44-EFT.WP.Data.ModelCards v1.0 | Chapter 5 Optional Extensions

Home ／ Docs-Technical WhitePaper ／ 44-EFT.WP.Data.ModelCards v1.0

Chapter 5 Optional Extensions

I. Chapter Purpose & Scope

II. Layering & Activation Principles

• Optional: Enable when higher explainability/auditability/privacy/operational observability is required. Once enabled, the field must satisfy this chapter’s type/regex/dependency anchors and metrology checks.
• Cross-volume dependencies: Data contract & export per Core.DataSpec v1.0; units/dimensions & uncertainty per Core.Metrology v1.0; path/arrival-time per Core.Equations v1.1; citations must include “Volume vX.Y:Anchor”.

III. Optional Extensions — Master Table

IV. Field Specs & Examples

1) explainability

explainability:

methods: ["grad", "ig", "lrp"]

coverage: {fraction: 0.25, policy: "topk-classes"}

faithfulness_tests: ["deletion", "insertion"]

artifacts:

- {path:"explain/samples/0001.png", sha256:"..."}

- {path:"explain/report.md", sha256:"..."}

2) compression

compression:

pruning: {method:"magnitude", sparsity:0.5}

quantization: {mode:"int8", scheme:"per-channel", calib:"minmax"}

distillation: {teacher:"eift.cls.large@v1.2", loss:["ce","mse"]}

impact:

accuracy_drop_rel: 0.012

latency_gain_rel: 0.35

power_drop_rel: 0.18

3) privacy_preserving

privacy_preserving:

training:

dp: {epsilon: 6.0, delta: 1e-6, accountant:"rdp"}

inference:

secure_mode: ["tee","he"]

pii_scan: true

data_minimization: true

notes: "No raw PII retained; hashed identifiers only."

4) audit_trails

audit_trails:

lineage:

code_sha256: "..."

data_refs: ["eift.obs.demo@v1.0"]

reports:

- {path:"eval/summary.csv", sha256:"..."}

- {path:"robustness/report.md", sha256:"..."}

dag: {path:"ci/pipeline.dag.json", sha256:"..."}

5) license_constraints

license_constraints:

license: "Apache-2.0"

allowed_use: ["academic","benchmark"]

prohibited_use: ["surveillance","biometric_identification"]

regional_limits: ["EU-GDPR"]

6) observability

observability:

metrics:

- {name:"latency_p99_ms", target: 20, window:"5m"}

- {name:"error_rate", target: 0.005}

- {name:"drift_kl", target: 0.10}

alerts:

- {name:"latency_p99_breach", rule:"latency_p99_ms>20 for 10m", severity:"high"}

dashboards: ["grafana:board/123"]

7) fallbacks

fallbacks:

routes:

- {if:"error_rate>0.01", then:"route_to_baseline"}

- {if:"latency_p99_ms>50", then:"drop_to_rule_based"}

baseline_ref: "eift.cls.baseline@v1.0"

8) ablation

ablation:

factors:

- {name:"augmentation_off", delta_f1_macro:-0.018}

- {name:"no_bn", delta_f1_macro:-0.037}

protocol: {seeds:[0,1,2], repeats:3}

artifacts: [{path:"ablation/table.csv", sha256:"..."}]

V. Machine-Readable Schema (Excerpt, Normative)

# I15-5 Optional Extensions (excerpt)

properties:

explainability:

type: object

properties:

methods: {type: array, items: {type: string}}

coverage: {type: object, properties:{fraction:{type:number}, policy:{type:string}}}

faithfulness_tests: {type: array, items:{type:string}}

artifacts: {type: array, items:{type: object, properties:{path:{type:string}, sha256:{type:string}}}}

compression:

type: object

properties:

pruning: {type: object}

quantization: {type: object}

distillation: {type: object}

impact: {type: object, properties:{accuracy_drop_rel:{type:number}, latency_gain_rel:{type:number}, power_drop_rel:{type:number}}}

privacy_preserving:

type: object

properties:

training: {type: object, properties:{dp:{type: object, properties:{epsilon:{type:number}, delta:{type:number}}}}}

inference: {type: object, properties:{secure_mode:{type: array, items:{type:string}}, pii_scan:{type:boolean}}}

data_minimization: {type: boolean}

audit_trails:

type: object

properties:

lineage: {type: object, properties:{code_sha256:{type:string}, data_refs:{type: array, items:{type:string}}}}

reports: {type: array, items:{type: object, properties:{path:{type:string}, sha256:{type:string}}}}

dag: {type: object, properties:{path:{type:string}, sha256:{type:string}}}

license_constraints:

type: object

properties:

license: {type: string}

allowed_use: {type: array, items:{type:string}}

prohibited_use: {type: array, items:{type:string}}

regional_limits: {type: array, items:{type:string}}

observability:

type: object

properties:

metrics: {type: array}

alerts: {type: array}

dashboards: {type: array, items:{type:string}}

fallbacks:

type: object

properties:

routes: {type: array}

baseline_ref: {type: string}

ablation:

type: object

properties:

factors: {type: array}

protocol: {type: object}

artifacts: {type: array}

(References/artifacts in these sections must appear in the export manifest and be verifiable; dimensions are checked by the Metrology chapter.)

VI. Coupling with Export Manifest (export_manifest supplement)

export_manifest:

artifacts:

- {path:"explain/report.md", sha256:"..."}

- {path:"compression/impact.csv", sha256:"..."}

- {path:"privacy/dp_config.yaml", sha256:"..."}

references:

- "EFT.WP.Core.DataSpec v1.0:EXPORT"

- "EFT.WP.Core.Metrology v1.0:check_dim"

(If any extension is enabled and introduces cross-volume dependencies or new artifacts, it must be reflected in references[] and artifacts[].)

VII. Metrology & Path-Dependence Consistency (when applicable)

1. For path-dependent quantities like T_arr, the Model Card must still register delta_form, path="gamma(ell)", measure="d ell". Two equivalent expressions coexist:
   • T_arr = ( 1 / c_ref ) * ( ∫ n_eff d ell )
   • T_arr = ( ∫ ( n_eff / c_ref ) d ell )
2. All numbers/units pass check_dim; no Chinese in formulas/symbols/definitions.

VIII. Chapter Compliance Checklist

• If any extension is enabled, corresponding schema/regex and dependency anchors are complete and pass Lint.
• see[]/export_manifest.references[] use “Volume vX.Y:Anchor”; no shortcodes or versionless cites.
• For path-dependent entries, delta_form/path/measure are complete and check_dim passes; compression/privacy impacts align with quality/evaluation sections.
• Math expression guardrails observed (backticks/parentheses, no Chinese).