Find research datasets worth reusing

opencc-by-4.0Jan 2020View details →

Stance in Replies and Quotes (SRQ): A New Dataset For Learning Stance inTwitter Conversations

Automated ways to extract stance (denying vs. supporting opinions) from conversations on social media are essential to advance opinion mining research. Recently, there is a renewed excitement in the field as we see new models attempting to improve the state-of-the-art. However, for training and evaluating the models, the datasets used are often small.  Additionally, these small datasets have uneven class distributions, i.e., only a tiny fraction of the examples in the dataset have favoring and denying stance, and most other examples have no clear stance. Moreover, the existing datasets do not distinguish between the different types of conversations on social media (e.g., replying vs. quoting on Twitter). Because of this, models trained on one event do not generalize to other events.  In the presented work, we create a new dataset by labeling stance in responses to posts on Twitter (both replies and quotes) on controversial issues. To the best of our knowledge, this is currently the largest human-labeled stance dataset for Twitter conversations with over 5200 stance labels. More importantly, we designed a tweet collection methodology that favours the selection of denial-type responses. This class is expected to be more useful in the identification of rumours and determining antagonistic relationships between users.

opencc-by-4.0Jun 1970View details →

FIGURE2 in Stance and gait in the flesh-eating dinosaur Tyrannosuurus

FIGURE2. The knee-joint of a megalosaurian to show the femoral condyle (indicated by an arrow) inserted between the tibia and fibula. This is similar to the knee-jointof Tyrannosaurus.

opencc-by-4.0Mar 2024View details →

Irish Rumour Stance Classification Dataset

We create a new public test set for rumour stance classification with substantial differences from the <a href="https://aclanthology.org/S19-2147/">RumourEval dataset</a> over both vocabulary and stance distribution, making it well-suited for studies on evaluation or adaptation under domain shift.

opencc-by-4.0Jun 2024View details →

Labeled data and models for COVID-19 vaccine related tweets with stance, location, and topics

The dataset contains Tweet IDs along with the location and tweet timestamp. The tweets are labeled based on motivating/demotivating status, stance towards the COVID-19 vaccine, and topic in the tweet text. To comply with Twitter guidelines, we removed the tweet texts and author information. You can use Hydrator API to hydrate the tweets. The repository also contains the machine-learning models for topic modeling, de/motivation classifier, and stance detection from the tweets.

opencc-by-4.0Sep 2021View details →

Same Side Stance Classification Resampled Datasets

The resampled datasets for the <a href="http://sameside.webis.de/">Same Side Stance Classification</a> problem used in the EMNLP'21 paper "On Classifying whether Two Texts are on the Same Side of an Argument". The data is based on the publicly available <a href="https://zenodo.org/record/4382353">S3C training datasets</a>. The data format is JSONlines. Python Load Example: (for every single task split) <pre><code class="language-python">import pandas as pd df_cross_dev = pd.read_json("cross_dev.jsonl", lines=True)</code></pre> For details on how the data was compiled, please refer to our <a href="https://github.com/webis-de/emnlp21-same-stance">code</a>.

opencc-by-4.0Jan 2023View details →

writer stance dataset

Repository: writer stance - 1000 German texts from the X-stance corpus (see References X-stance) annotated for writer stance - two formats: conll and BIO (sort of) - see below - goal: explicit and implicit writer stance wrt. to entities (nouns) and (some) events (verbs)   - implicit and explicit  are not distinguished (no indication which case was annotated) - labels:      BIO  Conll  - in favour: PRO    p   - against:   CON    c   - neutral:   O     no label - 3 Annotations (see References DeInStance) Formats: conll: (parzu parser) 1   Das die ART ART Def|Neut|Nom|Sg 2   det _   _ p2  Arbeitsgesetz   Arbeitsgesetz   N   NN  Neut|Nom|Sg 3   subj    _   _ 3   regelt  regeln  V   VVFIN   3|Sg|Pres|Ind   0   root    _   _ 4   die die ART ART Def|Fem|Acc|Pl  5   det _   _ 5   Arbeitszeiten   Arbeitszeit N   NN  Fem|Acc|Pl  3   obja    _   _ 6   und und KON KON _   3   kon _   _ 7   schützt schützen    V   VVFIN   _|_|Pres|Ind    6   cj  _   _ 8   den die ART ART Def|Masc|Acc|Sg 9   det _   _ p9  Arbeitnehmer    Arbeitnehmer    N   NN  Masc|Acc|Sg 7   obja    _   _ 10  .   .   $.  $.  _   0   root    _   _ head was annotated directly in front of the index (1. column) e.g. "p2  Arbeitsgesetz   Arbeitsgesetz   N   NN  Neut|Nom|Sg 3   subj    _   _" i.e. the writer is in favour of "Arbeitsgesetz" BIO: wordform lemma label Das die O Arbeitsgesetz Arbeitsgesetz PRO regelt regeln O die die O Arbeitszeiten Arbeitszeit O und und O schützt schützen O den die O Arbeitnehmer Arbeitnehmer PRO . . O head was annotated (last column) e.g. "Arbeitsgesetz Arbeitsgesetz PRO" for X-stance corpus see also https://vamvas.ch/more-general-stance-detection-with-x-stance References: @article{X-stance,   author    = {Jannis Vamvas and                Rico Sennrich},   title     = {X-Stance: {A} Multilingual Multi-Target Dataset for Stance Detection},   journal   = {CoRR},   volume    = {abs/2003.08385},   year      = {2020},   url       = {https://arxiv.org/abs/2003.08385},   eprinttype = {arXiv},   eprint    = {2003.08385},   timestamp = {Tue, 24 Mar 2020 16:42:29 +0100},   biburl    = {https://dblp.org/rec/journals/corr/abs-2003-08385.bib},   bibsource = {dblp computer science bibliography, https://dblp.org} } @inproceedings{DeInStance,        booktitle = {17th Conference on Natural Language Processing (KONVENS)},            month = {September},            title = {DeInStance: Creating and Evaluating a {G}erman Corpus for Fine-Grained Inferred Stance Detection},           author = {Anne Gohring and Manfred Klenner and Sophia Conrad},        publisher = {ACL Anthology},             year = {2021},            pages = {213--217},         language = {english},              url = {https://doi.org/10.5167/uzh-207940},         abstract = {We introduce deInStance, a corpus of 1000 politicians? answers in German (de) containing sentences labeled with explicitly expressed and inferred stances - pro and con relations - by 3 annotators. They achieved an acceptable inter-rater agreement given the inherent subjective nature of the task. A first baseline, a fine-tuned BERT-based token classifier, achieved F1-scores of around 70\% . Our focus is on the difficult subclass of sentences comprising only non-polar words, but still with an (implicit) pro or con perspective of the writer.} }

openother-atOct 2022View details →

Migration Reframed - Multilingual stance annotated Twitter news replies on migration in Europe in the context of the Ukrainian crisis

The corresponding paper for this dataset "Migration Reframed? A multilingual analysis on the stance shift in Europe during the Ukrainian crisis" has been published in the ACM Web Conference 2023 (WWW'23), and can be accessed here: <a href="https://doi.org/10.1145/3543507.3583442">https://doi.org/10.1145/3543507.3583442</a> . Please cite this when using the dataset. Twitter dataset of European news and replies to investigate public stance on refugees/migrants around the Ukrainian Crisis. September 2021 to August 2022. Countries: <ul> <li>France</li> <li>Germany</li> <li>Italy</li> <li>Poland</li> <li>Spain</li> </ul> Dataset contains: <ul> <li>Usernames of news outlet accounts on Twitter</li> <li>Tweet IDs of these news accounts during the mentioned period filtered for the migration topic + respective replies from the public</li> <li>Tweet IDs of stance annotated replies</li> <li>8,242 tweet/reply pairs labeled with the stance (positive / negative / neutral) on migrants/refugees (on request)</li> </ul> <table> <caption>Dataset overview by the numbers</caption> <thead> <tr> <th scope="col">Country</th> <th scope="col">News Outlets</th> <th scope="col">News Tweets</th> <th scope="col">Replies</th> <th scope="col">Stance Annotated</th> </tr> </thead> <tbody> <tr> <td>France</td> <td>37</td> <td>2,020</td> <td>32,839</td> <td>500</td> </tr> <tr> <td>Germany</td> <td>72</td> <td>3,752</td> <td>55,317</td> <td>500</td> </tr> <tr> <td>Italy</td> <td>21</td> <td>1,305</td> <td>9,892</td> <td>500</td> </tr> <tr> <td>Poland</td> <td>35</td> <td>3,138</td> <td>27,892</td> <td>6,242</td> </tr> <tr> <td>Spain</td> <td>35</td> <td>1,263</td> <td>20,771</td> <td>500</td> </tr> <tr> <td> </td> <td>200</td> <td>11,478</td> <td>146,711</td> <td>8,242</td> </tr> </tbody> </table> Please note: Stance labels are not included and are only available on request.

opencc-by-4.0Nov 2020View details →

Robot-aided Training of Propulsion During Walking: Effects of Torque Pulses Applied to the Hip and Knee Joints During Stance

Dataset linked with the manuscript "Robot-aided Training of Propulsion During Walking: Effects of Torque Pulses Applied to the Hip and Knee Joints During Stance". Please see attached readme document for details

opencc-by-4.0May 2024View details →

Dataset of Brazilian Federal Senate Session Transcriptions From 2023 with Relevant Topics and Stance Detection Annotations.

[pt-BR] Conjunto de dados de transcrições de sessões do Senado Federal brasileiro de 2023 com anotações de tópicos relevantes e de deteção de posicionamento. Dataset description This set contains transcript data from 203 Federal Senate sessions from the year 2023, with annotations of relevant topics and positioning detection. The file <a href="../api/records/11106904/draft/files/meetings.csv/content" target="_blank" rel="noopener">meetings.csv</a> has the following structure: session_id: Unique event identifier speaker_name: Name of the person who gave the speech party: Political party speech: Speech given The folder <a href="../api/records/11106904/draft/files/ground_truth.zip/content" target="_blank" rel="noopener noreferrer">ground_truth.zip</a> contains 6 JSON annotation files related to the detection of relevant topics and positions, and its main keys are the following: id_session: Unique event identifier response: Model response      list_latent_topics: List of latent topics considered by the model      stances: The stances of each person, according to the model, on a given topic model_response_evaluation: Evaluation of the model's response according to a human annotation     mapping: Mapping between the topic named by the model and the topic named in the annotation     list_latent_topics: It contains four keys which are the lists of topics considered true positives, false positives, true negatives and false negatives. Code used All the code used to process the data can be found at: <a href="https://github.com/helenbc/tcc-notas-taquigraficas">https://github.com/helenbc/tcc-notas-taquigraficas</a> [pt-BR] Descrição do conjunto de dados Este conjunto contém dados de transcrição de 203 sessões do Senado Federal do ano de 2023, com anotações de tópicos relevantes e detecção de posicionamento. O arquivo meetings.csv possui a seguinte estrutura: session_id: Identificador único do evento  speaker_name: Nome da pessoa que fez o discurso  party: Partido político speech: Discurso proferido A pasta ground_truth.zip contém 6 arquivos de anotação JSON relacionados à detecção de tópicos relevantes e posições, e suas principais chaves são as seguintes: id_session: Identificador único do evento response: Resposta do modelo     list_latent_topics: Lista de tópicos latentes considerados pelo modelo     stances: As posições de cada pessoa, de acordo com o modelo, sobre um determinado tópico model_response_evaluation: Avaliação da resposta do modelo de acordo com uma anotação humana     mapping: Mapeamento entre o tópico nomeado pelo modelo e o tópico nomeado na anotação     list_latent_topics: Contém quatro chaves que são as listas de tópicos considerados verdadeiros positivos, falsos positivos, verdadeiros negativos e falsos negativos. Código utilizado Todo o código utilizado para processar os dados pode ser encontrado em: <a href="https://github.com/helenbc/tcc-notas-taquigraficas" target="_new" rel="noreferrer">https://github.com/helenbc/tcc-notas-taquigraficas</a>

opencc-by-4.0Jan 2019View details →

Dataset from: Honert and Zelik (2019) "Foot and shoe responsible for majority of soft tissue work in early stance of walking"

Raw and processed full body kinematics and kinetics for Honert and Zelik, 2019, "Foot and shoe responsible for majority of soft tissue work in early stance of walking" Human Movement Science.

opencc-by-4.0Mar 2024View details →

DIVERSE: Deciphering Internet Views on the U.S. Military Through Video Comment Stance Analysis: A Novel Benchmark Dataset for Stance Classification

Paper citation: Cruickshank, Iain J., and Lynnette Hui Xian Ng. "DIVERSE: Deciphering Internet Views on the US Military Through Video Comment Stance Analysis, A Novel Benchmark Dataset for Stance Classification." arXiv preprint arXiv:2403.03334 (2024). Link to paper: https://arxiv.org/abs/2403.03334

opencc-by-4.0Sep 2021View details →

Same Side Stance Classification Adversarial Test Cases

This dataset contains 175 cases. It is manually created to reveal the ability of models to solve different types of adversarial cases for same side stance predictions more systematically. The examples selected here are derived from the dataset used in the <a href="https://sameside.webis.de/">Same Side Stance Classification</a> shared task. We have selected 25 distinct arguments from the "gay marriage" topic that are short and express their stance clearly. For each selected argument, we construct new arguments of four distinct types to obtain two pairs, one with the same stance, and one with an opposing stance: <ul> <li>Negation: a simple negation of the argument.</li> <li>Paraphrase: alters important words from the argument to synonymous expressions with the same stance.</li> <li>Argument: uses an argument from the same topic and stance, but semantically completely different regarding the first one.</li> <li>Citation: repeats or summarizes the first argument and then expresses agreement or rejection (a case frequently occurring in the dataset).</li> </ul> The types Paraphrase, Argument, and Citation are also formulated in a negated version to create additional test instances for the opposite stance.

opencc-by-4.0Jul 2019View details →

Same Side Stance Classification Challenge

The dataset contains argument pairs which are sampled from <a href="https://zenodo.org/record/4139439#.X-C6nOG-q6M">args.me dataset</a> and cover two topics: abortion and gay marriage. The dataset is used in the <a href="http://sameside.webis.de/">same side stance classification challenge</a> which consists of two experiments (cross-topics and within topics). Cite the dataset as <a href="https://webis.de/publications.html#stein_2021">Benno Stein, Yamen Ajjour, Roxanne El Baff, Khalid Al-Khatib, Philipp Cimiano, and Henning Wachsmuth. Same Side Stance Classification. In Yamen Ajjour et al, editors, Same Side Stance Classification Shared Task 2019, volume 2921, July 2021</a>

opencc-by-4.0Nov 2024View details →

Resources for the paper: "Social Context in Political Stance Detection: Impact and Extrapolation"

This repository contains the resources in our paper [Social Context in Political Stance Detection: Impact and Extrapolation] Ramon Villa-Cox, Evan Williams, Kathleen M. Carley In this work, we explore the performance and extrapolation power of political stance-detection models using an existing large-scale weakly-labeled Twitter dataset collected around the 2019 South American Protests [1]. We construct transformer-based user and tweet encoders to embed users in a low-dimensional space using their text and ego-networks. We then train heterogeneous graph attention networks to predict user stances and contrast their ability to extrapolate stance predictions to different country contexts. The protest dataset, which was collected between September 25 and December 24 of 2019, contains 550k labeled users split unevenly across the four countries and contains over 36 million labeled tweets. It contains an additional 1.1 million unlabeled neighbors and 40 million unlabeled tweets. This repository includes the anonymized datasets necessary to reproduce the results and tables of the paper. In addition, we include the corresponding anonymized resources for the new weakly-labeled dataset around the 2020 Chilean Referendum presented in our paper. Following Twitter's January 2023 User Protection Policy update, tweet or user IDs related to sensitive political events cannot be publicly shared. We respect this policy, and only share: <ul> <li>The anonymized user ID, their weak-stance label, the label predicted by each model and the data split (train, validation or test) the user was assigned to.</li> <li>Anonymized user network edges used by the different network classifiers</li> <li>The type of tweet the edge represents (Original, Reply, or Quote)</li> <li>The User Embeddings produced by the User Transformer and which serve as input for the different network models.</li> </ul> This repository is comprised of the following files: <ol> <li>Main_Predictions.7z: Compressed folder containing anonymized user IDs their stance label and each model’s prediction for the country it was trained on. The performance metrics for each model can be obtained based on the test split for each country. This folder includes the results for the Chilean Referendum.</li> <li>Cross_Predictions.7z: Compressed folder containing the results of the cross-country experiments for each anonymized user. The performance metrics for each model, when applied on a different country can be obtained based on each complete file. Users seen during the training of each model are excluded as described in the paper.</li> <li>Tweet_Level_Edgelists.7z: Compressed folder containing  anonymized tweet edge lists indicating its interaction type (Original, Reply, or Quote).</li> <li>User_Networks.7z: Compressed folder containing  different anonymized user edge lists for each interaction type.</li> <li>Embeddings.zip: Compressed Pytorch tensor files containing the User Embeddings produced by the User Transformer and which serve as input for the different network models. This are provided for the main results and the cross-country and referendum experiments.</li> </ol> The code developed for this study is available at: https://github.com/rvillaco/Protest_Stance_Detection

opencc-by-4.0Jun 2024View details →

Illustrations 6 DoF-Platform, Gait, Stance

Illustrations of 6 DoF-Motion Platform (Moog 6DOF2000E), stance, and gait.

opencc-by-4.0Jul 2024View details →

Examining the Influence of Political Bias on Large Language Model Performance in Stance Classification

Code and dataset for paper "Examining the Influence of Political Bias on Large Language Model Performance in Stance Classification". ICWSM 2025 Preprint: https://arxiv.org/abs/2407.17688 Citation:  @misc{ng2024examininginfluencepoliticalbias,       title={Examining the Influence of Political Bias on Large Language Model Performance in Stance Classification},        author={Lynnette Hui Xian Ng and Iain Cruickshank and Roy Ka-Wei Lee},       year={2024},       eprint={2407.17688},       archivePrefix={arXiv},       primaryClass={cs.CL},       url={https://arxiv.org/abs/2407.17688},  }