<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE article PUBLIC "-//TaxonX//DTD Taxonomic Treatment Publishing DTD v0 20100105//EN" "../../nlm/tax-treatment-NS0.dtd">
<article xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:tp="http://www.plazi.org/taxpub" article-type="research-article">
  <front>
    <journal-meta>
      <journal-id journal-id-type="publisher-id">63</journal-id>
      <journal-id journal-id-type="index">urn:lsid:arphahub.com:pub:0E0032F4-55AE-5263-8B3C-F4DD637C30C2</journal-id>
      <journal-title-group>
        <journal-title xml:lang="en">Biodiversity Information Science and Standards</journal-title>
        <abbrev-journal-title xml:lang="en">BISS</abbrev-journal-title>
      </journal-title-group>
      <issn pub-type="epub">2535-0897</issn>
      <publisher>
        <publisher-name>Pensoft Publishers</publisher-name>
      </publisher>
    </journal-meta>
    <article-meta>
      <article-id pub-id-type="doi">10.3897/biss.8.142382</article-id>
      <article-id pub-id-type="publisher-id">142382</article-id>
      <article-id pub-id-type="manuscript">26263</article-id>
      <article-categories>
        <subj-group subj-group-type="heading">
          <subject>Conference Abstract</subject>
        </subj-group>
        <subj-group subj-group-type="conference-part">
          <subject>SYM15 - Emergent AI Contributions to Biodiversity Data &amp; Data Standardisation: Opportunities, challenges, and a year in review</subject>
        </subj-group>
        <subj-group subj-group-type="biological_taxon">
          <subject>Plantae</subject>
        </subj-group>
        <subj-group subj-group-type="scientific_subject">
          <subject>Artificial intelligence</subject>
          <subject>Computer applications</subject>
          <subject>Computer &amp; Information sciences</subject>
        </subj-group>
        <subj-group subj-group-type="geographical_area">
          <subject>Americas</subject>
          <subject>Central America</subject>
          <subject>Central America and the Caribbean</subject>
          <subject>Costa Rica</subject>
        </subj-group>
      </article-categories>
      <title-group>
        <article-title>Relation Extraction From Unstructured Species Descriptions Using TaxonNERD and LLaMA 2 7B</article-title>
      </title-group>
      <contrib-group content-type="authors">
        <contrib contrib-type="author" corresp="yes">
          <name name-style="western">
            <surname>Rios Montero</surname>
            <given-names>Fabricio De Jesus</given-names>
          </name>
          <email xlink:type="simple">fabrirm15@gmail.com</email>
          <xref ref-type="aff" rid="A1">1</xref>
        </contrib>
        <contrib contrib-type="author" corresp="no">
          <name name-style="western">
            <surname>Rodríguez</surname>
            <given-names>Ervin</given-names>
          </name>
          <xref ref-type="aff" rid="A2">2</xref>
        </contrib>
        <contrib contrib-type="author" corresp="no">
          <name name-style="western">
            <surname>Mora Cross</surname>
            <given-names>Maria</given-names>
          </name>
          <uri content-type="orcid">https://orcid.org/0000-0002-3457-0963</uri>
          <xref ref-type="aff" rid="A2">2</xref>
        </contrib>
      </contrib-group>
      <aff id="A1">
        <label>1</label>
        <addr-line content-type="verbatim">Computer science, Heredia, Costa Rica</addr-line>
        <institution>Computer science</institution>
        <addr-line content-type="city">Heredia</addr-line>
        <country>Costa Rica</country>
      </aff>
      <aff id="A2">
        <label>2</label>
        <addr-line content-type="verbatim">Computer science, Alajuela, Costa Rica</addr-line>
        <institution>Computer science</institution>
        <addr-line content-type="city">Alajuela</addr-line>
        <country>Costa Rica</country>
      </aff>
      <author-notes>
        <fn fn-type="corresp">
          <p>Corresponding author: Fabricio De Jesus Rios Montero (<email xlink:type="simple">fabrirm15@gmail.com</email>).</p>
        </fn>
        <fn fn-type="edited-by">
          <p>Academic editor: </p>
        </fn>
      </author-notes>
      <pub-date pub-type="collection">
        <year>2024</year>
      </pub-date>
      <pub-date pub-type="epub">
        <day>22</day>
        <month>11</month>
        <year>2024</year>
      </pub-date>
      <volume>8</volume>
      <elocation-id>e142382</elocation-id>
      <uri content-type="arpha" xlink:href="http://openbiodiv.net/02B25165-9257-5C01-AB3D-B5963825AE32">02B25165-9257-5C01-AB3D-B5963825AE32</uri>
      <permissions>
        <copyright-statement>Fabricio De Jesus Rios Montero, Ervin Rodríguez, Maria Mora Cross</copyright-statement>
        <license license-type="creative-commons-attribution" xlink:href="http://creativecommons.org/licenses/by/4.0/" xlink:type="simple">
          <license-p>This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.</license-p>
        </license>
      </permissions>
      <abstract>
        <label>Abstract</label>
        <p>Ontologies are essential tools for organizing information on taxonomy, ecology, and inter-species relationships, helping to standardize ecological data and facilitate integration of large datasets. Combining ontologies with advanced Natural Language Processing (NLP) techniques, such as Named Entity Recognition (NER) and Relation Extraction (RE), has greatly improved the discovery of insights from unstructured scientific texts, particularly in biodiversity (<xref ref-type="bibr" rid="B12161765">Gabud et al. 2023</xref>, <xref ref-type="bibr" rid="B12161777">Abdelmageed et al. 2022</xref>, <xref ref-type="bibr" rid="B12161792">Hearst 1992</xref>).</p>
        <p>This study combines ontologies and NLP to analyze complex trophic interactions among animal species (<xref ref-type="bibr" rid="B12161765">Gabud et al. 2023</xref>), using a dataset (<xref ref-type="bibr" rid="B12254115">National Biodiversity Institute of Costa Rica (INBio) 2015</xref>) containing species descriptions in English and Spanish. We applied <ext-link ext-link-type="uri" xlink:href="https://github.com/nleguillarme/taxonerd">TaxoNERD</ext-link> to identify taxonomic entities (<xref ref-type="bibr" rid="B12161802">Le Guillarme and Thuiller 2021</xref>) and we fine-tuned the Large Language Model Meta AI (LLaMA 2 7B) to extract feeding interactions and predator-prey relationships <styled-content style="display: none;"> </styled-content><styled-content style="display: none;"> </styled-content> (<xref ref-type="bibr" rid="B12161897">CheeKean 2023</xref>), due to its effectiveness in handling complex language patterns and its adaptability to diverse scientific domains.</p>
        <p>Our results (Fig. <xref ref-type="fig" rid="F12245781">1</xref>) showed a recall of 0.73 and a precision of 0.68, indicating that the model effectively identifies feeding relationships in most cases. However, the lower precision suggests that the model may still capture some unrelated interactions, highlighting an area for improvement to reduce false positives and increase accuracy (<xref ref-type="bibr" rid="B12162104">Touvron et al. 2023</xref>). Previous studies also emphasize the need for further refinement of relation extraction models to enhance accuracy (<xref ref-type="bibr" rid="B12162175">Mora-Cross et al. 2023</xref>). The structured dataset offers valuable insights into species’ diets and roles, contributing to biodiversity research and conservation efforts (<xref ref-type="bibr" rid="B12162175">Mora-Cross et al. 2023</xref>, <xref ref-type="bibr" rid="B12162104">Touvron et al. 2023</xref>).</p>
        <p>Moreover, this research highlights the potential of integrating AI-driven tools with ontological frameworks to manage and analyze biodiversity data at scale (<xref ref-type="bibr" rid="B12161777">Abdelmageed et al. 2022</xref>). By transforming unstructured text into structured data, we make ecological information more accessible, supporting better decision-making in conservation strategies (<xref ref-type="bibr" rid="B12161777">Abdelmageed et al. 2022</xref>, <xref ref-type="bibr" rid="B12161792">Hearst 1992</xref>). This approach scales well with the growing volume of biodiversity data, offering a more efficient and accurate method for analyzing species interactions, which are crucial for ecosystem management and endangered species protection (<xref ref-type="bibr" rid="B12161765">Gabud et al. 2023</xref>).</p>
      </abstract>
      <kwd-group>
        <label>Keywords</label>
        <kwd>biodiversity</kwd>
        <kwd>ontologies</kwd>
        <kwd>Named Entity Recognition (NER)</kwd>
        <kwd>Relation Extraction (RE)</kwd>
        <kwd>LLaMA2-7b</kwd>
        <kwd>feeding relationships</kwd>
      </kwd-group>
      <counts>
        <fig-count count="1"/>
        <table-count count="0"/>
        <ref-count count="8"/>
      </counts>
    </article-meta>
    <notes>
      <sec sec-type="Presenting author">
        <title>Presenting author</title>
        <p>Fabricio Ríos Montero</p>
      </sec>
      <sec sec-type="Presented at">
        <title>Presented at</title>
        <p>SPNHC-TDWG 2024</p>
      </sec>
    </notes>
  </front>
  
  <back>
    <ack>
      <title>Acknowledgements</title>
      <p>This work was made possible through the support provided by the Instituto Nacional de Costa Rica (ITCR), the International Development Research Center (IDRC) through the Central American Higher University Council (CSUCA),  and the Costa Rican Innovation and Research Promoter of the Ministry of Science, Innovation, Technology, and Telecommunications (MICITT) of Costa Rica.</p>
    </ack>
    <sec sec-type="Conflicts of interest">
      <title>Conflicts of interest</title>
      <p>No conflict of interest to declare</p>
      <p>Disclaimer: This article is (co-)authored by any of the Editors-in-Chief, Managing Editors or their deputies in this journal.</p>
    </sec>
    <ref-list>
      <title>References</title>
      <ref id="B12161777">
        <element-citation publication-type="article">
          <person-group person-group-type="author">
            <name name-style="western">
              <surname>Abdelmageed</surname>
              <given-names>Nora</given-names>
            </name>
            <name name-style="western">
              <surname>Löffler</surname>
              <given-names>Felicitas</given-names>
            </name>
            <name name-style="western">
              <surname>Feddoul</surname>
              <given-names>Leila</given-names>
            </name>
            <name name-style="western">
              <surname>Algergawy</surname>
              <given-names>Alsayed</given-names>
            </name>
            <name name-style="western">
              <surname>Samuel</surname>
              <given-names>Sheeba</given-names>
            </name>
            <name name-style="western">
              <surname>Gaikwad</surname>
              <given-names>Jitendra</given-names>
            </name>
            <name name-style="western">
              <surname>Kazem</surname>
              <given-names>Anahita</given-names>
            </name>
            <name name-style="western">
              <surname>König-Ries</surname>
              <given-names>Birgitta</given-names>
            </name>
            <etal/>
          </person-group>
          <year>2022</year>
          <article-title>BiodivNERE: Gold standard corpora for named entity recognition and relation extraction in the biodiversity domain</article-title>
          <source>Biodiversity Data Journal</source>
          <volume>10</volume>
          <pub-id pub-id-type="doi">10.3897/bdj.10.e89481</pub-id>
        </element-citation>
      </ref>
      <ref id="B12161897">
        <element-citation publication-type="website">
          <person-group person-group-type="author">
            <name name-style="western">
              <surname>CheeKean</surname>
            </name>
          </person-group>
          <article-title>Understanding Llama2: KV Cache, Grouped Query Attention, Rotary Embedding and More</article-title>
          <uri>https://plainenglish.io/community/understanding-llama2-kv-cache-grouped-query-attention-rotary-embedding-and-more-9a79bd</uri>
        </element-citation>
      </ref>
      <ref id="B12161765">
        <element-citation publication-type="article">
          <person-group person-group-type="author">
            <name name-style="western">
              <surname>Gabud</surname>
              <given-names>Roselyn</given-names>
            </name>
            <name name-style="western">
              <surname>Lapitan</surname>
              <given-names>Portia</given-names>
            </name>
            <name name-style="western">
              <surname>Mariano</surname>
              <given-names>Vladimir</given-names>
            </name>
            <name name-style="western">
              <surname>Mendoza</surname>
              <given-names>Eduardo</given-names>
            </name>
            <name name-style="western">
              <surname>Pampolina</surname>
              <given-names>Nelson</given-names>
            </name>
            <name name-style="western">
              <surname>Clariño</surname>
              <given-names>Maria Art Antonette</given-names>
            </name>
            <name name-style="western">
              <surname>Batista-Navarro</surname>
              <given-names>Riza</given-names>
            </name>
            <etal/>
          </person-group>
          <year>2023</year>
          <article-title>A Hybrid of Rule-based and Transformer-based Approaches for Relation Extraction in Biodiversity Literature</article-title>
          <source>Proceedings of the 2nd Workshop on Pattern-based Approaches to NLP in the Age of Deep Learning</source>
          <fpage>103</fpage>
          <lpage>113</lpage>
          <pub-id pub-id-type="doi">10.18653/v1/2023.pandl-1.10</pub-id>
        </element-citation>
      </ref>
      <ref id="B12161792">
        <element-citation publication-type="article">
          <person-group person-group-type="author">
            <name name-style="western">
              <surname>Hearst</surname>
              <given-names>M. A.</given-names>
            </name>
            <etal/>
          </person-group>
          <year>1992</year>
          <article-title>Automatic acquisition of hyponyms from large text corpora</article-title>
          <source>Proceedings of the 14th conference on Computational linguistics  -</source>
          <volume>2</volume>
          <pub-id pub-id-type="doi">10.3115/992133.992154</pub-id>
        </element-citation>
      </ref>
      <ref id="B12161802">
        <element-citation publication-type="article">
          <person-group person-group-type="author">
            <name name-style="western">
              <surname>Le Guillarme</surname>
              <given-names>Nicolas</given-names>
            </name>
            <name name-style="western">
              <surname>Thuiller</surname>
              <given-names>Wilfried</given-names>
            </name>
            <etal/>
          </person-group>
          <year>2021</year>
          <article-title>TaxoNERD: Deep neural models for the recognition of taxonomic entities in the ecological and evolutionary literature</article-title>
          <source>Methods in Ecology and Evolution</source>
          <volume>13</volume>
          <issue>3</issue>
          <fpage>625</fpage>
          <lpage>641</lpage>
          <pub-id pub-id-type="doi">10.1111/2041-210x.13778</pub-id>
        </element-citation>
      </ref>
      <ref id="B12162175">
        <element-citation publication-type="article">
          <person-group person-group-type="author">
            <name name-style="western">
              <surname>Mora-Cross</surname>
              <given-names>Maria</given-names>
            </name>
            <name name-style="western">
              <surname>Ulate</surname>
              <given-names>William</given-names>
            </name>
            <name name-style="western">
              <surname>Retana Chacón</surname>
              <given-names>Brandon</given-names>
            </name>
            <name name-style="western">
              <surname>Biarreta Portillo</surname>
              <given-names>María</given-names>
            </name>
            <name name-style="western">
              <surname>Castro Ramírez</surname>
              <given-names>Josué David</given-names>
            </name>
            <name name-style="western">
              <surname>Chavarria Madriz</surname>
              <given-names>Jose</given-names>
            </name>
            <etal/>
          </person-group>
          <year>2023</year>
          <article-title>Structuring Information from Plant Morphological Descriptions using Open Information Extraction</article-title>
          <source>Biodiversity Information Science and Standards</source>
          <volume>7</volume>
          <pub-id pub-id-type="doi">10.3897/biss.7.113055</pub-id>
        </element-citation>
      </ref>
      <ref id="B12254115">
        <element-citation publication-type="software">
          <person-group person-group-type="author">
            <name name-style="western">
              <surname>(INBio)</surname>
              <given-names>National Biodiversity Institute of Costa Rica</given-names>
            </name>
          </person-group>
          <year>2015</year>
          <article-title><italic>Atta</italic>: Species Records from Costa Rica Documented between 1999 and 2015 [Dataset]</article-title>
          <publisher-name>National Biodiversity Institute of Costa Rica</publisher-name>
          <uri>https://docs.google.com/spreadsheets/d/1EVIjVQYE7gw5m4uGPWrn6rXIgcnXLkqr/edit?usp=sharing&amp;ouid=112437040868151967020&amp;rtpof=true&amp;sd=true</uri>
        </element-citation>
      </ref>
      <ref id="B12162104">
        <element-citation publication-type="article">
          <person-group person-group-type="author">
            <name name-style="western">
              <surname>Touvron</surname>
              <given-names>Hugo</given-names>
            </name>
            <name name-style="western">
              <surname>Bojanowski</surname>
              <given-names>Piotr</given-names>
            </name>
            <name name-style="western">
              <surname>Caron</surname>
              <given-names>Mathilde</given-names>
            </name>
            <name name-style="western">
              <surname>Cord</surname>
              <given-names>Matthieu</given-names>
            </name>
            <name name-style="western">
              <surname>El-Nouby</surname>
              <given-names>Alaaeldin</given-names>
            </name>
            <name name-style="western">
              <surname>Grave</surname>
              <given-names>Edouard</given-names>
            </name>
            <name name-style="western">
              <surname>Izacard</surname>
              <given-names>Gautier</given-names>
            </name>
            <name name-style="western">
              <surname>Joulin</surname>
              <given-names>Armand</given-names>
            </name>
            <name name-style="western">
              <surname>Synnaeve</surname>
              <given-names>Gabriel</given-names>
            </name>
            <name name-style="western">
              <surname>Verbeek</surname>
              <given-names>Jakob</given-names>
            </name>
            <name name-style="western">
              <surname>Jégou</surname>
              <given-names>Hervé</given-names>
            </name>
            <etal/>
          </person-group>
          <year>2023</year>
          <article-title>ResMLP: Feedforward Networks for Image Classification With Data-Efficient Training</article-title>
          <source>IEEE Transactions on Pattern Analysis and Machine Intelligence</source>
          <volume>45</volume>
          <issue>4</issue>
          <fpage>5314</fpage>
          <lpage>5321</lpage>
          <pub-id pub-id-type="doi">10.1109/tpami.2022.3206148</pub-id>
        </element-citation>
      </ref>
    </ref-list>
  </back>
  <floats-group>
    <fig id="F12245781" position="float" orientation="portrait">
      <object-id content-type="arpha">6CF997AE-DCA4-5433-ABE3-67FDCC815BED</object-id>
      
      <label>Figure 1.</label>
      <caption>
        <p>Distribution of <ext-link ext-link-type="uri" xlink:href="https://huggingface.co/spaces/evaluate-metric/bertscore">BERTScore</ext-link> (Bidirectional Encoder Representations from Transformers) metrics (F1, Precision, Recall) with most scores between 0.6 and 0.9. Outliers in Precision suggest areas for improving accuracy and reducing false positives.</p>
      </caption>
      <graphic xlink:href="biss-08-e142382-g001.png" position="float" id="oo_1173450.png" orientation="portrait" xlink:type="simple">
        <uri content-type="original_file">https://binary.pensoft.net/fig/1173450</uri>
      </graphic>
    </fig>
  </floats-group>
</article>
