The web contains an abundance of user- generated content. While this content is useful for many applications, it poses many challenges due to the presence of offensive, biased, and overall toxic language. In this work, we present a system that identifies and classifies sexist content at different levels of granularity. Using transformer-based models, we explore the value of data augmentation, use of ensemble methods, and leverage in-context learning using foundation models to tackle the task. We evaluate the different components of our system both quantitatively and qualitatively. Our best systems achieve an F1 score of 0.84 for the binary classification task aiming to identify whether a given content is sexist or not and 0.64 and 0.47 for the two multi-class tasks that aim to identify the coarse and fine-grained types of sexism present in the given content respectively.

We explore Interactive Post-Editing (IPE) models for human-in-loop translation to help correct translation errors and rephrase it with a desired style variation. We specifically study verbosity for style variations and build on top of multi-source transformers that can read source and hypothesis to improve the latter with user inputs. Token-level interaction inputs for error corrections and length interaction inputs for verbosity control are used by the model to generate a suitable translation. We report BERTScore to evaluate semantic quality with other relevant metrics for translations from English to German, French and Spanish languages. Our model achieves superior BERTScore over state-of-the-art machine translation models while maintaining the desired token-level and verbosity preference.

Automatic post-editing (APE) models are usedto correct machine translation (MT) system outputs by learning from human post-editing patterns. We present the system used in our submission to the WMT’21 Automatic Post-Editing (APE) English-German (En-De) shared task. We leverage the state-of-the-art MT system (Ng et al., 2019) for this task. For further improvements, we adapt the MT model to the task domain by using WikiMatrix (Schwenket al., 2021) followed by fine-tuning with additional APE samples from previous editions of the shared task (WMT-16,17,18) and ensembling the models. Our systems beat the baseline on TER scores on the WMT’21 test set.