Datasets:

ClarusC64
/

clinical-quad-infection-buffer-lag-coupling-sepsis-transition-v1.2

scenario_id string	infection_load float64	buffer_capacity float64	lag_burden float64	coupling_stress 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 float64	transition_direction string	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 float64	primary_intervention_path string	secondary_intervention_path string	intervention_uncertainty float64	pathway_switch_velocity float64	control_sequence_alignment_score float64	control_horizon int64	feedback_response_score float64	intervention_timing_score float64	adaptation_latency int64	control_stability_margin float64	sequence_divergence_margin float64	controller_confidence float64	recovery_consistency_score float64	control_recalibration_count int64	terminal_pathway_state string	feedback_noise_ratio float64	controller_oscillation_score float64	rollback_trigger_count int64	perturbation_radius float64	collapse_trigger string	recovery_distance float64	recovery_gradient float64	return_feasibility float64	policy_drift_sensitivity float64	adaptive_policy_margin float64	stabilization_half_life float64	policy_fragility_score float64	policy_reselection_trigger int64	policy_decay_rate float64	temporal_control_consistency float64	reselection_delay_cost float64	delta_infection_load float64	delta_buffer_capacity float64	delta_lag_burden float64	delta_coupling_stress float64	trajectory_shift float64	minimal_intervention_path string	stabilization_success int64	label_sepsis_transition_temporal_policy int64
seps_t001	0.84	0.31	0.67	0.71	0.56	0.75	0.47	0.21	0.32	0.7	0.19	0.22	0.76	0.65	toward_failure	0.42	0.24	0.61	0.83	0.8	0.39	0.34	0.65	antibiotics_then_fluids	source_control_then_antibiotics	0.21	0.39	0.78	4	0.74	0.72	2	0.7	0.4	0.75	0.7	1	stabilized	0.15	0.19	0	0.49	early_sepsis_burden	0.38	-0.24	0.56	0.27	0.35	0.8	0.25	0	0.21	0.77	0.05	-0.08	0.06	0.03	-0.02	-0.19	antibiotics_then_fluids	1	1
seps_t002	0.9	0.23	0.75	0.79	0.64	0.83	0.57	0.14	0.22	0.81	0.25	0.28	0.67	0.75	toward_failure	0.28	0.36	0.7	0.87	0.72	0.26	0.23	0.75	pressor_then_fluids	antibiotics_then_fluids	0.31	0.47	0.64	5	0.59	0.55	4	0.48	0.48	0.65	0.52	3	unstable_recovery	0.24	0.33	2	0.6	vasoplegic_sepsis	0.5	-0.11	0.35	0.65	0.11	0.36	0.6	1	0.62	0.41	0.28	-0.11	0.03	0.05	-0.03	-0.03	pressor_then_fluids	0	1
seps_t003	0.78	0.4	0.57	0.61	0.45	0.62	0.38	0.29	0.4	0.61	0.17	0.19	0.82	0.54	toward_failure	0.49	0.2	0.47	0.77	0.82	0.52	0.39	0.53	antibiotics_only	antibiotics_then_fluids	0.16	0.28	0.82	3	0.8	0.77	1	0.76	0.35	0.81	0.79	1	stabilized	0.11	0.14	0	0.35	contained_bacteremia	0.32	-0.34	0.68	0.3	0.3	0.75	0.23	0	0.23	0.8	0.04	-0.05	0.07	-0.02	0.01	-0.24	antibiotics_only	1	1
seps_t004	0.93	0.19	0.81	0.82	0.72	0.88	0.62	0.1	0.18	0.86	0.3	0.32	0.6	0.81	toward_failure	0.22	0.39	0.76	0.88	0.69	0.2	0.17	0.81	mechanical_support_then_source_control	pressor_then_fluids	0.35	0.5	0.57	6	0.51	0.47	5	0.4	0.54	0.61	0.45	4	relapse	0.28	0.37	2	0.68	septic_multiorgan_escalation	0.58	-0.08	0.22	0.72	0.08	0.28	0.66	1	0.69	0.38	0.35	-0.12	0.02	0.06	-0.04	-0.02	mechanical_support_then_source_control	0	1
seps_t005	0.74	0.44	0.53	0.56	0.38	0.57	0.34	0.33	0.45	0.57	0.15	0.17	0.84	0.48	toward_recovery	0.57	0.18	0.41	0.74	0.84	0.59	0.43	0.46	antibiotics_then_fluids	antibiotics_only	0.14	0.24	0.84	3	0.82	0.79	1	0.78	0.32	0.84	0.81	0	stabilized	0.1	0.12	0	0.32	early_source_control	0.31	-0.36	0.71	0.21	0.36	0.84	0.18	0	0.17	0.84	0.03	-0.04	0.08	-0.01	0	-0.26	antibiotics_then_fluids	1	1
seps_t006	0.83	0.28	0.66	0.7	0.54	0.74	0.46	0.21	0.31	0.7	0.2	0.23	0.75	0.64	toward_failure	0.42	0.25	0.61	0.82	0.78	0.37	0.33	0.65	source_control_then_antibiotics	antibiotics_then_fluids	0.23	0.38	0.76	4	0.73	0.7	2	0.68	0.44	0.74	0.67	2	unstable_recovery	0.17	0.22	1	0.51	delayed_source_control	0.41	-0.2	0.49	0.47	0.18	0.48	0.44	1	0.48	0.56	0.19	-0.08	0.05	0.04	-0.02	-0.06	source_control_then_antibiotics	0	1
seps_t007	0.71	0.47	0.5	0.54	0.34	0.49	0.29	0.37	0.47	0.54	0.13	0.15	0.86	0.43	toward_recovery	0.6	0.16	0.37	0.72	0.85	0.61	0.44	0.42	antibiotics_only	antibiotics_then_fluids	0.12	0.22	0.85	2	0.83	0.8	1	0.79	0.3	0.85	0.82	0	stabilized	0.09	0.11	0	0.3	contained_infection	0.3	-0.38	0.72	0.18	0.38	0.86	0.16	0	0.14	0.86	0.03	-0.04	0.08	-0.01	0	-0.27	antibiotics_only	1	1
seps_t008	0.87	0.24	0.73	0.77	0.6	0.81	0.54	0.16	0.24	0.79	0.25	0.27	0.68	0.72	toward_failure	0.31	0.33	0.68	0.84	0.74	0.29	0.25	0.73	pressor_then_fluids_then_antibiotics	antibiotics_then_fluids	0.29	0.44	0.66	5	0.61	0.57	4	0.5	0.46	0.67	0.54	3	relapse	0.22	0.31	2	0.6	vasoplegic_escalation	0.48	-0.13	0.37	0.59	0.14	0.4	0.53	1	0.58	0.46	0.24	-0.1	0.03	0.05	-0.03	-0.03	pressor_then_fluids_then_antibiotics	0	1
seps_t009	0.76	0.41	0.55	0.59	0.41	0.59	0.35	0.31	0.42	0.6	0.16	0.18	0.83	0.5	toward_recovery	0.54	0.18	0.43	0.75	0.83	0.56	0.41	0.49	antibiotics_then_fluids	source_control_then_antibiotics	0.15	0.26	0.83	3	0.81	0.78	1	0.77	0.33	0.83	0.8	1	stabilized	0.1	0.13	0	0.33	mixed_inflammatory_shift	0.33	-0.33	0.68	0.26	0.31	0.78	0.22	0	0.2	0.81	0.04	-0.05	0.07	-0.02	0.01	-0.23	antibiotics_then_fluids	1	1
seps_t010	0.84	0.27	0.67	0.71	0.56	0.75	0.47	0.2	0.29	0.72	0.21	0.24	0.74	0.65	toward_failure	0.39	0.27	0.62	0.83	0.77	0.35	0.32	0.67	source_control_then_fluids	antibiotics_then_fluids	0.24	0.39	0.72	5	0.69	0.65	3	0.61	0.43	0.72	0.62	2	unstable_recovery	0.18	0.25	1	0.54	drainage_delay	0.44	-0.18	0.45	0.44	0.19	0.46	0.4	1	0.45	0.57	0.18	-0.09	0.04	0.04	-0.02	-0.05	source_control_then_fluids	0	1

Clinical Quad Infection Buffer Lag Coupling Sepsis Transition v1.2

What this repo does

This repository contains a Clarus v1.2 benchmark dataset.

The v1.2 layer introduces Temporal Policy Stability Geometry.

Earlier versions evaluate:

system state
trajectory
intervention selection
closed-loop control
counterfactual policy quality

v1.2 extends the framework to evaluate policy handling across time.

The benchmark asks:

does the chosen policy remain correct as conditions evolve
when does a previously correct policy become unsafe
does the controller know when to maintain the current policy
does the controller know when to switch strategy

This is not only a durability benchmark.

It is a benchmark for temporally correct policy handling.

Core quad

The system is defined by four interacting variables.

infection_load
buffer_capacity
lag_burden
coupling_stress

These variables define the state geometry of the system.

All higher-level signals describe how this state changes under pressure, intervention, drift, and feedback.

Clinical variable mapping

Quad Variable	Clinical Measurements	Typical Indicators
infection_load	infectious burden, source severity, microbial pressure	source persistence, bacteremia load, uncontrolled infection
buffer_capacity	physiological reserve, perfusion reserve, immune reserve	remaining compensation, reserve to tolerate septic stress
lag_burden	delayed correction load, unresolved instability debt	late antibiotics, delayed source control, untreated progression
coupling_stress	cross-system destabilization linking infection, perfusion, and organs	vasoplegia, inflammatory spillover, organ strain

Prediction target

label_sepsis_transition_temporal_policy

Binary classification.

1 = temporally correct policy handling
0 = temporally incorrect policy handling

A positive label includes two valid cases:

the original policy remained correct over time and was correctly maintained
the original policy became invalid and the controller correctly identified that policy reselection was required

Label logic

A positive label means the controller handled temporal policy evolution correctly.

That includes:

Stable-policy case

The original policy remains valid over time.

Typical properties:

stabilization is real
trajectory improves materially
control remains aligned
policy drift sensitivity stays low
policy half-life stays high
no reselection trigger is required

Correct-switch case

The original policy degrades over time.

Typical properties:

drift sensitivity rises
policy half-life falls
fragility increases
a reselection trigger should fire
the controller correctly identifies that the policy must change

This means v1.2 rewards both:

correct policy retention
correct switch detection

What v1.2 adds

Earlier versions ask:

where the system is
where it is moving
which intervention is best
whether that intervention remains robust under alternatives

v1.2 adds a temporal question:

does the controller know when to stay with the current policy
and does it know when to switch

This introduces:

policy durability
policy decay
reselection timing
temporal control correctness

The shift is precise.

From:

choosing the correct policy

To:

handling policy evolution correctly over time

New v1.2 temporal policy signals

v1.2 introduces signals describing policy validity under drift.

Signal	Meaning
`policy_drift_sensitivity`	how rapidly the chosen policy becomes invalid as conditions change
`adaptive_policy_margin`	tolerance before the policy must be changed
`stabilization_half_life`	duration for which the chosen policy remains effective
`policy_fragility_score`	sensitivity of the policy to small temporal or state perturbations
`policy_reselection_trigger`	binary indicator that the controller should switch policy
`policy_decay_rate`	rate at which policy effectiveness declines over time
`temporal_control_consistency`	whether control remains coherent across successive temporal states
`reselection_delay_cost`	cost incurred by failing to switch when reselection is needed

These signals define Temporal Policy Stability Geometry.

Example scenario

A representative v1.2 switch case:

the initial intervention is correct at first
the septic state begins to drift
policy effectiveness decays
a reselection trigger should fire
the controller must switch rather than remain on the original path

In v1.2, this can still be a positive case.

The controller is rewarded not only for holding a stable policy, but also for correctly detecting when the policy must change.

Example numeric row pattern

Signal	Value
`control_sequence_alignment_score`	0.64
`policy_drift_sensitivity`	0.65
`stabilization_half_life`	0.36
`policy_fragility_score`	0.60
`policy_reselection_trigger`	1
`temporal_control_consistency`	0.41
`reselection_delay_cost`	0.28
`label_sepsis_transition_temporal_policy`	1

Interpretation:

the original policy does not remain valid
the system correctly signals that policy reselection is needed
the controller is rewarded for correct switch detection

Row structure

Each row includes:

core system variables
trajectory signals
boundary geometry
regime transition signals
intervention competition signals
control sequence signals
temporal policy signals
perturbation and recovery signals
quad delta signals
final outcome fields

This allows a model to evaluate not only whether a policy works, but whether it continues to work as the system changes.

Files

data/train.csv
Full training dataset with labels and all temporal policy signals.
data/tester.csv
Evaluation-style dataset with:
- stabilization_success
- label_sepsis_transition_temporal_policy
removed.
scorer.py
Reference scorer computing binary metrics, temporal policy diagnostics, temporal miss diagnostics, and control diagnostics.
benchmark_spec.json
Machine-readable benchmark specification.
dataset_schema.json
Machine-readable schema with types, ranges, and row order.
README.md
This document.

Evaluation

Primary metric:

recall_temporally_correct_policy_handling

This measures how often the model handles temporal policy evolution correctly.

That includes both:

keeping a policy that remains valid
switching when the policy becomes invalid

Secondary metric:

false_temporal_handling_rate

This measures how often the model predicts temporally correct handling when the policy handling was actually wrong.

Binary metrics:

accuracy
precision
recall
f1
confusion matrix

Temporal policy diagnostics:

temporal_policy_path_accuracy
policy_drift_sensitivity_error
adaptive_policy_margin_error
stabilization_half_life_error
policy_fragility_score_error
policy_reselection_trigger_accuracy
policy_decay_rate_error
temporal_control_consistency_error
reselection_delay_cost_error

Temporal miss diagnostics:

high_uncertainty_temporal_miss_rate
narrow_window_temporal_miss_rate

Control diagnostics:

control_sequence_alignment_accuracy
control_horizon_error
feedback_response_accuracy
intervention_timing_accuracy
adaptation_latency_error
control_stability_error
recovery_consistency_error
recalibration_overuse_rate
controller_oscillation_misread_rate
terminal_pathway_state_accuracy

Dataset construction

The dataset is generated using structured temporal scenarios.

Typical generation steps:

construct an initial system state from the quad variables
simulate one or more intervention policies
evolve the system through temporal drift
measure whether the original policy remains valid
identify whether and when reselection becomes necessary
compute temporal policy signals
assign the benchmark label based on correctness of temporal policy handling

Temporal scenarios are designed to capture two key classes:

Policy retention cases

The original policy remains valid across the relevant horizon.

Policy switch cases

The original policy becomes invalid and reselection is required.

This means v1.2 is designed to test the controller’s capacity for temporal judgment, not just policy endurance.

Running the scorer

python scorer.py data/train.csv predictions.csv

python scorer.py data/train.csv predictions.csv --verbose
Dataset limitations

This dataset evaluates structural temporal policy handling, not exact real-world treatment precision.

Important limits:

temporal drift is represented structurally rather than as a full mechanistic simulation
policy reselection logic is generated at dataset construction time
temporal signals represent control quality and decay, not literal bedside protocols
domain abstractions may omit real-world noise, staffing effects, or intervention delays outside the modeled geometry

These datasets should be treated as benchmarks for temporal control reasoning.

Intended use

This dataset is intended for:

control policy evaluation
temporal decision stability testing
reinforcement learning benchmarking
adaptive controller benchmarking
model robustness research under drift

This dataset is not intended for:

direct clinical decision making
diagnosis
treatment recommendation
deployment without external validation
use as a sole decision system
Structural note

v1.2 marks the move from:

choosing the correct policy

to:

handling policy evolution correctly across time

That means v1.2 evaluates both:

stable-policy retention
correct switch detection

This makes v1.2 the first Clarus layer that formally tests whether a controller knows:

when to stay
when to switch
Position in the Clarus ladder
v0.1 — detection
v0.2 — trajectory
v0.3 — cascade forecasting
v0.4 — boundary discovery
v0.5 — recovery geometry
v0.6 — intervention reasoning
v0.7 — uncertainty geometry
v0.8 — regime transition
v0.9 — intervention competition
v1.0 — closed-loop control
v1.1 — counterfactual and adversarial policy testing
v1.2 — temporal policy stability
Production deployment

This dataset format is suitable for controlled benchmarking in domains where policy validity changes over time.

Examples include:

ICU stabilization under evolving state
sepsis management under delayed source control
infectious escalation under shifting instability
adaptive control in distributed systems
sequential intervention planning in high-risk environments
Enterprise and research collaboration

Clarus evaluates stability and control in complex systems.

The goal is not simply to predict what happens next.

The goal is to determine:

whether the chosen action remains valid
whether the controller can track changing conditions
whether the system knows when the correct move is to switch

v1.2 extends Clarus from control correctness into temporal control correctness.

License

MIT

Downloads last month: 20

Number of rows:

Total file size:

67.3 kB