md_gender_bias

参考文献:

• コード
• ハグフェイス

性的な言葉

次のコマンドを使用して、このデータセットを TFDS にロードします。

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

• 説明：

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.

• ライセンス: MITライセンス
• バージョン: 1.0.0
• 分割:

• 特徴：

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

名前_性別

次のコマンドを使用して、このデータセットを TFDS にロードします。

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

• 説明：

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.

• ライセンス: MITライセンス
• バージョン: 1.0.0
• 分割:

• 特徴：

{
    "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"
    }
}

新しいデータ

次のコマンドを使用して、このデータセットを TFDS にロードします。

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

• 説明：

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.

• ライセンス: MITライセンス
• バージョン: 1.0.0
• 分割:

• 特徴：

{
    "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"
    }
}

ファンペディア

次のコマンドを使用して、このデータセットを TFDS にロードします。

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

• 説明：

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.

• ライセンス: MITライセンス
• バージョン: 1.0.0
• 分割:

• 特徴：

{
    "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"
    }
}

画像チャット

次のコマンドを使用して、このデータセットを TFDS にロードします。

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

• 説明：

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.

• ライセンス: MITライセンス
• バージョン: 1.0.0
• 分割:

• 特徴：

{
    "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"
    }
}

ウィザード

次のコマンドを使用して、このデータセットを TFDS にロードします。

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

• 説明：

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.

• ライセンス: MITライセンス
• バージョン: 1.0.0
• 分割:

• 特徴：

{
    "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_inferred

次のコマンドを使用して、このデータセットを TFDS にロードします。

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

• 説明：

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.

• ライセンス: MITライセンス
• バージョン: 1.0.0
• 分割:

• 特徴：

{
    "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_inferred

次のコマンドを使用して、このデータセットを TFDS にロードします。

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

• 説明：

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.

• ライセンス: MITライセンス
• バージョン: 1.0.0
• 分割:

• 特徴：

{
    "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

次のコマンドを使用して、このデータセットを TFDS にロードします。

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

• 説明：

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.

• ライセンス: MITライセンス
• バージョン: 1.0.0
• 分割:

• 特徴：

{
    "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

次のコマンドを使用して、このデータセットを TFDS にロードします。

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

• 説明：

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.

• ライセンス: MITライセンス
• バージョン: 1.0.0
• 分割:

• 特徴：

{
    "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"
    }
}