_systemic_coupling
• trajectory_shift
• minimal_intervention_path

Structural note

The Clarus dataset series evolves through progressively richer representations of cascade dynamics.

Version progression:

• v0.1 — cascade state detection
• v0.2 — trajectory-aware detection
• v0.3 — dynamic cascade forecasting
• v0.4 — boundary discovery
• v0.5 — recovery geometry
• v0.6 — intervention pathway reasoning

Earlier versions identify when instability is developing.

Version 0.6 evaluates whether a proposed intervention meaningfully alters the trajectory of a system approaching collapse.

This marks the transition from monitoring cascade dynamics to evaluating control pathways.

Production deployment

This dataset structure can support clinical decision environments where ventilator deterioration must be detected and corrected before irreversible transition occurs.

Example settings include:

• ICU ventilator surveillance
• ARDS escalation monitoring
• proning and vent-adjustment decision support
• secretion burden escalation review
• respiratory failure step-up monitoring

Enterprise and research collaboration

This dataset class supports benchmarking for:

• intervention-aware clinical AI
• ventilatory cascade modeling
• recovery feasibility prediction
• false-stability detection
• boundary-sensitive decision support systems

Contact

For dataset expansion, custom coherence scorers, or deployment architecture:

team@clarusinvariant.com

Instability is detectable. Governance determines whether it propagates.

License

MIT