The public sector faces several challenges, such as a number of external and internal demands for change, citizens' dissatisfaction and frustration with public sector organizations, that need to be addressed. An alternative to the traditional top-down development of public services is co-creation of public services. Co-creation promotes collaboration between stakeholders with the aim to create better public services and achieve public values. At the same time, data analytics has been fuelled by the availability of immense amounts of textual data. Whilst both co-creation and TA have been used in the private sector, we study existing works on the application of Text Analytics (TA) techniques on text data to support public service co-creation. We systematically review 75 of the 979 papers that focus directly or indirectly on the application of TA in the context of public service development. In our review, we analyze the TA techniques, the public service they support, public value outcomes, and the co-creation phase they are used in. Our findings indicate that the TA implementation for co-creation is still in its early stages and thus still limited. Our research framework promotes the concept and stimulates the strengthening of the role of Text Analytics techniques to support public sector organisations and their use of co-creation process. From policy-makers' and public administration managers' standpoints, our findings and the proposed research framework can be used as a guideline in developing a strategy for the designing co-created and user-centred public services.

These findings are documented based on interpretations from imaging examinations (e.g., fundus examination), complications or outcomes associated with surgeries (e.g., cataract surgery), and experiences or symptoms shared by patients. Such findings are oftentimes described along with their exact eye locations as well as other contextual information such as their timing and status. Thus, ophthalmology notes comprise of spatial relations between eye findings and their corresponding locations, and these findings are further described using different spatial characteristics such as laterality and size. Although there has been recent advancements in using natural language processing (NLP) methods in the ophthalmology domain, they are mainly targeted for specific ocular conditions. Some work leveraged electronic health record text data to identify conditions such as glaucoma [1], herpes zoster ophthalmicus [2], and exfoliation syndrome [3], while another set of work extracted quantitative measures particularly related to visual acuity [4, 5] and microbial keratitis [6]. In this work, we aim to extract more comprehensive information related to all eye findings, covering both spatial and contextual, from the ophthalmology notes. Besides automated screening and diagnosis of various ocular conditions, identifying such detailed information can aid in applications such as automated monitoring of eye findings or diseases and cohort retrieval for retrospective epidemiological studies. For this, we propose to extend our existing radiology spatial representation schema-Rad-SpatialNet [7] to the ophthalmology domain. We refer to this as the Eye-SpatialNet schema in this paper.

Information extraction (IE) systems aim to automatically extract structured information, such as named entities, relations between entities, and events, from unstructured texts. While most existing work addresses a particular IE task, universally modeling various IE tasks with one model has achieved great success recently. Despite their success, they employ a one-stage learning strategy, i.e., directly learning to extract the target structure given the input text, which contradicts the human learning process. In this paper, we propose a unified easy-to-hard learning framework consisting of three stages, i.e., the easy stage, the hard stage, and the main stage, for IE by mimicking the human learning process. By breaking down the learning process into multiple stages, our framework facilitates the model to acquire general IE task knowledge and improve its generalization ability. Extensive experiments across four IE tasks demonstrate the effectiveness of our framework. We achieve new state-of-the-art results on 13 out of 17 datasets. Our code is available at \url{https://github.com/DAMO-NLP-SG/IE-E2H}.

Sentiment analysis (SA) systems are used in many products and hundreds of languages. Gender and racial biases are well-studied in English SA systems, but understudied in other languages, with few resources for such studies. To remedy this, we build a counterfactual evaluation corpus for gender and racial/migrant bias in four languages. We demonstrate its usefulness by answering a simple but important question that an engineer might need to answer when deploying a system: What biases do systems import from pre-trained models when compared to a baseline with no pre-training? Our evaluation corpus, by virtue of being counterfactual, not only reveals which models have less bias, but also pinpoints changes in model bias behaviour, which enables more targeted mitigation strategies. We release our code and evaluation corpora to facilitate future research.

We explore how weak supervision on abundant unlabeled data can be leveraged to improve few-shot performance in aspect-based sentiment analysis (ABSA) tasks. We propose a pipeline approach to construct a noisy ABSA dataset, and we use it to adapt a pre-trained sequence-to-sequence model to the ABSA tasks. We test the resulting model on three widely used ABSA datasets, before and after fine-tuning. Our proposed method preserves the full fine-tuning performance while showing significant improvements (15.84% absolute F1) in the few-shot learning scenario for the harder tasks. In zero-shot (i.e., without fine-tuning), our method outperforms the previous state of the art on the aspect extraction sentiment classification (AESC) task and is, additionally, capable of performing the harder aspect sentiment triplet extraction (ASTE) task.

Prompt learning's fine-tune performance on text classification tasks has attracted the NLP community. This paper applies it to resume information extraction, improving existing methods for this task. We created manual templates and verbalizers tailored to resume texts and compared the performance of Masked Language Model (MLM) and Seq2Seq PLMs. Also, we enhanced the verbalizer design for Knowledgeable Prompt-tuning, contributing to prompt template design across NLP tasks. We present the Manual Knowledgeable Verbalizer (MKV), a rule for constructing verbalizers for specific applications. Our tests show that MKV rules yield more effective, robust templates and verbalizers than existing methods. Our MKV approach resolved sample imbalance, surpassing current automatic prompt methods. This study underscores the value of tailored prompt learning for resume extraction, stressing the importance of custom-designed templates and verbalizers.

Cross-domain aspect-based sentiment analysis (ABSA) aims to perform various fine-grained sentiment analysis tasks on a target domain by transferring knowledge from a source domain. Since labeled data only exists in the source domain, a model is expected to bridge the domain gap for tackling cross-domain ABSA. Though domain adaptation methods have proven to be effective, most of them are based on a discriminative model, which needs to be specifically designed for different ABSA tasks. To offer a more general solution, we propose a unified bidirectional generative framework to tackle various cross-domain ABSA tasks. Specifically, our framework trains a generative model in both text-to-label and label-to-text directions. The former transforms each task into a unified format to learn domain-agnostic features, and the latter generates natural sentences from noisy labels for data augmentation, with which a more accurate model can be trained. To investigate the effectiveness and generality of our framework, we conduct extensive experiments on four cross-domain ABSA tasks and present new state-of-the-art results on all tasks. Our data and code are publicly available at \url{https://github.com/DAMO-NLP-SG/BGCA}.

Aspect-based sentiment classification (ABSC) has revealed the potential dependency of sentiment polarities among different aspects. Our study further explores this phenomenon, positing that adjacent aspects often exhibit similar sentiments, a concept we term "aspect sentiment coherency." We argue that the current research landscape has not fully appreciated the significance of modeling aspect sentiment coherency. To address this gap, we introduce a local sentiment aggregation paradigm (LSA) that facilitates fine-grained sentiment coherency modeling. This approach enables the extraction of implicit sentiments for aspects lacking explicit sentiment descriptions. Leveraging gradient descent, we design a differential-weighted sentiment aggregation window that guides the modeling of aspect sentiment coherency. Experimental results affirm the efficacy of LSA in learning sentiment coherency, as it achieves state-of-the-art performance across three public datasets, thus significantly enhancing existing ABSC models. We have made our code available, providing a ready tool for existing methods to harness the potential of sentiment coherency information.

Understanding documents is central to many real-world tasks but remains a challenging topic. Unfortunately, there is no well-established consensus on how to comprehensively evaluate document understanding abilities, which significantly hinders the fair comparison and measuring the progress of the field. To benchmark document understanding researches, this paper summarizes four representative abilities, i.e., document classification, document structural analysis, document information extraction, and document transcription. Under the new evaluation framework, we propose \textbf{Document Language Understanding Evaluation} -- \textbf{DLUE}, a new task suite which covers a wide-range of tasks in various forms, domains and document genres. We also systematically evaluate six well-established transformer models on DLUE, and find that due to the lengthy content, complicated underlying structure and dispersed knowledge, document understanding is still far from being solved, and currently there is no neural architecture that dominates all tasks, raising requirements for a universal document understanding architecture.

Progress in NLP is increasingly measured through benchmarks; hence, contextualizing progress requires understanding when and why practitioners may disagree about the validity of benchmarks. We develop a taxonomy of disagreement, drawing on tools from measurement modeling, and distinguish between two types of disagreement: 1) how tasks are conceptualized and 2) how measurements of model performance are operationalized. To provide evidence for our taxonomy, we conduct a meta-analysis of relevant literature to understand how NLP tasks are conceptualized, as well as a survey of practitioners about their impressions of different factors that affect benchmark validity. Our meta-analysis and survey across eight tasks, ranging from coreference resolution to question answering, uncover that tasks are generally not clearly and consistently conceptualized and benchmarks suffer from operationalization disagreements. These findings support our proposed taxonomy of disagreement. Finally, based on our taxonomy, we present a framework for constructing benchmarks and documenting their limitations.