The recent advances in NLP have lead to exciting new language technologies being deployed at scale, affecting millions of users daily. However, these advances have been anything but equitable, as we observe a rich-get-richer scenario for high-resource languages, while whole language communities are left behind without access to language technology.
In this talk, I will focus on a sample of research threads towards the goal of building NLP that serves everyone. First, I'll describe work that is tailored to creating NLP tools for endangered language documentation, with neural models that take advantage of additional signals (in this case translations) in multi-source and multi-task settings. Second, I'll show how similarities across languages can be leveraged for building more accurate morphological inflection systems in numerous under-resourced languages, along with data augmentation through hallucination. I will also briefly describe our work towards building systems robust to grammatical errors made by non-native speakers. Last, I will suggest several directions for future research.