add example approaches

clef24/touche24-web/human-value-detection.html

@@ -94,6 +94,7 @@ <h2 id="synopsis">Synopsis</h2>
 <!-- <li>Input: [data <a href="https://doi.org/10.5281/zenodo.6814563">download</a>]</li> -->
 <!-- <li>Submission: [<a href="run-validator.html">validator</a>] [<a href="https://github.com/touche-webis-de/touche-code/tree/main/semeval23/human-value-detection/evaluator">evaluator</a>] [<a href="https://github.com/touche-webis-de/touche-code/tree/main/semeval23/human-value-detection">baselines</a>] <a href="https://www.tira.io/task/valueeval-at-semeval-2023-human-value-detection">submit</a>] [paper template: <a href="https://www.overleaf.com/latex/templates/valueeval23-paper-template/htprfrpkkfhh">overleaf</a>, <a href="https://github.com/touche-webis-de/touche-code/raw/main/semeval23/human-value-detection/participant-paper-template/participant-paper-template.zip">download</a>, <a href="https://github.com/touche-webis-de/touche-code/tree/main/semeval23/human-value-detection/participant-tables">tables</a>, <a href="https://github.com/touche-webis-de/touche-code/tree/main/semeval23/human-value-detection/value-radar-plot">radar plots</a>] [<a href="https://www.tira.io/task/valueeval-at-semeval-2023-human-value-detection">tira</a>] [<a href="https://softconf.com/acl2023/SemEval2023/">paper submission site</a>]</li> -->
 <li>Data: [<a href="https://zenodo.org/doi/10.5281/zenodo.10396293">download</a>] [<a href="https://knowledge4policy.ec.europa.eu/projects-activities/valuesml-unravelling-expressed-values-media-informed-policy-making_en">project</a>]</li>
+<li>Submission: [<a href="https://github.com/touche-webis-de/touche-code/tree/main/clef24/human-value-detection/approaches">example approaches</a>] [<a href="https://github.com/touche-webis-de/touche-code/tree/main/clef24/human-value-detection/evaluator">evaluator</a>]</li>
 <li>ValueEval'23: [<a href="https://touche.webis.de/semeval23/touche23-web/index.html">website</a>] [<a href="https://values.args.me/">demo</a>]</li>
 </ul>
 <a class="uk-button uk-button-primary" href="https://clef2024-labs-registration.dei.unipd.it/">Register now</a>
@@ -413,7 +414,7 @@ <h2 id="data">Data</h2>
 <p>For sub-task 1 an approach thus has to identify the values for each sentence where at least one is not 0. For sub-task 2 an approach has to identify for those values for a sentence where at least one is 1, whether the 1 is in the attained or constrained column.</p>
 
 <h2 id="submission">Submission</h2>
-<p>The submission system will open soon. Register on the <a href="https://groups.google.com/group/valueeval/subscribe">mailing list</a> to get notified. We recommend to start your approach from our random baselines, which include the code for reading and writing the files and make it easy to later deploy your approach as server or submit and distribute it as Docker image. [random baseline: <a href="https://github.com/touche-webis-de/touche-code/tree/main/clef24/human-value-detection/approaches/random-baseline">python script</a>, <a href="https://github.com/touche-webis-de/touche-code/tree/main/clef24/human-value-detection/approaches/random-baseline-notebook">python notebook</a>]</p>
+<p>The submission system will open soon. Register on the <a href="https://groups.google.com/group/valueeval/subscribe">mailing list</a> to get notified. We recommend to start your approach from one of our <a href="https://github.com/touche-webis-de/touche-code/tree/main/clef24/human-value-detection/approaches">example approaches</a> (in Python), which include the code for reading and writing the files and make it easy to later deploy your approach as server or submit and distribute it as Docker image. [random baseline: <a href="https://github.com/touche-webis-de/touche-code/tree/main/clef24/human-value-detection/approaches/random-baseline">script</a>, <a href="https://github.com/touche-webis-de/touche-code/tree/main/clef24/human-value-detection/approaches/random-baseline-notebook">notebook</a>] [<a href="https://github.com/touche-webis-de/touche-code/tree/main/clef24/human-value-detection/approaches/bert-baseline">bert baseline</a>] [<a href="https://github.com/touche-webis-de/touche-code/tree/main/clef24/human-value-detection/approaches/ollama-baseline">ollama baseline</a>]</p>
 <p>Approaches need to produce run files that have the same format as the <code>labels.tsv</code>, but the numbers can be between 0 and 1 and are interpreted as the confidence of the approach (employed for evaluation via ROC-curves): [<a href="https://github.com/touche-webis-de/touche-code/blob/main/clef24/human-value-detection/toy-dataset/run.tsv">toy example</a>]</p>
 <ul>
 <li><p>For sub-task 1: For each sentence and value, the sum of the numbers in the attained and constrained columns should be the confidence of your approach in that the sentence references the value. A sum &GreaterEqual; 0.5 is treated as a positive prediction for purposes of evaluation with precision, recall, and F1-score.<br>
@@ -424,7 +425,7 @@ <h2 id="submission">Submission</h2>
 <p>Note that you submit for both sub-tasks with the same file. If you want to participate only in sub-task 1, always set the number for constrained to 0. If you want to participate only in sub-task 2, the sum for attained and constrained for a value does not matter, only their ratio. If you want to participate in both sub-tasks, avoid submitting the same number for attained or constrained for a value even if your approach is certain that the value is not referenced: if the approach is wrong and the value is actually referenced, it still matters for sub-task 2 which number is the larger one.</p>
 
 <h2 id="evaluation">Evaluation</h2>
-<p>For both sub-tasks, the submission system will evaluate runs automatically using F1-score, Precision, Recall, and ROC-curves (for each value and averaged). Runs on the sub-task leaderboards are ranked according to averaged F1-score.</p>
+<p>For both sub-tasks, the submission system will evaluate runs automatically using F1-score, Precision, Recall, and ROC-curves (for each value and averaged). Runs on the sub-task leaderboards are ranked according to averaged F1-score. [<a href="https://github.com/touche-webis-de/touche-code/tree/main/clef24/human-value-detection/evaluator">evaluator</a>]</p>
 
 
 <h2><a id="related-work"></a>Related Work</h2>