Ethics regarding social bias has recently thrown striking issues in natural language processing. Especially for gender-related topics, the need for a system that reduces the model bias has grown in areas such as image captioning, content recommendation, and automated employment. However, detection and evaluation of gender bias in the machine translation systems are not yet thoroughly investigated, for the task being cross-lingual and challenging to define. In this paper, we propose a scheme for making up a test set that evaluates the gender bias in a machine translation system, with Korean, a language with gender-neutral pronouns. Three word/phrase sets are primarily constructed, each incorporating positive/negative expressions or occupations; all the terms are gender-independent or at least not biased to one side severely. Then, additional sentence lists are constructed concerning formality of the pronouns and politeness of the sentences. With the generated sentence set of size 4,236 in total, we evaluate gender bias in conventional machine translation systems utilizing the proposed measure, which is termed here as translation gender bias index (TGBI). The corpus and the code for evaluation is available on-line.
Machine translation systems with inadequate document understanding can make errors when translating dropped or neutral pronouns into languages with gendered pronouns (e.g., English). Predicting the underlying gender of these pronouns is difficult since it is not marked textually and must instead be inferred from coreferent mentions in the context. We propose a novel cross-lingual pivoting technique for automatically producing high-quality gender labels, and show that this data can be used to fine-tune a BERT classifier with 92% F1 for Spanish dropped feminine pronouns, compared with 30-51% for…
University of Groningen（格罗宁根大学）；University of Malta（马耳他大学）
Contextualized word embeddings have been replacing standard embeddings as the representational knowledge source of choice in NLP systems. Since a variety of biases have previously been found in standard word embeddings, it is crucial to assess biases encoded in their replacements as well. Focusing on BERT (Devlin et al., 2018), we measure gender bias by studying associations between gender-denoting target words and names of professions in English and German, comparing the findings with real-world workforce statistics. We mitigate bias by fine-tuning BERT on the GAP corpus (Webster et al., 2018…
University of Pennsylvania（宾夕法尼亚大学）；Moscow Institute of Physics and Technology（莫斯科物理技术学院）；Kitsong；PJSC Sberbank（俄罗斯联邦储蓄银行）；Jet Infosystems；Sofia University（索非亚大学）
Pronoun resolution is part of coreference resolution, the task of pairing an expression to its referring entity. This is an important task for natural language understanding and a necessary component of machine translation systems, chat bots and assistants. Neural machine learning systems perform far from ideally in this task, reaching as low as 73% F1 scores on modern benchmark datasets. Moreover, they tend to perform better for masculine pronouns than for feminine ones. Thus, the problem is both challenging and important for NLP researchers and practitioners. In this project, we describe our…
George Washington University（乔治华盛顿大学）；University of Washington（华盛顿大学）
Does the grammatical gender of a language interfere when measuring the semantic gender information captured by its word embeddings? A number of anomalous gender bias measurements in the embeddings of gendered languages suggest this possibility. We demonstrate that word embeddings learn the association between a noun and its grammatical gender in grammatically gendered languages, which can skew social gender bias measurements. Consequently, word embedding post-processing methods are introduced to quantify, disentangle, and evaluate grammatical gender signals. The evaluation is performed on five…
Machine learning models allow us to compare languages by showing how hard a task in each language might be to learn and perform well on. Following this line of investigation, we explore what makes a language "hard to pronounce" by modelling the task of grapheme-to-phoneme (g2p) transliteration. By training a character-level transformer model on this task across 22 languages and measuring the model's proficiency against its grapheme and phoneme inventories, we show that certain characteristics emerge that separate easier and harder languages with respect to learning to pronounce. Namely the com…
ADAPT Centre, Trinity College Dublin（都柏林圣三一学院ADAPT中心）；ADAPT Centre, Dublin City University（都柏林城市大学ADAPT中心）；Microsoft（微软）；Tilburg University（蒂尔堡大学）
Gender bias is a frequent occurrence in NLP-based applications, especially pronounced in gender-inflected languages. Bias can appear through associations of certain adjectives and animate nouns with the natural gender of referents, but also due to unbalanced grammatical gender frequencies of inflected words. This type of bias becomes more evident in generating conversational utterances where gender is not specified within the sentence, because most current NLP applications still work on a sentence-level context. As a step towards more inclusive NLP, this paper proposes an automatic and general…
Pronouns are important determinants of a text's meaning but difficult to translate. This is because pronoun choice can depend on entities described in previous sentences, and in some languages pronouns may be dropped when the referent is inferrable from the context. These issues can lead Neural Machine Translation (NMT) systems to make critical errors on pronouns that impair intelligibility and even reinforce gender bias. We investigate the severity of this pronoun issue, showing that (1) in some domains, pronoun choice can account for more than half of a NMT systems' errors, and (2) pronouns …
Transformer based models are the modern work horses for neural machine translation (NMT), reaching state of the art across several benchmarks. Despite their impressive accuracy, we observe a systemic and rudimentary class of errors made by transformer based models with regards to translating from a language that doesn't mark gender on nouns into others that do. We find that even when the surrounding context provides unambiguous evidence of the appropriate grammatical gender marking, no transformer based model we tested was able to accurately gender occupation nouns systematically. We release a…
Gender-neutral pronouns have recently been introduced in many languages to a) include non-binary people and b) as a generic singular. Recent results from psycholinguistics suggest that gender-neutral pronouns (in Swedish) are not associated with human processing difficulties. This, we show, is in sharp contrast with automated processing. We show that gender-neutral pronouns in Danish, English, and Swedish are associated with higher perplexity, more dispersed attention patterns, and worse downstream performance. We argue that such conservativity in language models may limit widespread adoption …
Coreference resolution is an important task for natural language understanding, and the resolution of ambiguous pronouns a longstanding challenge. Nonetheless, existing corpora do not capture ambiguous pronouns in sufficient volume or diversity to accurately indicate the practical utility of models. Furthermore, we find gender bias in existing corpora and systems favoring masculine entities. To address this, we present and release GAP, a gender-balanced labeled corpus of 8,908 ambiguous pronoun-name pairs sampled to provide diverse coverage of challenges posed by real-world text. We explore a …
This paper presents a strong set of results for resolving gendered ambiguous pronouns on the Gendered Ambiguous Pronouns shared task. The model presented here draws upon the strengths of state-of-the-art language and coreference resolution models, and introduces a novel evidence-based deep learning architecture. Injecting evidence from the coreference models compliments the base architecture, and analysis shows that the model is not hindered by their weaknesses, specifically gender bias. The modularity and simplicity of the architecture make it very easy to extend for further improvement and a…
Università Luigi Bocconi（博科尼大学）；University of South Carolina（南卡罗来纳大学）
The world of pronouns is changing. From a closed class of words with few members to a much more open set of terms to reflect identities. However, Natural Language Processing (NLP) is barely reflecting this linguistic shift, even though recent work outlined the harms of gender-exclusive language technology. Particularly problematic is the current modeling 3rd person pronouns, as it largely ignores various phenomena like neopronouns, i.e., pronoun sets that are novel and not (yet) widely established. This omission contributes to the discrimination of marginalized and underrepresented groups, e.g…
A Winograd schema is a pair of sentences that differ in a single word and that contain an ambiguous pronoun whose referent is different in the two sentences and requires the use of commonsense knowledge or world knowledge to disambiguate. This paper discusses how Winograd schemas and other sentence pairs could be used as challenges for machine translation using distinctions between pronouns, such as gender, that appear in the target language but not in the source.
University of Pennsylvania（宾夕法尼亚大学）；Yale University（耶鲁大学）
Gender bias has been found in existing coreference resolvers. In order to eliminate gender bias, a gender-balanced dataset Gendered Ambiguous Pronouns (GAP) has been released and the best baseline model achieves only 66.9% F1. Bidirectional Encoder Representations from Transformers (BERT) has broken several NLP task records and can be used on GAP dataset. However, fine-tune BERT on a specific task is computationally expensive. In this paper, we propose an end-to-end resolver by combining pre-trained BERT with Relational Graph Convolutional Network (R-GCN). R-GCN is used for digesting structura…
Neural Machine Translation (NMT) systems are typically evaluated using automated metrics that assess the agreement between generated translations and ground truth candidates. To improve systems with respect to these metrics, NLP researchers employ a variety of heuristic techniques, including searching for the conditional mode (vs. sampling) and incorporating various training heuristics (e.g., label smoothing). While search strategies significantly improve BLEU score, they yield deterministic outputs that lack the diversity of human translations. Moreover, search tends to bias the distribution …
Federal University of Rio Grande do Sul（南里奥格兰德州联邦大学）
Recently there has been a growing concern about machine bias, where trained statistical models grow to reflect controversial societal asymmetries, such as gender or racial bias. A significant number of AI tools have recently been suggested to be harmfully biased towards some minority, with reports of racist criminal behavior predictors, Iphone X failing to differentiate between two Asian people and Google photos' mistakenly classifying black people as gorillas. Although a systematic study of such biases can be difficult, we believe that automated translation tools can be exploited through gend…
Manipal Institute of Technology（马尼帕尔理工学院）；MAHE（马尼帕尔高等教育学院）
With language models being deployed increasingly in the real world, it is essential to address the issue of the fairness of their outputs. The word embedding representations of these language models often implicitly draw unwanted associations that form a social bias within the model. The nature of gendered languages like Hindi, poses an additional problem to the quantification and mitigation of bias, owing to the change in the form of the words in the sentence, based on the gender of the subject. Additionally, there is sparse work done in the realm of measuring and debiasing systems for Indic …
Institute of Formal and Applied Linguistics, Faculty of Mathematics and Physics（数学与物理学院形式与应用语言学研究所）；Charles University（查理大学）
The representations in large language models contain multiple types of gender information. We focus on two types of such signals in English texts: factual gender information, which is a grammatical or semantic property, and gender bias, which is the correlation between a word and specific gender. We can disentangle the model's embeddings and identify components encoding both types of information with probing. We aim to diminish the stereotypical bias in the representations while preserving the factual gender signal. Our filtering method shows that it is possible to decrease the bias of gender-…
Probing Social Identity Bias in Chinese LLMs with Gendered Pronouns and Social Groups
Department of Electronics, Information and Bioengineering, Politecnico di Milano（电子、信息与生物工程系，米兰理工大学）；University of Science and Technology of China（中国科学技术大学）
The resolution of ambiguous pronouns is a longstanding challenge in Natural Language Understanding. Recent studies have suggested gender bias among state-of-the-art coreference resolution systems. As an example, Google AI Language team recently released a gender-balanced dataset and showed that performance of these coreference resolvers is significantly limited on the dataset. In this paper, we propose an extractive question answering (QA) formulation of pronoun resolution task that overcomes this limitation and shows much lower gender bias (0.99) on their dataset. This system uses fine-tuned …
ChatGPT Perpetuates Gender Bias in Machine Translation and Ignores Non-Gendered Pronouns: Findings across Bengali and Five other Low-Resource Languages
We present an empirical study of gender bias in coreference resolution systems. We first introduce a novel, Winograd schema-style set of minimal pair sentences that differ only by pronoun gender. With these "Winogender schemas," we evaluate and confirm systematic gender bias in three publicly-available coreference resolution systems, and correlate this bias with real-world and textual gender statistics.
Many text corpora exhibit socially problematic biases, which can be propagated or amplified in the models trained on such data. For example, doctor cooccurs more frequently with male pronouns than female pronouns. In this study we (i) propose a metric to measure gender bias; (ii) measure bias in a text corpus and the text generated from a recurrent neural network language model trained on the text corpus; (iii) propose a regularization loss term for the language model that minimizes the projection of encoder-trained embeddings onto an embedding subspace that encodes gender; (iv) finally, evalu…
University of Oxford（牛津大学）；TU Wien（维也纳工业大学）；University College London（伦敦大学学院）
Bias-measuring datasets play a critical role in detecting biased behavior of language models and in evaluating progress of bias mitigation methods. In this work, we focus on evaluating gender bias through coreference resolution, where previous datasets are either hand-crafted or fail to reliably measure an explicitly defined bias. To overcome these shortcomings, we propose a novel method to collect diverse, natural, and minimally distant text pairs via counterfactual generation, and construct Counter-GAP, an annotated dataset consisting of 4008 instances grouped into 1002 quadruples. We furthe…
University of Bergen（卑尔根大学）；Bocconi University（博科尼大学）
Scandinavian countries are perceived as role-models when it comes to gender equality. With the advent of pre-trained language models and their widespread usage, we investigate to what extent gender-based harmful and toxic content exist in selected Scandinavian language models. We examine nine models, covering Danish, Swedish, and Norwegian, by manually creating template-based sentences and probing the models for completion. We evaluate the completions using two methods for measuring harmful and toxic completions and provide a thorough analysis of the results. We show that Scandinavian pre-trai…
This study analyzes the usage of Japanese gendered language on Twitter. Starting from a collection of 408 million Japanese tweets from 2015 till 2019 and an additional sample of 2355 manually classified Twitter accounts timelines into gender and categories (politicians, musicians, etc). A large scale textual analysis is performed on this corpus to identify and examine sentence-final particles (SFPs) and first-person pronouns appearing in the texts. It turns out that gendered language is in fact used also on Twitter, in about 6% of the tweets, and that the prescriptive classification into "male…
Contextual word embeddings such as BERT have achieved state of the art performance in numerous NLP tasks. Since they are optimized to capture the statistical properties of training data, they tend to pick up on and amplify social stereotypes present in the data as well. In this study, we (1)~propose a template-based method to quantify bias in BERT; (2)~show that this method obtains more consistent results in capturing social biases than the traditional cosine based method; and (3)~conduct a case study, evaluating gender bias in a downstream task of Gender Pronoun Resolution. Although our case …
University of Copenhagen（哥本哈根大学）；Alexandra Instituttet（亚历山德拉研究所）；Google Research（谷歌研究院）
The one-sided focus on English in previous studies of gender bias in NLP misses out on opportunities in other languages: English challenge datasets such as GAP and WinoGender highlight model preferences that are "hallucinatory", e.g., disambiguating gender-ambiguous occurrences of 'doctor' as male doctors. We show that for languages with type B reflexivization, e.g., Swedish and Russian, we can construct multi-task challenge datasets for detecting gender bias that lead to unambiguously wrong model predictions: In these languages, the direct translation of 'the doctor removed his mask' is not a…
Gender discrimination in hiring is a pertinent and persistent bias in society, and a common motivating example for exploring bias in NLP. However, the manifestation of gendered language in application materials has received limited attention. This paper investigates the framing of skills and background in CVs of self-identified men and women. We introduce a data set of 1.8K authentic, English-language, CVs from the US, covering 16 occupations, allowing us to partially control for the confound occupation-specific gender base rates. We find that (1) women use more verbs evoking impressions of lo…
Masked language models pick up gender biases during pre-training. Such biases are usually attributed to a certain model architecture and its pre-training corpora, with the implicit assumption that other variations in the pre-training process, such as the choices of the random seed or the stopping point, have no effect on the biases measured. However, we show that severe fluctuations exist at the fundamental level of individual templates, invalidating the assumption. Further against the intuition of how humans acquire biases, these fluctuations are not correlated with the certainty of the predi…
Insight SFI Research Centre for Data Analytics（Insight SFI数据分析研究中心）；School of ICS（信息与计算科学学院）；University College Dublin（都柏林大学学院）
This paper presents a new method for automatically detecting words with lexical gender in large-scale language datasets. Currently, the evaluation of gender bias in natural language processing relies on manually compiled lexicons of gendered expressions, such as pronouns ('he', 'she', etc.) and nouns with lexical gender ('mother', 'boyfriend', 'policewoman', etc.). However, manual compilation of such lists can lead to static information if they are not periodically updated and often involve value judgments by individual annotators and researchers. Moreover, terms not included in the list fall …
