{"id":27879393,"url":"https://github.com/gesistsa/sweater","last_synced_at":"2025-07-22T04:04:35.523Z","repository":{"id":42663826,"uuid":"324136564","full_name":"gesistsa/sweater","owner":"gesistsa","description":"👚 Speedy Word Embedding Association Test \u0026 Extras using R","archived":false,"fork":false,"pushed_at":"2025-02-01T22:39:43.000Z","size":24564,"stargazers_count":30,"open_issues_count":6,"forks_count":5,"subscribers_count":3,"default_branch":"v0.1","last_synced_at":"2025-07-14T19:28:04.250Z","etag":null,"topics":["bias-detection","r","textanalysis","wordembedding"],"latest_commit_sha":null,"homepage":"","language":"R","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"gpl-3.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/gesistsa.png","metadata":{"files":{"readme":"README.Rmd","changelog":null,"contributing":null,"funding":null,"license":"LICENSE.md","code_of_conduct":"CODE_OF_CONDUCT.md","threat_model":null,"audit":null,"citation":"CITATION.cff","codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null}},"created_at":"2020-12-24T11:06:39.000Z","updated_at":"2025-02-27T23:23:04.000Z","dependencies_parsed_at":"2023-01-22T02:45:13.798Z","dependency_job_id":"1bd82597-4c9c-4ce0-9f62-05ce3c3b0fda","html_url":"https://github.com/gesistsa/sweater","commit_stats":{"total_commits":123,"total_committers":2,"mean_commits":61.5,"dds":0.008130081300813052,"last_synced_commit":"5a233e804cea5e526e161b2420d3f443fac97bbc"},"previous_names":["gesistsa/sweater","chainsawriot/sweater"],"tags_count":5,"template":false,"template_full_name":null,"purl":"pkg:github/gesistsa/sweater","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/gesistsa%2Fsweater","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/gesistsa%2Fsweater/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/gesistsa%2Fsweater/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/gesistsa%2Fsweater/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/gesistsa","download_url":"https://codeload.github.com/gesistsa/sweater/tar.gz/refs/heads/v0.1","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/gesistsa%2Fsweater/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":265534518,"owners_count":23783844,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["bias-detection","r","textanalysis","wordembedding"],"created_at":"2025-05-05T03:21:25.108Z","updated_at":"2025-07-22T04:04:35.494Z","avatar_url":"https://github.com/gesistsa.png","language":"R","readme":"---\noutput: github_document\n---\n\n\u003c!-- README.md is generated from README.Rmd. Please edit that file --\u003e\n\n```{r, include = FALSE}\nknitr::opts_chunk$set(\n  collapse = TRUE,\n  comment = \"#\u003e\",\n  fig.path = \"man/figures/README-\",\n  out.width = \"100%\"\n  )\nset.seed(46709394)\ndevtools::load_all()\n```\n\n# sweater \u003cimg src=\"man/figures/sweater_logo.svg\" align=\"right\" height=\"200\" /\u003e\n\n\u003c!-- badges: start --\u003e\n[![CRAN status](https://www.r-pkg.org/badges/version/sweater)](https://CRAN.R-project.org/package=sweater)\n[![DOI](https://joss.theoj.org/papers/10.21105/joss.04036/status.svg)](https://doi.org/10.21105/joss.04036)\n[![R-CMD-check](https://github.com/gesistsa/sweater/actions/workflows/R-CMD-check.yaml/badge.svg)](https://github.com/gesistsa/sweater/actions/workflows/R-CMD-check.yaml)\n\u003c!-- badges: end --\u003e\n\nThe goal of sweater (**S**peedy **W**ord **E**mbedding **A**ssociation **T**est \u0026 **E**xtras using **R**) is to test for associations among words in word embedding spaces. The methods provided by this package can also be used to test for unwanted associations, or biases.\n\nThe package provides functions that are speedy. They are either implemented in C++, or are speedy but accurate approximation of the original implementation proposed by Caliskan et al (2017). See the benchmark [here](https://github.com/gesistsa/sweater/blob/master/paper/benchmark.md).\n\nThis package provides extra methods such as Relative Norm Distance, Embedding Coherence Test, SemAxis and Relative Negative Sentiment Bias.\n\nIf your goal is to reproduce the analysis in Caliskan et al (2017), please consider using the [original Java program](https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/DX4VWP\u0026version=2.0) or the R package [cbn](https://github.com/conjugateprior/cbn) by Lowe. To reproduce the analysis in Garg et al (2018), please consider using the [original Python program](https://github.com/nikhgarg/EmbeddingDynamicStereotypes). To reproduce the analysis in Mazini et al (2019), please consider using the [original Python program](https://github.com/TManzini/DebiasMulticlassWordEmbedding/).\n\nPlease cite this software as:\n\nChan, C., (2022). sweater: Speedy Word Embedding Association Test and Extras Using R. Journal of Open Source Software, 7(72), 4036, https://doi.org/10.21105/joss.04036\n\nFor a BibTeX entry, use the output from `citation(package = \"sweater\")`.\n\n## Installation\n\nRecommended: install the latest development version\n\n``` r\nremotes::install_github(\"gesistsa/sweater\")\n```\n\nor the \"stable\" release\n\n```r\ninstall.packages(\"sweater\")\n```\n\n## Notation of a query\n\nAll tests in this package use the concept of queries (see Badilla et al., 2020) to study associations in the input word embeddings `w`. This package uses the \"STAB\" notation from Brunet et al (2019). [^1]\n\n[^1]: In the pre 0.1.0 version of this package, the package used `S`, `T`, `A`, and `B` as the main parameters. It was later rejected because the symbol `T` is hardlinked to the logical value `TRUE` [as a global variable](https://stat.ethz.ch/R-manual/R-devel/library/base/html/logical.html); and it is considered to be a [bad style](https://style.tidyverse.org/syntax.html) to use the symbol `T`. Accordingly, they were renamed to `S_words`, `T_words`, `A_words`, and `B_words` respectively. But in general, please stop using the symbol `T` to represent `TRUE`!\n\nAll tests depend on two types of words. The first type, namely, `S_words` and `T_words`, is *target words* (or *neutral words* in Garg et al). In the case of studying biases, these are words that **should** have no bias. For instance, the words such as \"nurse\" and \"professor\" can be used as target words to study the gender bias in word embeddings. One can also separate these words into two sets, `S_words` and `T_words`, to group words by their perceived bias. For example, Caliskan et al. (2017) grouped target words into two groups: mathematics (\"math\", \"algebra\", \"geometry\", \"calculus\", \"equations\", \"computation\", \"numbers\", \"addition\") and arts (\"poetry\", \"art\", \"dance\", \"literature\", \"novel\", \"symphony\", \"drama\", \"sculpture\"). Please note that also `T_words` is not always required.\n\nThe second type, namely `A_words` and `B_words`, is *attribute words* (or *group words* in Garg et al). These are words with known properties in relation to the bias that one is studying. For example, Caliskan et al. (2017) used gender-related words such as \"male\", \"man\", \"boy\", \"brother\", \"he\", \"him\", \"his\", \"son\" to study gender bias. These words qualify as attribute words because we know they are related to a certain gender.\n\nIt is recommended using the function `query()` to make a query and `calculate_es()` to calculate the effect size.\n\n## Available methods\n\n| Target words     | Attribute words  | Method                                                      | `method` argument | Suggested by `query`? | legacy functions [^legacy]                         |\n|------------------|------------------|-------------------------------------------------------------|-------------------|-----------------------|----------------------------------------------------|\n| S_words          | A_words          | Mean Average Cosine Similarity (Mazini et al. 2019)         | \"mac\"             | yes                   | mac(), mac_es()                                    |\n| S_words          | A_words, B_words | Relative Norm Distance (Garg et al. 2018)                   | \"rnd\"             | yes                   | rnd(), rnd_es()                                    |\n| S_words          | A_words, B_words | Relative Negative Sentiment Bias (Sweeney \u0026 Najafian. 2019) | \"rnsb\"            | no                    | rnsb(), rnsb_es()                                  |\n| S_words          | A_words, B_words | Embedding Coherence Test (Dev \u0026 Phillips. 2019)             | \"ect\"             | no                    | ect(), ect_es(), plot_ect()                        |\n| S_words          | A_words, B_words | SemAxis (An et al. 2018)                                    | \"semaxis\"         | no                    | semaxis()                                          |\n| S_words          | A_words, B_words | Normalized Association Score (Caliskan et al. 2017)         | \"nas\"             | no                    | nas()                                              |\n| S_words, T_words | A_words, B_words | Word Embedding Association Test (Caliskan et al. 2017)      | \"weat\"            | yes                   | weat(), weat_es(), weat_resampling(), weat_exact() |\n| S_words, T_words | A_words, B_words | Word Embeddings Fairness Evaluation (Badilla et al. 2020)   | To be implemented |                       |                                                    |\n\n## Example: Mean Average Cosine Similarity\n\nThe simplest form of bias detection is Mean Average Cosine Similarity (Mazini et al. 2019). The same method was used in Kroon et al. (2020). `googlenews` is a subset of [the pretrained word2vec word embeddings provided by Google](https://code.google.com/archive/p/word2vec/).\n\nBy default, the `query()` function guesses the method you want to use based on the combination of target words and attribute words provided (see the \"Suggested?\" column in the above table). You can also make this explicit by specifying the `method` argument. Printing the returned object shows the effect size (if available) as well as the functions that can further process the object: `calculate_es` and `plot`. Please read the help file of `calculate_es` (`?calculate_es`) on what is the meaning of the effect size for a specific test.\n\n```{r, eval = FALSE}\nrequire(sweater)\n```\n\n```{r mac_neg}\nS1 \u003c- c(\"janitor\", \"statistician\", \"midwife\", \"bailiff\", \"auctioneer\",\n\"photographer\", \"geologist\", \"shoemaker\", \"athlete\", \"cashier\",\n\"dancer\", \"housekeeper\", \"accountant\", \"physicist\", \"gardener\",\n\"dentist\", \"weaver\", \"blacksmith\", \"psychologist\", \"supervisor\",\n\"mathematician\", \"surveyor\", \"tailor\", \"designer\", \"economist\",\n\"mechanic\", \"laborer\", \"postmaster\", \"broker\", \"chemist\", \"librarian\",\n\"attendant\", \"clerical\", \"musician\", \"porter\", \"scientist\", \"carpenter\",\n\"sailor\", \"instructor\", \"sheriff\", \"pilot\", \"inspector\", \"mason\",\n\"baker\", \"administrator\", \"architect\", \"collector\", \"operator\",\n\"surgeon\", \"driver\", \"painter\", \"conductor\", \"nurse\", \"cook\",\n\"engineer\", \"retired\", \"sales\", \"lawyer\", \"clergy\", \"physician\",\n\"farmer\", \"clerk\", \"manager\", \"guard\", \"artist\", \"smith\", \"official\",\n\"police\", \"doctor\", \"professor\", \"student\", \"judge\", \"teacher\",\n\"author\", \"secretary\", \"soldier\")\n\nA1 \u003c- c(\"he\", \"son\", \"his\", \"him\", \"father\", \"man\", \"boy\", \"himself\",\n\"male\", \"brother\", \"sons\", \"fathers\", \"men\", \"boys\", \"males\",\n\"brothers\", \"uncle\", \"uncles\", \"nephew\", \"nephews\")\n\n## The same as:\n## mac_neg \u003c- query(googlenews, S_words = S1, A_words = A1, method = \"mac\")\nmac_neg \u003c- query(googlenews, S_words = S1, A_words = A1)\nmac_neg\n```\n\nThe returned object is an S3 object. Please refer to the help file of the method for the definition of all slots (in this case: `?mac`). For example, the magnitude of bias for each word in `S1` is available in the `P` slot.\n\n```{r mac_neg2}\nsort(mac_neg$P)\n```\n\n## Example: Relative Norm Distance\n\nThis analysis reproduces the analysis in Garg et al (2018), namely Figure 1.\n\n```{r}\nB1 \u003c- c(\"she\", \"daughter\", \"hers\", \"her\", \"mother\", \"woman\", \"girl\",\n\"herself\", \"female\", \"sister\", \"daughters\", \"mothers\", \"women\",\n\"girls\", \"females\", \"sisters\", \"aunt\", \"aunts\", \"niece\", \"nieces\"\n)\n\ngarg_f1 \u003c- query(googlenews, S_words = S1, A_words = A1, B_words = B1)\ngarg_f1\n```\n\nThe object can be plotted by the function `plot` to show the bias of each word in S. Words such as \"nurse\", \"midwife\" and \"librarian\" are more associated with female, as indicated by the positive relative norm distance.\n\n```{r rndplot, fig.height = 12}\nplot(garg_f1)\n```\n\nThe effect size is simply the sum of all relative norm distance values (Equation 3 in Garg et al. 2018). It is displayed simply by printing the object. You can also use the function `calculate_es` to obtain the numeric result.\n\nThe more positive effect size indicates that words in `S_words` are more associated with `B_words`. As the effect size is negative, it indicates that the concept of occupation is more associated with `A_words`, i.e. male.\n\n```{r}\ncalculate_es(garg_f1)\n```\n\n## Example: SemAxis\n\nThis analysis attempts to reproduce the analysis in An et al. (2018).\n\nYou may obtain the word2vec word vectors trained with Trump supporters Reddit from [here](https://github.com/ghdi6758/SemAxis). This package provides a tiny version of the data `small_reddit` for reproducing the analysis.\n\n```{r semxaxisplot}\nS2 \u003c- c(\"mexicans\", \"asians\", \"whites\", \"blacks\", \"latinos\")\nA2 \u003c- c(\"respect\")\nB2 \u003c- c(\"disrespect\")\nres \u003c- query(small_reddit, S_words = S2, A_words = A2, B_words = B2, method = \"semaxis\", l = 1)\nplot(res)\n```\n\n## Example: Embedding Coherence Test\n\nEmbedding Coherence Test (Dev \u0026 Phillips, 2019) is similar to SemAxis. The only significant difference is that no \"SemAxis\" is calculated (the difference between the average word vectors of `A_words` and `B_words`). Instead, it calculates two separate axes for `A_words` and `B_words`. Then it calculates the proximity of each word in `S_words` with the two axes. It is like doing two separate `mac`, but `ect` averages the word vectors of `A_words` / `B_words` first.\n\nIt is important to note that `P` is a 2-D matrix. Hence, the plot is 2-dimensional. Words above the equality line are more associated with `B_words` and vice versa.\n\n```{r ectplot}\nres \u003c- query(googlenews, S_words = S1, A_words = A1, B_words = B1, method = \"ect\")\nres$P\nplot(res)\n```\n\nEffect size can also be calculated. It is the Spearman Correlation Coefficient of the two rows in `P`. Higher value indicates more \"coherent\", i.e. less bias.\n\n```{r}\nres\n```\n\n\n## Example: Relative Negative Sentiment Bias\n\nThis analysis attempts to reproduce the analysis in Sweeney \u0026 Najafian (2019).\n\nPlease note that the datasets `glove_sweeney`, `bing_pos` and `bing_neg` are not included in the package. If you are interested in reproducing the analysis, the 3 datasets are available from [here](https://github.com/gesistsa/sweater/tree/master/tests/testdata).\n\n```{r}\nload(\"tests/testdata/bing_neg.rda\")\nload(\"tests/testdata/bing_pos.rda\")\nload(\"tests/testdata/glove_sweeney.rda\")\n\nS3 \u003c- c(\"swedish\", \"irish\", \"mexican\", \"chinese\", \"filipino\",\n        \"german\", \"english\", \"french\", \"norwegian\", \"american\",\n        \"indian\", \"dutch\", \"russian\", \"scottish\", \"italian\")\nsn \u003c- query(glove_sweeney, S_words = S3, A_words = bing_pos, B_words = bing_neg, method = \"rnsb\")\n```\n\nThe analysis shows that `indian`, `mexican`, and `russian` are more likely to be associated with negative sentiment.\n\n```{r rnsbplot}\nplot(sn)\n```\n\nThe effect size from the analysis is the Kullback–Leibler divergence of P from the uniform distribution. It is extremely close to the value reported in the original paper (0.6225).\n\n```{r}\nsn\n```\n\n## Support for Quanteda Dictionaries\n\n`rnsb` supports [quanteda](https://github.com/quanteda/quanteda) dictionaries as `S_words`. This support will be expanded to other methods later.\n\nThis analysis uses the data from [here](https://github.com/gesistsa/sweater/tree/master/tests/testdata).\n\nFor example, `newsmap_europe` is an abridged dictionary from the package newsmap (Watanabe, 2018). The dictionary contains keywords of European countries and has two levels: regional level (e.g. Eastern Europe) and country level (e.g. Germany).\n\n```{r}\nload(\"tests/testdata/newsmap_europe.rda\")\nload(\"tests/testdata/dictionary_demo.rda\")\n\nrequire(quanteda)\nnewsmap_europe\n```\n\nCountry-level analysis\n\n```{r rnsb2, fig.height = 10}\ncountry_level \u003c- rnsb(w = dictionary_demo, S_words = newsmap_europe, A_words = bing_pos, B_words = bing_neg, levels = 2)\nplot(country_level)\n```\n\nRegion-level analysis\n\n```{r rnsb3}\nregion_level \u003c- rnsb(w = dictionary_demo, S_words = newsmap_europe, A_words = bing_pos, B_words = bing_neg, levels = 1)\nplot(region_level)\n```\n\nComparison of the two effect sizes. Please note the much smaller effect size from region-level analysis. It reflects the evener distribution of P across regions than across countries.\n\n```{r}\ncalculate_es(country_level)\ncalculate_es(region_level)\n```\n\n## Example: Normalized Association Score\n\nNormalized Association Score (Caliskan et al., 2017) is similar to Relative Norm Distance above. It was used in Müller et al. (2023).\n\n```{r nasplot, fig.height = 12}\nS3 \u003c- c(\"janitor\", \"statistician\", \"midwife\", \"bailiff\", \"auctioneer\",\n\"photographer\", \"geologist\", \"shoemaker\", \"athlete\", \"cashier\",\n\"dancer\", \"housekeeper\", \"accountant\", \"physicist\", \"gardener\",\n\"dentist\", \"weaver\", \"blacksmith\", \"psychologist\", \"supervisor\",\n\"mathematician\", \"surveyor\", \"tailor\", \"designer\", \"economist\",\n\"mechanic\", \"laborer\", \"postmaster\", \"broker\", \"chemist\", \"librarian\",\n\"attendant\", \"clerical\", \"musician\", \"porter\", \"scientist\", \"carpenter\",\n\"sailor\", \"instructor\", \"sheriff\", \"pilot\", \"inspector\", \"mason\",\n\"baker\", \"administrator\", \"architect\", \"collector\", \"operator\",\n\"surgeon\", \"driver\", \"painter\", \"conductor\", \"nurse\", \"cook\",\n\"engineer\", \"retired\", \"sales\", \"lawyer\", \"clergy\", \"physician\",\n\"farmer\", \"clerk\", \"manager\", \"guard\", \"artist\", \"smith\", \"official\",\n\"police\", \"doctor\", \"professor\", \"student\", \"judge\", \"teacher\",\n\"author\", \"secretary\", \"soldier\")\nA3 \u003c- c(\"he\", \"son\", \"his\", \"him\", \"father\", \"man\", \"boy\", \"himself\",\n\"male\", \"brother\", \"sons\", \"fathers\", \"men\", \"boys\", \"males\",\n\"brothers\", \"uncle\", \"uncles\", \"nephew\", \"nephews\")\nB3 \u003c- c(\"she\", \"daughter\", \"hers\", \"her\", \"mother\", \"woman\", \"girl\",\n\"herself\", \"female\", \"sister\", \"daughters\", \"mothers\", \"women\",\n\"girls\", \"females\", \"sisters\", \"aunt\", \"aunts\", \"niece\", \"nieces\"\n)\n\nnas_f1 \u003c- query(googlenews, S_words= S3, A_words = A3, B_words = B3, method = \"nas\")\nplot(nas_f1)\n```\n\nThere is a very strong correlation between NAS and RND.\n\n```{r}\ncor.test(nas_f1$P, garg_f1$P)\n```\n\n## Example: Word Embedding Association Test\n\nThis example reproduces the detection of \"Math. vs Arts\" gender bias in Caliskan et al (2017).\n\n```{r maths}\ndata(glove_math) # a subset of the original GLoVE word vectors\n\nS4 \u003c- c(\"math\", \"algebra\", \"geometry\", \"calculus\", \"equations\", \"computation\", \"numbers\", \"addition\")\nT4 \u003c- c(\"poetry\", \"art\", \"dance\", \"literature\", \"novel\", \"symphony\", \"drama\", \"sculpture\")\nA4 \u003c- c(\"male\", \"man\", \"boy\", \"brother\", \"he\", \"him\", \"his\", \"son\")\nB4 \u003c- c(\"female\", \"woman\", \"girl\", \"sister\", \"she\", \"her\", \"hers\", \"daughter\")\nsw \u003c- query(glove_math, S4, T4, A4, B4)\n\n# extraction of effect size\nsw\n```\n\n## A note about the effect size\n\nBy default, the effect size from the function `weat_es` is adjusted by the pooled standard deviaion (see Page 2 of Caliskan et al. 2007). The standardized effect size can be interpreted the way as Cohen's d (Cohen, 1988).\n\nOne can also get the unstandardized version (aka. test statistic in the original paper):\n\n```{r}\n## weat_es\ncalculate_es(sw, standardize = FALSE)\n```\n\nThe original implementation assumes equal size of `S` and `T`. This assumption can be relaxed by pooling the standard deviaion with sample size adjustment. The function `weat_es` does it when `S` and `T` are of different length.\n\nAlso, the effect size can be converted to point-biserial correlation (mathematically equivalent to the Pearson's product moment correlation).\n\n```{r}\nweat_es(sw, r = TRUE)\n```\n\n## Exact test\n\nThe exact test described in Caliskan et al. (2017) is also available. But it takes a long time to calculate.\n\n```r\n## Don't do it. It takes a long time and is almost always significant.\nweat_exact(sw)\n```\n\nInstead, please use the resampling approximation of the exact test. The p-value is very close to the reported 0.018.\n\n```{r}\nweat_resampling(sw)\n```\n\n## How to get help\n\n* Read the documentation\n* Search for [issues](https://github.com/gesistsa/sweater/issues)\n\n## Contributing\n\nContributions in the form of feedback, comments, code, and bug report are welcome.\n\n* Fork the source code, modify, and issue a [pull request](https://docs.github.com/en/github/collaborating-with-issues-and-pull-requests/creating-a-pull-request-from-a-fork).\n* Issues, bug reports: [File a Github issue](https://github.com/gesistsa/sweater/issues).\n\n## Code of Conduct\n\nPlease note that the sweater project is released with a [Contributor Code of Conduct](https://contributor-covenant.org/version/2/0/CODE_OF_CONDUCT.html). By contributing to this project, you agree to abide by its terms.\n\n## References\n\n1. An, J., Kwak, H., \u0026 Ahn, Y. Y. (2018). SemAxis: A lightweight framework to characterize domain-specific word semantics beyond sentiment. arXiv preprint arXiv:1806.05521.\n2. Badilla, P., Bravo-Marquez, F., \u0026 Pérez, J. (2020). WEFE: The word embeddings fairness evaluation framework. In Proceedings of the 29 th Intern. Joint Conf. Artificial Intelligence.\n3. Brunet, M. E., Alkalay-Houlihan, C., Anderson, A., \u0026 Zemel, R. (2019, May). Understanding the origins of bias in word embeddings. In International Conference on Machine Learning (pp. 803-811). PMLR.\n4. Caliskan, Aylin, Joanna J. Bryson, and Arvind Narayanan. \"Semantics derived automatically from language corpora contain human-like biases.\" Science 356.6334 (2017): 183-186.\n5. Cohen, J. (1988), Statistical Power Analysis for the Behavioral Sciences, 2nd Edition. Hillsdale: Lawrence Erlbaum.\n6. Dev, S., \u0026 Phillips, J. (2019, April). Attenuating bias in word vectors. In The 22nd International Conference on Artificial Intelligence and Statistics (pp. 879-887). PMLR.\n7. Garg, N., Schiebinger, L., Jurafsky, D., \u0026 Zou, J. (2018). Word embeddings quantify 100 years of gender and ethnic stereotypes. Proceedings of the National Academy of Sciences, 115(16), E3635-E3644.\n8. Manzini, T., Lim, Y. C., Tsvetkov, Y., \u0026 Black, A. W. (2019). Black is to criminal as caucasian is to police: Detecting and removing multiclass bias in word embeddings. arXiv preprint arXiv:1904.04047.\n9. McGrath, R. E., \u0026 Meyer, G. J. (2006). When effect sizes disagree: the case of r and d. Psychological methods, 11(4), 386.\n10. Müller, P., Chan, C. H., Ludwig, K., Freudenthaler, R., \u0026 Wessler, H. (2023). Differential Racism in the News: Using Semi-Supervised Machine Learning to Distinguish Explicit and Implicit Stigmatization of Ethnic and Religious Groups in Journalistic Discourse. Political Communication, 1-19.\n11. Rosenthal, R. (1991), Meta-Analytic Procedures for Social Research. Newbury Park: Sage\n12. Sweeney, C., \u0026 Najafian, M. (2019, July). A transparent framework for evaluating unintended demographic bias in word embeddings. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (pp. 1662-1667).\n13. Watanabe, K. (2018). Newsmap: A semi-supervised approach to geographical news classification. Digital Journalism, 6(3), 294-309.\n\n---\n\n[^legacy]: Please use the `query` function. These functions are kept for backward compatibility.\n","funding_links":[],"categories":[],"sub_categories":[],"project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fgesistsa%2Fsweater","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fgesistsa%2Fsweater","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fgesistsa%2Fsweater/lists"}