@prefix this: <https://w3id.org/np/RARv-ABCr9uAsxKeqU4U_I7aX2d5v-JtgMOfZatwmlmM8> .
@prefix sub: <https://w3id.org/np/RARv-ABCr9uAsxKeqU4U_I7aX2d5v-JtgMOfZatwmlmM8/> .
@prefix np: <http://www.nanopub.org/nschema#> .
@prefix dct: <http://purl.org/dc/terms/> .
@prefix nt: <https://w3id.org/np/o/ntemplate/> .
@prefix npx: <http://purl.org/nanopub/x/> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix orcid: <https://orcid.org/> .
@prefix prov: <http://www.w3.org/ns/prov#> .
@prefix foaf: <http://xmlns.com/foaf/0.1/> .

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

sub:assertion {
  sub:few-shot-eurosat-synthesis a <https://w3id.org/sciencelive/o/terms/Research-Synthesis>;
    dct:subject <http://www.wikidata.org/entity/Q110797734>, <http://www.wikidata.org/entity/Q199687>,
      <http://www.wikidata.org/entity/Q3001793>, <http://www.wikidata.org/entity/Q6027324>;
    <http://purl.org/spar/cito/isSupportedBy> <https://w3id.org/sciencelive/np/RA7OZmOmun07jDm8q6lq7Ris1W0MU3rptcp3bphOWUJj8>,
      <https://w3id.org/sciencelive/np/RA9PP1TVbvRwEv9UNHYfGWvfCXtxCe3f6_xJnf-YbAgSc>, <https://w3id.org/sciencelive/np/RAUS6GbT3Bu-Np0Ue73q58G_c2HilLhh95Y2b8W18o--M>,
      <https://w3id.org/sciencelive/np/RAqs99x7CDi1tutYKJ6J1zes8PEbUhfjbpf3dkdqSoffQ>;
    <http://schema.org/endDate> "2026-04-18"^^xsd:date;
    rdfs:label "Within-domain few-shot learning is recommended over cross-domain transfer for satellite imagery classification";
    <https://w3id.org/sciencelive/o/terms/hasConditionsDescription> "Applicable to optical satellite imagery (Sentinel-2, Landsat) for land cover and habitat classification tasks where labeled examples  are scarce. Tested on 10 broad EuroSAT land cover categories at 10 m ground resolution using RGB bands only.";
    <https://w3id.org/sciencelive/o/terms/hasLimitationsDescription> "Not tested on fine-grained habitat discrimination such as distinguishing vegetation subtypes within Natura 2000 sites. Only RGB bands used — near-infrared and shortwave infrared bands, which are critical for vegetation analysis, were not included. Results may differ for non-optical sensors (SAR, LiDAR) or very high resolution imagery. The within-domain experiment used a different base/novel class split than a real monitoring scenario would require.";
    <https://w3id.org/sciencelive/o/terms/hasRecommendationDescription> "For Earth observation researchers with limited labeled satellite data: if any labeled satellite imagery is available for your region (even for different classes than your target), use within-domain few-shot learning — train on common classes and classify rare ones. If no satellite training data exists at all, use an off-the-shelf ImageNet-pretrained model as a feature extractor for initial screening, then invest in labeling a small satellite dataset to improve accuracy. Complex meta-learning pipelines are not necessary — standard supervised pretraining achieves comparable results.";
    <https://w3id.org/sciencelive/o/terms/hasSynthesisDescription> "Four experiments comparing within-domain and cross-domain few-shot learning on Sentinel-2 satellite imagery show that within-domain transfer (training on common satellite land cover classes, classifying rare ones) achieves 82% accuracy with 5 labeled examples per class, while cross-domain transfer from everyday photographs achieves 67–76% depending on backbone architecture and training method. The domain gap between photographs and satellite imagery reduces accuracy by 6–15 percentage points. Supervised pretraining on everyday photographs matches episodic meta-learning with 12 times less training time, but both cross-domain approaches remain below within-domain accuracy." .
}

sub:provenance {
  sub:assertion prov:wasAttributedTo orcid:0000-0002-1784-2920 .
}

sub:pubinfo {
  <http://www.wikidata.org/entity/Q110797734> nt:hasLabelFromApi "few-shot learning - machine learning approach that enables a system to learn new tasks or recognize new objects from only a few examples or demonstrations, rather than requiring extensive data" .
  
  <http://www.wikidata.org/entity/Q199687> nt:hasLabelFromApi "remote sensing - acquisition of information about an object or phenomenon without making physical contact with the object, especially the Earth" .
  
  <http://www.wikidata.org/entity/Q3001793> nt:hasLabelFromApi "land cover - nature of the physical material at the surface of the earth" .
  
  <http://www.wikidata.org/entity/Q6027324> nt:hasLabelFromApi "transfer learning - research problem in machine learning (ML) that focuses on storing knowledge gained while solving one problem and applying it to a different but related problem" .
  
  orcid:0000-0002-1784-2920 foaf:name "Anne Fouilloux" .
  
  this: dct:created "2026-04-18T18:55:18.984Z"^^xsd:dateTime;
    dct:creator orcid:0000-0002-1784-2920;
    dct:license <https://creativecommons.org/licenses/by/4.0/>;
    npx:introduces sub:few-shot-eurosat-synthesis;
    npx:wasCreatedAt <https://nanodash.knowledgepixels.com/>;
    nt:wasCreatedFromProvenanceTemplate <https://w3id.org/np/RA7lSq6MuK_TIC6JMSHvLtee3lpLoZDOqLJCLXevnrPoU>;
    nt:wasCreatedFromPubinfoTemplate <https://w3id.org/np/RACJ58Gvyn91LqCKIO9zu1eijDQIeEff28iyDrJgjSJF8>,
      <https://w3id.org/np/RAukAcWHRDlkqxk7H2XNSegc1WnHI569INvNr-xdptDGI>;
    nt:wasCreatedFromTemplate <https://w3id.org/np/RApmrqOEr4f5bJC2vayrTnzhwnuEfAU_I4Pdg8K5JxeBw> .
  
  sub:sig npx:hasAlgorithm "RSA";
    npx:hasPublicKey "MIGfMA0GCSqGSIb3DQEBAQUAA4GNADCBiQKBgQDWv2pJnmDsBOq8OlT1aSvYXSuWT34WOp4FYqEzdnn2F0kqzcFevBqWGZDxJWC0lqCrDEuNfp2QFyPe/+nES9dlHGYIhqPi68fwK6ZiNUotRFxXou+rjFznVvUxtCL8Ede79EBHwWN61QtwSIcU12bLoZsNPFlqQASQ93BJuKlihwIDAQAB";
    npx:hasSignature "FTGbwGUbqi+nCLOqOznThPR9MOLOcTesabQLyyR5desYZTqLy/St4qAbY50Y781/RcoQGvufwkJEp73g0bXf1UKhSykRwfC/NrC01/kh5emK/uuwxgbVd4+avSqwx3cXSa0i5Esx2Xv3J4WFDm8qXw2GhtmWcdDXkthZHQ6MiW4=";
    npx:hasSignatureTarget this:;
    npx:signedBy orcid:0000-0002-1784-2920 .
}