# Wake-word detection on Wake-word detection dataset (test) Source: [Wake-word detection on Wake-word detection dataset (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T06:01:55.796Z Evaluation of Wake-word detection performance on the test split of the Wake-word detection dataset, measured by relative False Alarm rates across five different devices. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Wake-word detection | | Dataset (including any stated split) | Wake-word detection dataset (test) | | Metric | Relative FA (Device A) | | Unit | PERCENT | | Direction | Lower is better | | Results loaded | 10 of 10 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Relative FA (Device A)](): Lower is better; unit: PERCENT. - [Relative FA (Device B)](): Lower is better; unit: PERCENT. - [Relative FA (Device C)](): Lower is better; unit: PERCENT. - [Relative FA (Device D)](): Lower is better; unit: PERCENT. - [Relative FA (Device E)](): Lower is better; unit: PERCENT. ## Recorded results | Method | Relative FA (Device A) | Result date | Report paper | | --- | --- | --- | --- | | Self-FL | 0.92 | Not recorded | 68fa6c5c1e209131cea3395a | | FedAvg-e | 0.97 | Not recorded | 68fa6c5c1e209131cea3395a | | DITTO-e | 0.97 | Not recorded | 68fa6c5c1e209131cea3395a | | PerFedAvg | 1.06 | Not recorded | 68fa6c5c1e209131cea3395a | | DITTO-w | 8.38 | Not recorded | 68fa6c5c1e209131cea3395a | | FedAvg-w | 8.39 | Not recorded | 68fa6c5c1e209131cea3395a | | Self-FL | 93 | Not recorded | 68fa6c5c1e209131cea3395a | | FedAvg-e | 95 | Not recorded | 68fa6c5c1e209131cea3395a | | DITTO-e | 97 | Not recorded | 68fa6c5c1e209131cea3395a | | PerFedAvg | 102 | Not recorded | 68fa6c5c1e209131cea3395a | ### Self-FL Relative FA (Device A): 0.92. Result ID: 698f0b953a6c314849f536bd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scope":"Global","Operating State":"Normal"}. This record does not assert independent reproduction. ### FedAvg-e Relative FA (Device A): 0.97. Result ID: 698f0b953a6c314849f536bf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scope":"Global","Aggregation Weighting":"equal-weighted model averaging","Operating State":"Normal"}. This record does not assert independent reproduction. ### DITTO-e Relative FA (Device A): 0.97. Result ID: 698f0b953a6c314849f536c1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scope":"Global","Aggregation Weighting":"equal-weighted model averaging","Operating State":"Normal"}. This record does not assert independent reproduction. ### PerFedAvg Relative FA (Device A): 1.06. Result ID: 698f0b953a6c314849f536c2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scope":"Global","Aggregation Weighting":"equal-weighted aggregation","Operating State":"Normal"}. This record does not assert independent reproduction. ### DITTO-w Relative FA (Device A): 8.38. Result ID: 698f0b953a6c314849f536c0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scope":"Global","Aggregation Weighting":"sample size-based weighted averaging","Operating State":"Normal"}. This record does not assert independent reproduction. ### FedAvg-w Relative FA (Device A): 8.39. Result ID: 698f0b953a6c314849f536be. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scope":"Global","Aggregation Weighting":"sample size-based weighted averaging","Operating State":"Normal"}. This record does not assert independent reproduction. ### Self-FL Relative FA (Device A): 93. Result ID: 698f0b953a6c314849f536b9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Type":"Personalized","Device State":"Normal"}. This record does not assert independent reproduction. ### FedAvg-e Relative FA (Device A): 95. Result ID: 698f0b953a6c314849f536ba. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Type":"Personalized","Device State":"Normal","Aggregation":"equal-weighted"}. This record does not assert independent reproduction. ### DITTO-e Relative FA (Device A): 97. Result ID: 698f0b953a6c314849f536bb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Type":"Personalized","Device State":"Normal","Aggregation":"equal-weighted"}. This record does not assert independent reproduction. ### PerFedAvg Relative FA (Device A): 102. Result ID: 698f0b953a6c314849f536bc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Type":"Personalized","Device State":"Normal","Aggregation":"equal-weighted"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.