@prefix this: <https://w3id.org/sciencelive/np/RAUS6GbT3Bu-Np0Ue73q58G_c2HilLhh95Y2b8W18o--M> .
@prefix sub: <https://w3id.org/sciencelive/np/RAUS6GbT3Bu-Np0Ue73q58G_c2HilLhh95Y2b8W18o--M/> .
@prefix np: <http://www.nanopub.org/nschema#> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix prov: <http://www.w3.org/ns/prov#> .
@prefix npx: <http://purl.org/nanopub/x/> .
@prefix dc: <http://purl.org/dc/terms/> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .

sub:Head {
  this: a np:Nanopublication;
    np:hasAssertion sub:assertion;
    np:hasProvenance sub:provenance;
    np:hasPublicationInfo sub:pubinfo .
}

sub:assertion {
  sub:outcome-few-shot-sentinel-2 a <https://w3id.org/sciencelive/o/terms/FORRT-Replication-Outcome>;
    <http://schema.org/endDate> "2026-04-16";
    <http://www.w3.org/2000/01/rdf-schema#label> "ProtoNet achieves 82% on mixed but only 54% on novel-only satellite land cover classes";
    <https://w3id.org/sciencelive/o/terms/hasConclusionDescription> "Prototypical Networks — an AI method that learns to classify new categories from a few labeled examples by computing distances to class prototypes in a learned feature space — successfully classifies Sentinel-2 satellite land cover types when given just 5 labeled images per type (82% accuracy). This shows that Prototypical Networks, originally designed and tested on everyday photographs (mini-ImageNet), transfer to Earth observation data. However, when tested exclusively on land cover types that Prototypical Networks never encountered during training, accuracy drops to 54% (random guessing would be 33%). Prototypical Networks help classify rare land cover types but are not yet reliable enough for operational monitoring of rare habitats from satellite imagery.";
    <https://w3id.org/sciencelive/o/terms/hasConfidenceLevel> <https://w3id.org/sciencelive/o/terms/HighConfidence>;
    <https://w3id.org/sciencelive/o/terms/hasEvidenceDescription> "We trained Prototypical Networks on 7 common land cover classes from the EuroSAT dataset (27,000 real Sentinel-2 satellite image patches, 10 m resolution, across Europe) and tested classification of 3 unseen land cover classes. With 1 labeled example per class: 71.5% accuracy. With 5 examples: 82.1%. With 20 examples: 84.3%. When all classes are unseen (never encountered during training): 53.8%. Evaluated over 600 random classification tasks with 95% confidence intervals.";
    <https://w3id.org/sciencelive/o/terms/hasLimitationsDescription> "Only visible light bands (RGB) from Sentinel-2 were used, discarding 10 additional spectral bands (near-infrared, shortwave infrared) that carry important information for distinguishing vegetation types. The EuroSAT dataset contains 10 broad land cover categories — real Natura 2000 habitat types are more fine-grained and visually similar to each other. Different choices of which classes are treated as \"common\" versus \"rare\" could change the results. Prototypical Networks used a simple 4-block convolutional neural network as backbone, not a larger pre-trained model.";
    <https://w3id.org/sciencelive/o/terms/hasOutcomeRepository> <https://doi.org/10.5281/zenodo.19607662>;
    <https://w3id.org/sciencelive/o/terms/hasValidationStatus> <https://w3id.org/sciencelive/o/terms/PartiallySupported>;
    <https://w3id.org/sciencelive/o/terms/isOutcomeOf> <https://w3id.org/sciencelive/np/RALGj6JatoumnOiiTUie18OUEVGpeAKuhvUkLpAa2Wjjc/few-shot-sentinel-2> .
}

sub:provenance {
  sub:assertion prov:wasAttributedTo <https://orcid.org/0000-0002-1784-2920> .
}

sub:pubinfo {
  <https://orcid.org/0000-0002-1784-2920> <http://xmlns.com/foaf/0.1/name> "Anne Fouilloux" .
  
  this: dc:created "2026-04-17T12:51:01.394Z"^^xsd:dateTime;
    dc:creator <https://orcid.org/0000-0002-1784-2920>;
    dc:license <https://creativecommons.org/licenses/by/4.0/>;
    npx:introduces sub:outcome-few-shot-sentinel-2;
    npx:wasCreatedAt <https://platform.sciencelive4all.org>;
    <http://www.w3.org/2000/01/rdf-schema#label> "NP created using Declaring a replication study outcome according to FORRT";
    <https://w3id.org/np/o/ntemplate/wasCreatedFromTemplate> <https://w3id.org/np/RA2zljn0Nw9SadppOyxZoh-_Rxosslrq-vYG-p9SttnJE> .
  
  sub:sig npx:hasAlgorithm "RSA";
    npx:hasPublicKey "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";
    npx:hasSignature "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";
    npx:hasSignatureTarget this:;
    npx:signedBy <https://orcid.org/0000-0002-1784-2920> .
}