Various efforts in the Natural Language Processing (NLP) community have been made to accommodate linguistic diversity and serve speakers of many different languages. However, it is important to acknowledge that speakers and the content they produce and require, vary not just by language, but also by culture. Although language and culture are tightly linked, there are important differences. Analogous to cross-lingual and multilingual NLP, cross-cultural and multicultural NLP considers these differences in order to better serve users of NLP systems. We propose a principled framework to frame these efforts, and survey existing and potential strategies.

Unresolved coreference is a bottleneck for relation extraction, and high-quality coreference resolvers may produce an output that makes it a lot easier to extract knowledge triples. We show how to improve coreference resolvers by forwarding their input to a relation extraction system and reward the resolvers for producing triples that are found in knowledge bases. Since relation extraction systems can rely on different forms of supervision and be biased in different ways, we obtain the best performance, improving over the state of the art, using multi-task reinforcement learning.

Creole languages such as Nigerian Pidgin English and Haitian Creole are under-resourced and largely ignored in the NLP literature. Creoles typically result from the fusion of a foreign language with multiple local languages, and what grammatical and lexical features are transferred to the creole is a complex process (Sessarego, 2020). While creoles are generally stable, the prominence of some features may be much stronger with certain demographics or in some linguistic situations (Winford, 1999;Patrick, 1999). This paper makes several contributions: We collect existing corpora and release models for Haitian Creole, Nigerian Pidgin English, and Singaporean Colloquial English. We evaluate these models on intrinsic and extrinsic tasks. Motivated by the above literature, we compare standard language models with distributionally robust ones and find that, somewhat surprisingly, the standard language models are superior to the distributionally robust ones. We investigate whether this is an effect of overparameterization or relative distributional stability, and find that the difference persists in the absence of over-parameterization, and that drift is limited, confirming the relative stability of creole languages.

The importance of academic publications is often evaluated by the number of and impact of its subsequent citing works. These citing works build upon the referenced material, representing both further intellectual insights and additional derived uses. As such, reading peer-reviewed articles which cite one's work can serve as a way for authors to understand how their research is being adopted and extended by the greater scientific community, further develop the broader impacts of their research, and even find new collaborators. Unfortunately, in today's rapidly growing and shifting scientific landscape, it is unlikely that a researcher has enough time to read through all articles citing their works, especially in the case of highly-cited broad-impact studies. To address this challenge, we developed the Science Citation Knowledge Extractor (SCKE), a web tool to provide biological and biomedical researchers with an overview of how their work is being utilized by the broader scientific community. SCKE is a web-based tool which utilizes natural language processing and machine learning to retrieve key information from scientific publications citing a given work, analyze the citing material, and present users with interactive data visualizations which illustrate how their works are contributing to greater scientific pursuits. Results are generally grouped into two categories, aimed at (1) understanding the broad scientific areas which one's work is impacting and (2) assessing the breadth and impact of one's work within these areas. As a web application, SCKE is easy to use, with a single input of PubMed ID(s) to analyze. SCKE is available for immediate use by the scientific community as a hosted web application at https:// geco.iplantcollaborative.org/scke/. SCKE can also be self-hosted by taking advantage of a fully-integrated VM Image (https://tinyurl.com/y7ggpvaa), Docker container (https:// tinyurl.com/y95u9dhw), or open-source code (GPL license) available on GitHub (https:// tinyurl.com/yaesue5e).

Building causal models of complicated phenomena such as food insecurity is currently a slow and labor-intensive manual process. In this paper, we introduce an approach that builds executable probabilistic models from raw, free text. The proposed approach is implemented through three systems: Eidos 1 , IN-DRA 2 and Delphi 3. Eidos is an open-domain machine reading system designed to extract causal relations from natural language. It is rule-based, allowing for rapid domain transfer, customizability, and interpretability. IN-DRA aggregates multiple sources of causal information and performs assembly to create a coherent knowledge base and assess its reliability. This assembled knowledge serves as the starting point for modeling. Delphi is a modeling framework that assembles quantified causal fragments and their contexts into executable probabilistic models that respect the semantics of the original text and can be used to support decision making.

No abstract

Coreference resolution and semantic role labeling are NLP tasks that capture different aspects of semantics, indicating respectively, which expressions refer to the same entity, and what semantic roles expressions serve in the sentence. However, they are often closely interdependent, and both generally necessitate natural language understanding. Do they form a coherent abstract representation of documents? We present a neural network architecture for joint coreference resolution and semantic role labeling for English, and train graph neural networks to model the 'coherence' of the combined shallow semantic graph. Using the resulting coherence score as a reward for our joint semantic analyzer, we use reinforcement learning to encourage global coherence over the document and between semantic annotations. This leads to improvements on both tasks in multiple datasets from different domains, and across a range of encoders of different expressivity, calling, we believe, for a more holistic approach for semantics in NLP.

We aim to learn language models for Creole languages for which large volumes of data are not readily available, and therefore explore the potential transfer from ancestor languages (the 'Ancestry Transfer Hypothesis'). We find that standard transfer methods do not facilitate ancestry transfer. Surprisingly, different from other non-Creole languages, a very distinct two-phase pattern emerges for Creoles: As our training losses plateau, and language models begin to overfit on their source languages, perplexity on the Creoles drop. We explore if this compression phase can lead to practically useful language models (the 'Ancestry Bottleneck Hypothesis'), but also falsify this. Moreover, we show that Creoles even exhibit this twophase pattern even when training on random, unrelated languages. Thus Creoles seem to be typological outliers and we speculate whether there is a link between the two observations.