To appear in IEEE Transactions on Visualization and Computer Graphics. Figure 1: We present a workspace that enables the evaluation and comparison of adapters - lightweight alternatives for language model fine-tuning. After data pre-processing (e.g., embedding extraction), users can select pre-trained adapters, create explanations, and explore model differences through three types of visualizations: Concept Embedding Similarity, Concept Embedding Projection, and Concept Prediction Similarity. The explanations are provided for single models as well as model comparisons. For each explanation, we provide further explanation details, such as the word contexts as well as embedding vectors themselves. Abstract--Neural language models are widely used; however, their model parameters often need to be adapted to the specific domains and tasks of an application, which is time-and resource-consuming. Thus, adapters have recently been introduced as a lightweight alternative for model adaptation. They consist of a small set of task-specific parameters with a reduced training time and simple parameter composition. The simplicity of adapter training and composition comes along with new challenges, such as maintaining an overview of adapter properties and effectively comparing their produced embedding spaces. To help developers overcome these challenges, we provide a twofold contribution. First, in close collaboration with NLP researchers, we conducted a requirement analysis for an approach supporting adapter evaluation and detected, among others, the need for both intrinsic (i.e., embedding similaritybased) and extrinsic (i.e., prediction-based) explanation methods. Second, motivated by the gathered requirements, we designed a flexible visual analytics workspace that enables the comparison of adapter properties. In this paper, we discuss several design iterations and alternatives for interactive, comparative visual explanation methods. Our comparative visualizations show the differences in the adapted embedding vectors and prediction outcomes for diverse human-interpretable concepts (e.g., person names, human qualities).

Neural Machine Translation (NMT) is an open vocabulary problem. As a result, dealing with the words not occurring during training (a.k.a. out-of-vocabulary (OOV) words) have long been a fundamental challenge for NMT systems. The predominant method to tackle this problem is Byte Pair Encoding (BPE) which splits words, including OOV words, into sub-word segments. BPE has achieved impressive results for a wide range of translation tasks in terms of automatic evaluation metrics. While it is often assumed that by using BPE, NMT systems are capable of handling OOV words, the effectiveness of BPE in translating OOV words has not been explicitly measured. In this paper, we study to what extent BPE is successful in translating OOV words at the word-level. We analyze the translation quality of OOV words based on word type, number of segments, cross-attention weights, and the frequency of segment n-grams in the training data. Our experiments show that while careful BPE settings seem to be fairly useful in translating OOV words across datasets, a considerable percentage of OOV words are translated incorrectly. Furthermore, we highlight the slightly higher effectiveness of BPE in translating OOV words for special cases, such as named-entities and when the languages involved are linguistically close to each other.

People often utilise online media (e.g., Facebook, Reddit) as a platform to express their psychological distress and seek support. State-of-the-art NLP techniques demonstrate strong potential to automatically detect mental health issues from text. Research suggests that mental health issues are reflected in emotions (e.g., sadness) indicated in a person's choice of language. Therefore, we developed a novel emotion-annotated mental health corpus (EmoMent), consisting of 2802 Facebook posts (14845 sentences) extracted from two South Asian countries - Sri Lanka and India. Three clinical psychology postgraduates were involved in annotating these posts into eight categories, including 'mental illness' (e.g., depression) and emotions (e.g., 'sadness', 'anger'). EmoMent corpus achieved 'very good' inter-annotator agreement of 98.3% (i.e. % with two or more agreement) and Fleiss' Kappa of 0.82. Our RoBERTa based models achieved an F1 score of 0.76 and a macro-averaged F1 score of 0.77 for the first task (i.e. predicting a mental health condition from a post) and the second task (i.e. extent of association of relevant posts with the categories defined in our taxonomy), respectively.

Maximum Common induced Subgraph (MCS) is an important NP-hard problem with wide real-world applications. Branch-and-Bound (BnB) is the basis of a class of efficient algorithms for MCS, consisting in successively selecting vertices to match and pruning when it is discovered that a solution better than the best solution found so far does not exist. The method of selecting the vertices to match is essential for the performance of BnB. In this paper, we propose a new value function and a hybrid selection strategy used in reinforcement learning to define a new vertex selection method, and propose a new BnB algorithm, called McSplitDAL, for MCS. Extensive experiments show that McSplitDAL significantly improves the current best BnB algorithms, McSplit+LL and McSplit+RL. An empirical analysis is also performed to illustrate why the new value function and the hybrid selection strategy are effective.

Knowledge Graphs (KGs) have shown to be very important for applications such as personal assistants, question-answering systems, and search engines. Therefore, it is crucial to ensure their high quality. However, KGs inevitably contain errors, duplicates, and missing values, which may hinder their adoption and utility in business applications, as they are not curated, e.g., low-quality KGs produce low-quality applications that are built on top of them. In this vision paper, we propose a practical knowledge graph curation framework for improving the quality of KGs. First, we define a set of quality metrics for assessing the status of KGs, Second, we describe the verification and validation of KGs as cleaning tasks, Third, we present duplicate detection and knowledge fusion strategies for enriching KGs. Furthermore, we give insights and directions toward a better architecture for curating KGs.

In an age where a vast amount of data is being Alas, in the field of NLP, where the core unit of produced daily, the opportunities created by this data is unstructured, fuzzy text rather than a structured proliferation increase concurrently. The availability data point, an initial attempt to apply Differential of big data enables countless downstream tasks Privacy poses some challenges. Chief among whose accuracy and utility seem to increase with these is the challenge of how to transfer the core the amount of data used. Specifically, the fields concepts of Differential Privacy, namely the "individual" of Machine Learning (ML) and Deep Learning and adjacency, to the textual domain where (DL) have profited from such data. Particularly in these concepts are not easily perceivable. Thus, it the case of Natural Language Processing (NLP), becomes the goal to find new ways of reasoning the tasks at hand more often than not concern the about Differential Privacy in order to adapt it to handling of unstructured data, meaning data that the unstructured data domain of NLP. Through the is not neatly organized into a traditional row-like course of this paper, the foundations of Differential database structure, and furthermore, data that is Privacy in the lens of NLP will be investigated, motivated not necessarily static. In fact, it is estimated that by some privacy vulnerabilities that surface data on the order of zettabytes (ZB) is being produced from NLP techniques. Afterwards, the limitations every day (Begum and Nausheen, 2018), and and open questions of Differential Privacy with within this amount, roughly 80% is unstructured, NLP will be analyzed with an in-depth discussion.

This research provides the first comprehensive analysis of the performance of pre-trained language models for Sinhala text classification. We test on a set of different Sinhala text classification tasks and our analysis shows that out of the pre-trained multilingual models that include Sinhala (XLM-R, LaBSE, and LASER), XLM-R is the best model by far for Sinhala text classification. We also pre-train two RoBERTa-based monolingual Sinhala models, which are far superior to the existing pre-trained language models for Sinhala. We show that when fine-tuned, these pre-trained language models set a very strong baseline for Sinhala text classification and are robust in situations where labeled data is insufficient for fine-tuning. We further provide a set of recommendations for using pre-trained models for Sinhala text classification. We also introduce new annotated datasets useful for future research in Sinhala text classification and publicly release our pre-trained models.

This month, Amazon inked a deal to acquire smart vacuum company iRobot – the makers of Roomba – for a tidy US$1.7bn. As some see it, if the purchase goes through, that should worry us. "It's all about the data," says David Vaile from the Australian Privacy Foundation. Privacy advocates such as Vaile are concerned the robot vacuum cleaner will give Amazon access to floor plans of users' homes, using mapping features some iRobot products already offer. Amazon are yet to release details about what existing and future iRobot data will be used for; and the company told Reuters that they safeguard customer privacy and do not sell their data.

Relation Extraction (RE) refers to extracting the relation triples in the input text. Existing neural work based systems for RE rely heavily on manually labeled training data, but there are still a lot of domains where sufficient labeled data does not exist. Inspired by the distance-based few-shot named entity recognition methods, we put forward the definition of the few-shot RE task based on the sequence tagging joint extraction approaches, and propose a few-shot RE framework for the task. Besides, we apply two actual sequence tagging models to our framework (called Few-shot TPLinker and Few-shot BiTT), and achieves solid results on two few-shot RE tasks constructed from a public dataset.

Client selection schemes are widely adopted to handle the communication-efficient problems in recent studies of Federated Learning (FL). However, the large variance of the model updates aggregated from the randomly-selected unrepresentative subsets directly slows the FL convergence. We present a novel clustering-based client selection scheme to accelerate the FL convergence by variance reduction. Simple yet effective schemes are designed to improve the clustering effect and control the effect fluctuation, therefore, generating the client subset with certain representativeness of sampling. Theoretically, we demonstrate the improvement of the proposed scheme in variance reduction. We also present the tighter convergence guarantee of the proposed method thanks to the variance reduction. Experimental results confirm the exceed efficiency of our scheme compared to alternatives.