md_gender_bias

Les références:

• Code
• Visage câlin

mots_genrés

Utilisez la commande suivante pour charger cet ensemble de données dans TFDS :

ds = tfds.load('huggingface:md_gender_bias/gendered_words')

• Description :

Machine learning models are trained to find patterns in data.
NLP models can inadvertently learn socially undesirable patterns when training on gender biased text.
In this work, we propose a general framework that decomposes gender bias in text along several pragmatic and semantic dimensions:
bias from the gender of the person being spoken about, bias from the gender of the person being spoken to, and bias from the gender of the speaker.
Using this fine-grained framework, we automatically annotate eight large scale datasets with gender information.
In addition, we collect a novel, crowdsourced evaluation benchmark of utterance-level gender rewrites.
Distinguishing between gender bias along multiple dimensions is important, as it enables us to train finer-grained gender bias classifiers.
We show our classifiers prove valuable for a variety of important applications, such as controlling for gender bias in generative models,
detecting gender bias in arbitrary text, and shed light on offensive language in terms of genderedness.

• Licence : Licence MIT
• Version : 1.0.0
• Divisions :

• Caractéristiques :

{
    "word_masculine": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "word_feminine": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

nom_genres

Utilisez la commande suivante pour charger cet ensemble de données dans TFDS :

ds = tfds.load('huggingface:md_gender_bias/name_genders')

• Description :

Machine learning models are trained to find patterns in data.
NLP models can inadvertently learn socially undesirable patterns when training on gender biased text.
In this work, we propose a general framework that decomposes gender bias in text along several pragmatic and semantic dimensions:
bias from the gender of the person being spoken about, bias from the gender of the person being spoken to, and bias from the gender of the speaker.
Using this fine-grained framework, we automatically annotate eight large scale datasets with gender information.
In addition, we collect a novel, crowdsourced evaluation benchmark of utterance-level gender rewrites.
Distinguishing between gender bias along multiple dimensions is important, as it enables us to train finer-grained gender bias classifiers.
We show our classifiers prove valuable for a variety of important applications, such as controlling for gender bias in generative models,
detecting gender bias in arbitrary text, and shed light on offensive language in terms of genderedness.

• Licence : Licence MIT
• Version : 1.0.0
• Divisions :

• Caractéristiques :

{
    "name": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "assigned_gender": {
        "num_classes": 2,
        "names": [
            "M",
            "F"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    },
    "count": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

nouvelles données

Utilisez la commande suivante pour charger cet ensemble de données dans TFDS :

ds = tfds.load('huggingface:md_gender_bias/new_data')

• Description :

Machine learning models are trained to find patterns in data.
NLP models can inadvertently learn socially undesirable patterns when training on gender biased text.
In this work, we propose a general framework that decomposes gender bias in text along several pragmatic and semantic dimensions:
bias from the gender of the person being spoken about, bias from the gender of the person being spoken to, and bias from the gender of the speaker.
Using this fine-grained framework, we automatically annotate eight large scale datasets with gender information.
In addition, we collect a novel, crowdsourced evaluation benchmark of utterance-level gender rewrites.
Distinguishing between gender bias along multiple dimensions is important, as it enables us to train finer-grained gender bias classifiers.
We show our classifiers prove valuable for a variety of important applications, such as controlling for gender bias in generative models,
detecting gender bias in arbitrary text, and shed light on offensive language in terms of genderedness.

• Licence : Licence MIT
• Version : 1.0.0
• Divisions :

• Caractéristiques :

{
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "original": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "labels": [
        {
            "num_classes": 6,
            "names": [
                "ABOUT:female",
                "ABOUT:male",
                "PARTNER:female",
                "PARTNER:male",
                "SELF:female",
                "SELF:male"
            ],
            "names_file": null,
            "id": null,
            "_type": "ClassLabel"
        }
    ],
    "class_type": {
        "num_classes": 3,
        "names": [
            "about",
            "partner",
            "self"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    },
    "turker_gender": {
        "num_classes": 5,
        "names": [
            "man",
            "woman",
            "nonbinary",
            "prefer not to say",
            "no answer"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    },
    "episode_done": {
        "dtype": "bool_",
        "id": null,
        "_type": "Value"
    },
    "confidence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

funpédia

Utilisez la commande suivante pour charger cet ensemble de données dans TFDS :

ds = tfds.load('huggingface:md_gender_bias/funpedia')

• Description :

Machine learning models are trained to find patterns in data.
NLP models can inadvertently learn socially undesirable patterns when training on gender biased text.
In this work, we propose a general framework that decomposes gender bias in text along several pragmatic and semantic dimensions:
bias from the gender of the person being spoken about, bias from the gender of the person being spoken to, and bias from the gender of the speaker.
Using this fine-grained framework, we automatically annotate eight large scale datasets with gender information.
In addition, we collect a novel, crowdsourced evaluation benchmark of utterance-level gender rewrites.
Distinguishing between gender bias along multiple dimensions is important, as it enables us to train finer-grained gender bias classifiers.
We show our classifiers prove valuable for a variety of important applications, such as controlling for gender bias in generative models,
detecting gender bias in arbitrary text, and shed light on offensive language in terms of genderedness.

• Licence : Licence MIT
• Version : 1.0.0
• Divisions :

• Caractéristiques :

{
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "persona": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gender": {
        "num_classes": 3,
        "names": [
            "gender-neutral",
            "female",
            "male"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    }
}

image_chat

Utilisez la commande suivante pour charger cet ensemble de données dans TFDS :

ds = tfds.load('huggingface:md_gender_bias/image_chat')

• Description :

Machine learning models are trained to find patterns in data.
NLP models can inadvertently learn socially undesirable patterns when training on gender biased text.
In this work, we propose a general framework that decomposes gender bias in text along several pragmatic and semantic dimensions:
bias from the gender of the person being spoken about, bias from the gender of the person being spoken to, and bias from the gender of the speaker.
Using this fine-grained framework, we automatically annotate eight large scale datasets with gender information.
In addition, we collect a novel, crowdsourced evaluation benchmark of utterance-level gender rewrites.
Distinguishing between gender bias along multiple dimensions is important, as it enables us to train finer-grained gender bias classifiers.
We show our classifiers prove valuable for a variety of important applications, such as controlling for gender bias in generative models,
detecting gender bias in arbitrary text, and shed light on offensive language in terms of genderedness.

• Licence : Licence MIT
• Version : 1.0.0
• Divisions :

• Caractéristiques :

{
    "caption": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "male": {
        "dtype": "bool_",
        "id": null,
        "_type": "Value"
    },
    "female": {
        "dtype": "bool_",
        "id": null,
        "_type": "Value"
    }
}

magicien

Utilisez la commande suivante pour charger cet ensemble de données dans TFDS :

ds = tfds.load('huggingface:md_gender_bias/wizard')

• Description :

Machine learning models are trained to find patterns in data.
NLP models can inadvertently learn socially undesirable patterns when training on gender biased text.
In this work, we propose a general framework that decomposes gender bias in text along several pragmatic and semantic dimensions:
bias from the gender of the person being spoken about, bias from the gender of the person being spoken to, and bias from the gender of the speaker.
Using this fine-grained framework, we automatically annotate eight large scale datasets with gender information.
In addition, we collect a novel, crowdsourced evaluation benchmark of utterance-level gender rewrites.
Distinguishing between gender bias along multiple dimensions is important, as it enables us to train finer-grained gender bias classifiers.
We show our classifiers prove valuable for a variety of important applications, such as controlling for gender bias in generative models,
detecting gender bias in arbitrary text, and shed light on offensive language in terms of genderedness.

• Licence : Licence MIT
• Version : 1.0.0
• Divisions :

• Caractéristiques :

{
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "chosen_topic": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gender": {
        "num_classes": 3,
        "names": [
            "gender-neutral",
            "female",
            "male"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    }
}

convai2_inféré

Utilisez la commande suivante pour charger cet ensemble de données dans TFDS :

ds = tfds.load('huggingface:md_gender_bias/convai2_inferred')

• Description :

Machine learning models are trained to find patterns in data.
NLP models can inadvertently learn socially undesirable patterns when training on gender biased text.
In this work, we propose a general framework that decomposes gender bias in text along several pragmatic and semantic dimensions:
bias from the gender of the person being spoken about, bias from the gender of the person being spoken to, and bias from the gender of the speaker.
Using this fine-grained framework, we automatically annotate eight large scale datasets with gender information.
In addition, we collect a novel, crowdsourced evaluation benchmark of utterance-level gender rewrites.
Distinguishing between gender bias along multiple dimensions is important, as it enables us to train finer-grained gender bias classifiers.
We show our classifiers prove valuable for a variety of important applications, such as controlling for gender bias in generative models,
detecting gender bias in arbitrary text, and shed light on offensive language in terms of genderedness.

• Licence : Licence MIT
• Version : 1.0.0
• Divisions :

• Caractéristiques :

{
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "binary_label": {
        "num_classes": 2,
        "names": [
            "ABOUT:female",
            "ABOUT:male"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    },
    "binary_score": {
        "dtype": "float32",
        "id": null,
        "_type": "Value"
    },
    "ternary_label": {
        "num_classes": 3,
        "names": [
            "ABOUT:female",
            "ABOUT:male",
            "ABOUT:gender-neutral"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    },
    "ternary_score": {
        "dtype": "float32",
        "id": null,
        "_type": "Value"
    }
}

light_inféré

Utilisez la commande suivante pour charger cet ensemble de données dans TFDS :

ds = tfds.load('huggingface:md_gender_bias/light_inferred')

• Description :

Machine learning models are trained to find patterns in data.
NLP models can inadvertently learn socially undesirable patterns when training on gender biased text.
In this work, we propose a general framework that decomposes gender bias in text along several pragmatic and semantic dimensions:
bias from the gender of the person being spoken about, bias from the gender of the person being spoken to, and bias from the gender of the speaker.
Using this fine-grained framework, we automatically annotate eight large scale datasets with gender information.
In addition, we collect a novel, crowdsourced evaluation benchmark of utterance-level gender rewrites.
Distinguishing between gender bias along multiple dimensions is important, as it enables us to train finer-grained gender bias classifiers.
We show our classifiers prove valuable for a variety of important applications, such as controlling for gender bias in generative models,
detecting gender bias in arbitrary text, and shed light on offensive language in terms of genderedness.

• Licence : Licence MIT
• Version : 1.0.0
• Divisions :

• Caractéristiques :

{
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "binary_label": {
        "num_classes": 2,
        "names": [
            "ABOUT:female",
            "ABOUT:male"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    },
    "binary_score": {
        "dtype": "float32",
        "id": null,
        "_type": "Value"
    },
    "ternary_label": {
        "num_classes": 3,
        "names": [
            "ABOUT:female",
            "ABOUT:male",
            "ABOUT:gender-neutral"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    },
    "ternary_score": {
        "dtype": "float32",
        "id": null,
        "_type": "Value"
    }
}

opensubtitles_inferred

Utilisez la commande suivante pour charger cet ensemble de données dans TFDS :

ds = tfds.load('huggingface:md_gender_bias/opensubtitles_inferred')

• Description :

Machine learning models are trained to find patterns in data.
NLP models can inadvertently learn socially undesirable patterns when training on gender biased text.
In this work, we propose a general framework that decomposes gender bias in text along several pragmatic and semantic dimensions:
bias from the gender of the person being spoken about, bias from the gender of the person being spoken to, and bias from the gender of the speaker.
Using this fine-grained framework, we automatically annotate eight large scale datasets with gender information.
In addition, we collect a novel, crowdsourced evaluation benchmark of utterance-level gender rewrites.
Distinguishing between gender bias along multiple dimensions is important, as it enables us to train finer-grained gender bias classifiers.
We show our classifiers prove valuable for a variety of important applications, such as controlling for gender bias in generative models,
detecting gender bias in arbitrary text, and shed light on offensive language in terms of genderedness.

• Licence : Licence MIT
• Version : 1.0.0
• Divisions :

• Caractéristiques :

{
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "binary_label": {
        "num_classes": 2,
        "names": [
            "ABOUT:female",
            "ABOUT:male"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    },
    "binary_score": {
        "dtype": "float32",
        "id": null,
        "_type": "Value"
    },
    "ternary_label": {
        "num_classes": 3,
        "names": [
            "ABOUT:female",
            "ABOUT:male",
            "ABOUT:gender-neutral"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    },
    "ternary_score": {
        "dtype": "float32",
        "id": null,
        "_type": "Value"
    }
}

yelp_inferred

Utilisez la commande suivante pour charger cet ensemble de données dans TFDS :

ds = tfds.load('huggingface:md_gender_bias/yelp_inferred')

• Description :

Machine learning models are trained to find patterns in data.
NLP models can inadvertently learn socially undesirable patterns when training on gender biased text.
In this work, we propose a general framework that decomposes gender bias in text along several pragmatic and semantic dimensions:
bias from the gender of the person being spoken about, bias from the gender of the person being spoken to, and bias from the gender of the speaker.
Using this fine-grained framework, we automatically annotate eight large scale datasets with gender information.
In addition, we collect a novel, crowdsourced evaluation benchmark of utterance-level gender rewrites.
Distinguishing between gender bias along multiple dimensions is important, as it enables us to train finer-grained gender bias classifiers.
We show our classifiers prove valuable for a variety of important applications, such as controlling for gender bias in generative models,
detecting gender bias in arbitrary text, and shed light on offensive language in terms of genderedness.

• Licence : Licence MIT
• Version : 1.0.0
• Divisions :

• Caractéristiques :

{
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "binary_label": {
        "num_classes": 2,
        "names": [
            "ABOUT:female",
            "ABOUT:male"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    },
    "binary_score": {
        "dtype": "float32",
        "id": null,
        "_type": "Value"
    }
}