United States

This paper proposes an approach to the design of an ethical human-AI reasoning support system for decision makers in refugee law. In the context of refugee status determination, practitioners mostly rely on text data. We therefore investigate human-AI cooperation in legal natural language processing. Specifically, we want to determine which design methods can be transposed to legal text analytics. Although little work has been done so far on human-centered design methods applicable to the legal domain, we assume that introducing iterative cooperation and user engagement in the design process is (1) a method to reduce technical limitations of an NLP system and (2) that it will help design more ethical and effective applications by taking users’ preferences and feedback into account. The proposed methodology is based on three main design steps: cognitive process formalization in models understandable by both humans and computers, speculative design of prototypes, and semi-directed interviews with a sample of potential users.

NAACL 2022

Human-Centered Computing in Legal NLP: An Application to Refugee Status Determination

refugee law

human-centered computing

decision making

human-computer interaction

## Welcome to NAACL 2022!
The Annual Meeting of the North American Chapter of the Association for Computational Linguistics (NAACL) is taking place July 10-15, 2022 as a hybrid event, in Seattle, WA and online. We are happy to welcome all of you to conference! 
 
The main conference program features oral presentations, in-person and virtual posters and demo sessions, a plenary session for our best paper presentations and awards, keynote presentations, a plenary panel on the place of linguistics and symbolic structures in NLP, and an Industrial Track panel on Careers in NLP.
 
Posters (including Findings of NAACL 2022) and demos are grouped by areas for both the in-person and the virtual sessions. For the virtual component, the talks will be on Zoom and the posters and the demos will be in GatherTown. The Student Research Workshop and will have an oral session and a poster session.
 
The program also features 6 Tutorials and 26 Workshops.
 
 We wish you a wonderful conference!
 
[The NAACL 2022 Organizing Committee](https://2022.naacl.org/committees/organization/)

2022 Annual Conference of the North American Chapter of the Association for Computational Linguistics

workshop paper

With Human-Centric Research (HCR) we can steer research activities so that the research outcome is beneficial for human stakeholders, such as end users. But what exactly makes research human-centric? We address this question by providing a working definition and define how a research pipeline can be split into different stages in which human-centric components can be added. Additionally, we discuss existing NLP with HCR components and define a series of guiding questions, which can serve as starting points for researchers interested in exploring human-centric research approaches. We hope that this work would inspire researchers to refine the proposed definition and to pose other questions that might be meaningful for achieving HCR.

Human-Centric Research for NLP: Towards a Definition and Guiding Questions

Motivated by practical problems, we outline a plan to sutdy the effectiveness of various families of methods in assisting users complete a downstream task of text matching. Insights from a systematic evaluation of these methods will suggest potential ways to improve user experiences in practical systems.

Understanding the Effectiveness of Model Explanations in Text Matching Tasks

Motivated by prior literature, we provide a proof of concept simulation study for an understudied interactive machine learning method, machine teaching (MT), for the text-based emotion prediction task. We compare this method experimentally against a more well-studied technique, active learning (AL). Results show the strengths of both approaches over more resource-intensive offline supervised learning. Additionally, applying AL and MT to fine-tune a pre-trained model offers further efficiency gain. We end by recommending research directions which aim to empower users in the learning process.

Teaching Interactively to Learn Emotions in Natural Language

Students learn more from doing activities and practicing their skills on assessments, yet it can be challenging and time consuming to generate such practice opportunities. In this work, we present a pipeline for generating and evaluating questions from text-based learning materials in an introductory data science course. The pipeline includes applying a T5 question generation model and a concept hierarchy extraction model on the text content, then scoring the generated questions based on their relevance to the extracted key concepts. We further classified the generated questions as either useful to learning or not with two different approaches: automated labeling by a trained GPT-3 model and manual review by expert human judges. Our results showed that the generated questions were rated favorably by all three evaluation methods. We conclude with a discussion of the strengths and weaknesses of the generated questions and outline the next steps towards refining the pipeline and promoting NLP research in educational domains.

Towards Automated Generation and Evaluation of Questions in Educational Domains

Reliability concerns around Machine Translation (MT) system deployment have been well-documented, a dominant example being medical language which has shown potential for grave harm from translation errors. In this extended abstract, we sketch a research proposal for automatically estimating the quality of machine translation systems in medical settings to ensure they convey the message a clinician seeks to communicate to a patient. We briefly summarize findings from an interview study that is targeted towards understanding the strategies clinicians use to assess the quality of human and machine translations, and the challenges they incur in the process. We then discuss the limitations of existing quality estimation techniques from the NLP literature. Finally, we propose how these findings can inform future work on designing quality estimation systems that are more directly tailored towards clinician needs.

Evaluating the Quality of Machine Translation in Medical Settings

Exposure to ideas in domains outside a scientist's own may benefit her in reformulating existing research problems in novel ways and discovering new application domains for existing solution ideas. While improved performance in scholarly search engines can help scientists efficiently identify relevant advances in domains they may already be familiar with, it may fall short of helping them explore diverse ideas outside such domains. In this paper we explore the design of systems aimed at augmenting the end-user ability in cross-domain exploration with flexible query specification. To this end, we develop an exploratory search system in which end-users can select a portion of text core to their interest from a paper abstract and retrieve papers that have a high similarity to the user-selected core aspect but differ in terms of domains. Furthermore, end-users can `zoom in' to specific domain clusters to retrieve more papers from them and understand nuanced differences within the clusters. Our case studies with scientists uncover opportunities and design implications for systems aimed at facilitating cross-domain exploration and inspiration.

Augmenting Scientific Creativity with Retrieval across Knowledge Domains

Large Language Models (LMs) have achieved state-of-the-art performance on many Natural Language Processing (NLP) benchmarks. With the growing number of new benchmarks, we build bigger and more complex LMs. However, building new LMs may not be an ideal option owing to the cost, time and environmental impact associated with it. We explore an alternative route: can we modify data by expressing it in terms of the model's strengths, so that a question becomes easier for models to answer? We investigate if humans can decompose a hard question into a set of simpler questions that are relatively easier for models to solve. We analyze a range of datasets involving various forms of reasoning and find that it is indeed possible to significantly improve model performance (24\% for GPT3 and 29\% for RoBERTa-SQuAD along with a symbolic calculator) via decomposition. Our approach provides a viable option to involve people in NLP research in a meaningful way. Our findings indicate that Human-in-the-loop Question Decomposition (HQD) can potentially provide an alternate path to building large LMs.

Is a Question Decomposition Unit All We Need?

While Indic NLP has made rapid advances recently in terms of the availability of corpora and pre-trained models, benchmark datasets on standard NLU tasks are limited. To this end, we introduce INDICXNLI, an NLI dataset for
11 Indic languages. It has been created by high-quality machine translation of the original English XNLI dataset and our analysis attests to the quality of INDICXNLI. By finetuning different pre-trained LMs on this INDICXNLI, we analyze various cross-lingual transfer techniques with respect to the impact of the choice of language models, languages, multi-linguality, mix-language input, etc. These experiments provide us with useful insights into the behaviour of pre-trained models for a diverse set of languages

INDICXNLI: Evaluating Multilingual Inference for Indian Languages

The modes of discourse aid in comprehending the convention and purpose of various forms of languages used during communication. In this study, we introduce a discourse mode annotated corpus for the low-resource Bengali (also referred to as Bangla) language. The corpus consists of sentence-level annotation of three discourse modes, narrative, descriptive, and informative of the text excerpted from a number of Bengali novels. We analyze the annotated corpus to expose various linguistic aspects of discourse modes, such as class distributions and average sentence lengths. To automatically determine the mode of discourse, we apply CML (classical machine learning) classifiers with n-gram based statistical features and a fine-tuned BERT (Bidirectional Encoder Representations from Transformers) based language model. We observe that fine-tuned BERT-based model yields better results than CML classifiers. Our created discourse mode annotated dataset, the first of its kind in Bengali, and the evaluation, provide baselines for the automatic discourse mode identification in Bengali and can assist various downstream natural language processing tasks.

An Annotated Dataset and Automatic Approaches for Discourse Mode Identification in Low-resource Bengali Language

Existing methods for open-retrieval question answering in lower resource languages (LRLs) lag significantly behind English. They not only suffer from the shortcomings of non-English document retrieval, but are reliant on language-specific supervision for either the task or translation. We formulate a task setup more realistic to available resources, that circumvents document retrieval to reliably transfer knowledge from English to lower resource languages. Assuming a strong English question answering model or database, we compare and analyze methods that pivot through English: to map foreign queries to English and then English answers back to target language answers. Within this task setup we propose Reranked Multilingual Maximal Inner Product Search (RM-MIPS), akin to semantic similarity retrieval over the English training set with reranking, which outperforms the strongest baselines by 2.7\% on XQuAD and 6.2\% on MKQA. Analysis demonstrates the particular efficacy of this strategy over state-of-the-art alternatives in challenging settings: low-resource languages, with extensive distractor data and query distribution misalignment. Circumventing retrieval, our analysis shows this approach offers rapid answer generation to many other languages off-the-shelf, without necessitating additional training data in the target language.

Downloads

Next from NAACL 2022

Human-Centric Research for NLP: Towards a Definition and Guiding Questions

Similar lecture

GisPy: A Tool for Measuring Gist Inference Score in Text

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from NAACL 2022

Human-Centric Research for NLP: Towards a Definition and Guiding Questions

Similar lecture

GisPy: A Tool for Measuring Gist Inference Score in Text

Downloads