Towards Robust and Privacy-preserving Text Representations

By Yitong Li, Timothy Baldwin, Trevor Cohn. 2018

181 citations recorded by the source corpus.

Written text often provides sufficient clues to identify the author, their gender, age, and other important attributes. Consequently, the authorship of training and evaluation corpora can have unforeseen impacts, including differing model performance for different user groups, as well as privacy implications. In this paper, we propose an approach to explicitly obscure important author characteristics at training time, such that representations learned are invariant to these attributes. Evaluating on two tasks, we show that this leads to increased privacy in the learned representations, as well as more robust models to varying evaluation conditions, including out-of-domain corpora.

Subfields: Artificial Intelligence

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Mitigating Gender Bias in Natural Language Processing: Literature Review (2019, background): “Other important aspects such as model/data trans-parency (Mitchell et al., 2019; Bender and Fried-man, 2018) and privacy preservation (Reddy and Knight, 2016; Elazar and Goldberg, 2018; Li et al., 2018) are also not covered in this literature survey.”
Adversarial Removal of Demographic Attributes from Text Data (2018, background): “Recent work on creating private representation in the text domain (Li et al., 2018) share our motivation of removing unintended demographic attributes from the learned representation using adversarial training.”
Predictive Biases in Natural Language Processing Models: A Conceptual Framework and Overview (2019, background): “Li et al. (2018) introduced a model-based counter measure.”
Demoting Racial Bias in Hate Speech Detection (2020, background): “…for successfully adapting models to learn representations that are invariant to undesired attributes, such as demographics and topics, though they rarely disentangle attributes completely (Li et al., 2018; Elazar and Goldberg, 2018; Kumar et al., 2019; Lample et al., 2019; Landeiro et al., 2019).”
Information Leakage in Embedding Models (2020, methodology): “The same approach as been applied in NLP models to remove sensitive attributes from the representation for privacy or fairness concern [9, 18, 38].”
Gender Bias in Machine Translation (2021, background): “Accordingly, MT models could be brought to avoid binary forms and move toward gender-unspecified solutions, for example, adversarial networks including a discriminator that classifies speaker’s linguistic expression of gender (masculine or feminine) could be employed to ‘‘neutralize’’ speakerrelated”
Overlearning Reveals Sensitive Attributes (2019, methodology): “Techniques include adversarial training (Edwards & Storkey, 2016), which has been applied to census and health records (Xie et al., 2017), text (Li et al., 2018; Coavoux et al., 2018; Elazar & Goldberg, 2018), images (Hamm, 2017) and sensor data of wearables (Iwasawa et al., 2016).”
You Don’t Know My Favorite Color: Preventing Dialogue Representations from Revealing Speakers’ Private Personas (2022, background): “Adversarial training is widely used to protect sensitive features in natural language processing (Elazar and Goldberg, 2018; Coavoux et al., 2018; Li et al., 2018).”
The Importance of Modeling Social Factors of Language: Theory and Practice (2021, background): “, 2017), or via neural models for multi-task learning (Benton et al., 2017; Li et al., 2018).”
Topics to Avoid: Demoting Latent Confounds in Text Classification (2019, background): “, 2017), either to remove sensitive attributes such as demographic information (Li et al., 2018; Elazar and Goldberg, 2018; Beutel et al., 2017), or to understand costumer behavior for social science applications (Pryzant et al.”
Towards Identifying Social Bias in Dialog Systems: Framework, Dataset, and Benchmark (2022, background): “…of tasks, including identifying suspicious correlations (e.g., be-tween gender and toxicity labels) learned by embeddings or pre-trained models (Li et al., 2018; Zhao et al., 2019; Basta et al., 2019; Zhang et al., 2020; Nadeem et al., 2021; Zhou et al., 2021b; Du et al., 2021; Smith et al.,…”
Pile of Law: Learning Responsible Data Filtering from the Law and a 256GB Open-Source Legal Dataset (2022, background): “Instead, pre-training papers have focused their attention on alternatives to ﬁltering, like deduplication [62], federated learning [108, 50], differential privacy [80, 42], and other approaches [32, 74, 65, 127, 82].”
Privacy Regularization: Joint Privacy-Utility Optimization in LanguageModels (2021, unknown): “Our work is closely related to (Coavoux et al., 2018) and (Li et al., 2018).”
Adversarial Removal of Demographic Attributes Revisited (2019, background): “…training has been used to learn data representations that are invariant to demographic attributes (Raff and Sylvester, 2018; Beutel et al., 2017; Li et al., 2018), as well as representations invariant to domain differences (Ganin and Lempitsky, 2015), clean or noisy speech (Sriram et al.,…”
Gender-preserving Debiasing for Pre-trained Word Embeddings (2019, background): “…can be seen as a problem of hiding information related to a protected attribute such as gender, for which adversarial learning meth-ods (Xie et al., 2017; Elazar and Goldberg, 2018; Li et al., 2018) have been proposed in the fairness-aware machine learning community (Kamiran and Calders, 2009).”
Privacy-preserving Neural Representations of Text (2018, methodology): “In a recent study, Li et al. (2018) proposed a method based on GAN designed to improve the robustness and privacy of neural representations, applied to part-of-speech tagging and sentiment analysis.”
Submix: Practical Private Prediction for Large-Scale Language Models (2022, background): “Other related works exploring privacy in natural language processing include (Gopi et al., 2020; Lyu et al., 2020; Xu et al., 2020; Li et al., 2018; Kim et al., 2021).”
Towards Robustifying NLI Models Against Lexical Dataset Biases (2020, methodology): “Li et al. (2018) used a similar approach to learn privacy-preserving representations.”