Datasets:

ClarusC64
/

clinical-five-node-sepsis-cascade-boundary-v1.0

scenario_id float64	infection_load float64	immune_strain float64	perfusion_deficit float64	organ_drift float64	buffer_capacity float64	drift_gradient float64	drift_velocity float64	drift_acceleration float64	boundary_distance float64	secondary_boundary_distance float64	boundary_competition_ratio float64	boundary_uncertainty float64	trajectory_uncertainty float64	regime_confidence float64	regime_transition_score string	transition_direction float64	regime_separation_margin float64	transition_uncertainty float64	transition_velocity float64	intervention_leverage_score float64	intervention_alignment_score float64	rescue_window_width float64	pathway_divergence_margin float64	intervention_competition_ratio string	primary_intervention_path string	secondary_intervention_path float64	intervention_uncertainty float64	pathway_switch_velocity float64	control_sequence_alignment_score int64	control_horizon float64	feedback_response_score float64	intervention_timing_score int64	adaptation_latency float64	control_stability_margin float64	sequence_divergence_margin float64	controller_confidence float64	recovery_consistency_score int64	control_recalibration_count string	terminal_pathway_state float64	feedback_noise_ratio float64	controller_oscillation_score int64	rollback_trigger_count float64	perturbation_radius string	collapse_trigger float64	recovery_distance float64	recovery_gradient float64	return_feasibility float64	delta_infection_load float64	delta_immune_strain float64	delta_perfusion_deficit float64	delta_organ_drift float64	delta_buffer_capacity float64	trajectory_shift float64	minimal_intervention_path string	stabilization_success int64	label_sepsis_boundary int64	__index_level_0__ string
0.88	0.79	0.74	0.7	0.27	0.71	0.64	0.24	0.16	0.33	0.74	0.18	0.24	0.82	0.81	toward_failure	0.41	0.21	0.57	0.84	0.79	0.39	0.25	0.64	source_control_then_antibiotics_then_fluids	pressor_then_imaging_then_support	0.19	0.16	0.81	3	0.77	0.8	1	0.73	0.23	0.82	0.78	1	stabilized	0.11	0.1	0	0.44	bacteremic_escalation	0.3	-0.4	0.76	-0.18	-0.16	-0.15	-0.13	-0.14	0.22	-0.23	source_control_then_antibiotics_then_fluids	1	1	sepc001
0.93	0.85	0.81	0.78	0.21	0.79	0.72	0.29	0.11	0.27	0.82	0.24	0.29	0.78	0.88	toward_failure	0.35	0.25	0.64	0.8	0.68	0.29	0.19	0.68	pressor_then_imaging_then_support	source_control_then_antibiotics_then_fluids	0.28	0.24	0.55	1	0.5	0.54	3	0.35	0.19	0.65	0.43	3	unstable_recovery	0.17	0.3	1	0.58	septic_shock_acceleration	0.46	-0.17	0.4	-0.08	-0.05	-0.04	-0.06	-0.05	0.04	-0.05	pressor_then_imaging_then_support	0	0	sepc002
0.8	0.71	0.66	0.61	0.36	0.53	0.47	0.14	0.27	0.42	0.58	0.15	0.18	0.85	0.67	toward_failure	0.48	0.17	0.41	0.75	0.72	0.53	0.17	0.57	antibiotics_then_fluids_then_monitoring	source_control_then_antibiotics_then_fluids	0.14	0.1	0.75	2	0.76	0.77	1	0.7	0.16	0.79	0.74	1	partially_stabilized	0.08	0.12	0	0.34	occult_source_persistence	0.27	-0.3	0.71	-0.13	-0.11	-0.1	-0.09	-0.08	0.15	-0.14	antibiotics_then_fluids_then_monitoring	1	1	sepc003
0.96	0.89	0.86	0.84	0.17	0.85	0.77	0.31	0.08	0.23	0.87	0.28	0.33	0.75	0.93	toward_failure	0.29	0.27	0.71	0.77	0.58	0.21	0.15	0.73	mechanical_support_then_relook_then_rescue	pressor_then_imaging_then_support	0.34	0.3	0.44	0	0.39	0.41	4	0.24	0.14	0.58	0.31	4	relapse	0.22	0.38	2	0.66	refractory_septic_boundary	0.55	-0.1	0.25	-0.03	-0.02	-0.02	-0.02	-0.03	0.01	-0.03	mechanical_support_then_relook_then_rescue	0	0	sepc004
0.84	0.75	0.69	0.65	0.32	0.6	0.53	0.16	0.22	0.39	0.63	0.16	0.21	0.84	0.74	toward_failure	0.45	0.18	0.47	0.81	0.76	0.46	0.22	0.61	source_control_then_antibiotics_then_fluids	antibiotics_then_fluids_then_monitoring	0.17	0.13	0.78	2	0.8	0.79	1	0.73	0.2	0.81	0.78	1	stabilized	0.1	0.09	0	0.4	hemodynamic_sepsis_breakpoint	0.32	-0.34	0.75	-0.15	-0.13	-0.12	-0.11	-0.1	0.18	-0.17	source_control_then_antibiotics_then_fluids	1	1	sepc005
0.89	0.81	0.77	0.73	0.26	0.72	0.65	0.24	0.15	0.3	0.75	0.2	0.25	0.81	0.82	toward_failure	0.38	0.22	0.58	0.82	0.65	0.32	0.19	0.65	fluids_then_pressor_support_then_source_control	source_control_then_antibiotics_then_fluids	0.24	0.2	0.62	1	0.58	0.61	2	0.49	0.19	0.69	0.59	2	unstable_recovery	0.15	0.26	1	0.49	delayed_source_control	0.41	-0.21	0.46	-0.07	-0.06	-0.05	-0.05	-0.04	0.06	-0.07	fluids_then_pressor_support_then_source_control	1	0	sepc006
0.76	0.63	0.57	0.52	0.4	0.44	0.38	0.1	0.33	0.48	0.52	0.12	0.16	0.87	0.58	toward_failure	0.52	0.14	0.34	0.71	0.68	0.58	0.16	0.55	monitoring_then_antibiotics_then_fluids	antibiotics_then_fluids_then_monitoring	0.13	0.09	0.72	2	0.73	0.75	1	0.68	0.15	0.77	0.71	1	partially_stabilized	0.08	0.12	0	0.3	transient_inflammatory_spike	0.26	-0.26	0.69	-0.12	-0.09	-0.08	-0.08	-0.07	0.14	-0.13	monitoring_then_antibiotics_then_fluids	1	1	sepc007
0.92	0.84	0.8	0.79	0.22	0.8	0.73	0.28	0.1	0.26	0.83	0.25	0.29	0.77	0.89	toward_failure	0.33	0.25	0.65	0.77	0.54	0.26	0.16	0.7	pressor_then_imaging_then_support	mechanical_support_then_relook_then_rescue	0.31	0.28	0.45	0	0.4	0.4	4	0.23	0.14	0.55	0.3	4	irreversible_collapse	0.23	0.41	2	0.67	refractory_lactate_spiral	0.6	-0.08	0.2	-0.02	-0.01	-0.01	-0.01	-0.02	0.01	-0.03	pressor_then_imaging_then_support	0	0	sepc008
0.81	0.72	0.66	0.61	0.35	0.55	0.49	0.15	0.24	0.41	0.59	0.15	0.19	0.85	0.69	toward_failure	0.47	0.17	0.44	0.8	0.74	0.49	0.2	0.59	source_control_then_antibiotics_then_fluids	fluids_then_pressor_support_then_source_control	0.17	0.12	0.76	3	0.77	0.78	1	0.71	0.18	0.8	0.75	1	stabilized	0.09	0.1	0	0.37	mixed_sepsis_shift	0.29	-0.32	0.74	-0.14	-0.12	-0.11	-0.1	-0.09	0.17	-0.15	source_control_then_antibiotics_then_fluids	1	1	sepc009

ClarusC64/clinical-five-node-sepsis-cascade-boundary-v1.0

What this repo does

This repository provides a Clarus v1.0 benchmark for sepsis cascade boundary dynamics using a five-node clinical cascade:

infection_load
immune_strain
perfusion_deficit
organ_drift
buffer_capacity

The v1.0 upgrade is Closed-Loop Control Geometry.

The task is no longer limited to detecting deterioration, forecasting cascade spread, or ranking one intervention against another.

It tests whether a controller can:

choose the right cascade rescue path
apply that path in the right sequence
read feedback from the system
adapt in time
maintain durable recovery across the cascade

Concept ladder

Version	Capability
v0.1	cascade detection
v0.2	trajectory awareness
v0.3	cascade forecasting
v0.4	boundary discovery
v0.5	recovery geometry
v0.6	intervention reasoning
v0.7	uncertainty-aware intervention
v0.8	regime transition geometry
v0.9	intervention competition geometry
v1.0	closed-loop control geometry

Core five-node cascade

infection_load

Primary infectious burden initiating the cascade.

immune_strain

Inflammatory and immune activation burden.

perfusion_deficit

Circulatory under-delivery and tissue oxygen deficit.

organ_drift

Downstream organ instability and movement toward failure.

buffer_capacity

Remaining reserve available to absorb cascade stress.

Clinical variable mapping

Variable	Clinical interpretation	Typical measurement proxies
infection_load	infectious burden	source severity, bacteremia load, microbial persistence
immune_strain	inflammatory stress	CRP trend, cytokine surge, white-cell dysregulation
perfusion_deficit	circulatory compromise	lactate, MAP instability, capillary refill deficit
organ_drift	downstream organ destabilization	creatinine drift, bilirubin drift, ventilation burden
buffer_capacity	remaining reserve	perfusion reserve, organ tolerance, metabolic reserve

These five variables define the structural state of the septic cascade rather than isolated bedside readings.

Prediction target

The target is label_sepsis_boundary.

Default stronger v1.0 rule

label = 1 if all of the following hold:

stabilization_success = 1
trajectory_shift < -0.10
intervention_alignment_score >= 0.60
control_sequence_alignment_score >= 0.60
recovery_consistency_score >= 0.60

Mid-strength variant

label = 1 if:

stabilization_success = 1
trajectory_shift < -0.10
control_sequence_alignment_score >= 0.60

Relaxed variant

label = 1 if stabilization_success = 1

Example row

The following simplified row shows a near-miss under the strict v1.0 label rule.

Field	Value
stabilization_success	1
trajectory_shift	-0.11
intervention_alignment_score	0.64
control_sequence_alignment_score	0.59
recovery_consistency_score	0.74

Four of the five conditions are satisfied.

But control_sequence_alignment_score = 0.59 falls below the threshold of 0.60.

Therefore:

label = 0

This shows why v1.0 rewards true closed-loop cascade control, not temporary improvement alone.

Row structure

Each row represents a sepsis cascade boundary scenario with:

five cascade nodes
trajectory signals
boundary geometry
regime transition signals
intervention competition signals
closed-loop control signals
perturbation and recovery signals
delta signals
intervention path and final label

Signal groups

Five-node cascade variables

infection_load
immune_strain
perfusion_deficit
organ_drift
buffer_capacity

Trajectory signals

drift_gradient
drift_velocity
drift_acceleration
trajectory_shift

Boundary geometry

boundary_distance
secondary_boundary_distance
boundary_competition_ratio

Uncertainty signals

boundary_uncertainty
trajectory_uncertainty
regime_confidence
transition_uncertainty
intervention_uncertainty
controller_confidence
feedback_noise_ratio

Regime transition signals

regime_transition_score
transition_direction
regime_separation_margin
transition_velocity

Intervention signals

intervention_leverage_score
intervention_alignment_score
rescue_window_width
pathway_divergence_margin
intervention_competition_ratio
primary_intervention_path
secondary_intervention_path
pathway_switch_velocity
minimal_intervention_path

Closed-loop control signals (v1.0)

control_sequence_alignment_score
control_horizon
feedback_response_score
intervention_timing_score
adaptation_latency
control_stability_margin
sequence_divergence_margin
controller_confidence
recovery_consistency_score
control_recalibration_count
terminal_pathway_state

Possible terminal pathway states include:

stabilized
partially_stabilized
unstable_recovery
relapse
irreversible_collapse

Optional control diagnostics included

feedback_noise_ratio
controller_oscillation_score
rollback_trigger_count

Recovery signals

recovery_distance
recovery_gradient
return_feasibility

Perturbation signals

perturbation_radius
collapse_trigger

Delta signals

delta_infection_load
delta_immune_strain
delta_perfusion_deficit
delta_organ_drift
delta_buffer_capacity

Dataset construction

Each scenario is generated using a structured simulation of five-node sepsis cascade deterioration and intervention sequences.

1. Cascade initialization

A baseline septic cascade state is sampled across the five nodes.

2. Instability evolution

The cascade evolves using trajectory signals that determine movement toward collapse or recovery boundaries.

3. Intervention competition

Candidate rescue sequences are evaluated using intervention competition geometry.

4. Closed-loop control execution

A selected rescue path is applied through a sequence of actions.

Control signals measure:

sequence alignment
timing
feedback interpretation
adaptation speed
durability of recovery across the cascade

The final state determines:

stabilization_success
terminal_pathway_state
label_sepsis_boundary

Files

data/train.csv
Labeled training set with the full v1.0 schema.
data/tester.csv
Test-style file. stabilization_success is withheld.
scorer.py
Reference scorer for binary metrics and v1.0 control diagnostics.
benchmark_spec.json
Canonical machine-readable benchmark spec.
dataset_schema.json
Machine-readable structural schema with column groups, types, ranges, and row order.

Evaluation

Primary metric

recall_correct_control_sequence_selection

Secondary metric

false_effective_control_rate

Binary metrics

accuracy
precision
recall
f1
confusion matrix

Closed-loop diagnostics

primary_intervention_path_accuracy
secondary_intervention_path_accuracy
control_sequence_alignment_accuracy
control_horizon_error
feedback_response_accuracy
intervention_timing_accuracy
high_uncertainty_control_miss_rate
narrow_window_control_miss_rate
adaptation_latency_error
control_stability_error
recovery_consistency_error
recalibration_overuse_rate
controller_oscillation_misread_rate
terminal_pathway_state_accuracy

Structural interpretation

Earlier Clarus datasets asked:

Which rescue path is best?

v1.0 asks a harder question:

Can the controller stay aligned with reality while a five-node cascade evolves?

Real systems fail not only because the first action is wrong.

They also fail because:

feedback is misread
adaptation is delayed
interventions are mistimed
control oscillations destabilize recovery

v1.0 measures these failure modes directly.

Dataset limitations

This dataset models structural control dynamics, not detailed clinical treatment protocols.

Important limitations:

intervention paths are simplified abstractions
control signals represent structural decision quality, not pharmacological precision
physiological variables are normalized system indicators rather than raw bedside measurements
the dataset does not capture the full biological variability of real sepsis cascades

The benchmark evaluates control reasoning, not medical safety.

Intended use

This dataset is intended for research on:

instability prediction
sequential decision reasoning
closed-loop cascade control modeling
intervention planning under uncertainty
AI robustness in dynamic clinical-like environments

Not intended for

This dataset must not be used for:

real clinical decision support
medical diagnosis
treatment recommendation systems
automated ICU control systems
deployment in patient care environments

Structural note

This v1.0 dataset marks the move from intervention competition to actual control logic across a five-node cascade.

The benchmark asks whether the controller stays aligned with reality across time.

That is the threshold where Clarus becomes a control-layer instrument rather than only a detection or ranking layer.

Production deployment

This dataset format is suitable for controlled benchmarking in domains where sequential intervention quality matters more than one-shot classification.

Examples include:

sepsis cascade stabilization
ICU escalation pathway planning
multistage rescue benchmarking
distributed system control
multi-step recovery planning

Enterprise and research collaboration

This repo is part of the broader Clarus ladder for modeling instability, recovery, and control under feedback.

It is designed for:

benchmark development
model evaluation
intervention policy testing
control-sequence auditing
future cross-domain transfer into other high-stakes systems

License

MIT

Downloads last month: 22

Number of rows:

Total file size:

49.9 kB