Knowledge graphs (KGs) represent structured real-world information composed by triplets of head entity, relation, and tail entity. These graphs can be constructed automatically from text or manually curated. However, regardless of the construction method, KGs often suffer from misinformation, incompleteness, and noise, which hinder their reliability and utility. This study addresses the challenge of noisy KGs, where incorrect or misaligned entities and relations degrade graph quality. Leveraging recent advancements in large language models (LLMs) with strong capabilities across diverse tasks, we explore their potential to detect and refine noise in KGs. Specifically, we propose a novel method, LLM_sim, to enhance the detection and refinement of noisy triples. Our results confirm the effectiveness of this approach in elevating KG quality in noisy environments. Additionally, we apply our proposed method to Knowledge Graph Completion (KGC), a downstream KG task that aims to predict missing links and improve graph completeness. Traditional KGC methods assume that KGs are noise-free, which is unrealistic in practical scenarios. Our experiments analyze the impact of varying noise levels on KGC performance, revealing that LLMs can mitigate noise by identifying and refining incorrect entries, thus enhancing KG quality.

This paper proposes ProtoBox, a novel method to learn contextualized box embeddings. Unlike an ordinary word embedding, which represents a word as a single vector, a box embedding represents the meaning of a word as a box in a high-dimensional space: that is suitable for representing semantic relations between words. In addition, our method aims to obtain a “contextualized” box embedding, which is an abstract representation of a word in a specific context. ProtoBox is based on Prototypical Networks, which is a robust method for classification problems, especially focusing on learning the hypernym–hyponym relation between senses. ProtoBox is evaluated on three tasks: Word Sense Disambiguation (WSD), New Sense Classification (NSC), and Hypernym Identification (HI). Experimental results show that ProtoBox outperforms baselines for the HI task and is comparable for the WSD and NSC tasks.

Aspect Category Sentiment Analysis (ACSA) is one of the main subtasks of sentiment analysis, which aims at predicting polarity over a given aspect category. Recently, generative methods emerge as an efficient way to utilize a pre-trained language model for solving ACSA. However, those methods fail to model relations of target words and opinion words in a sentence including multiple aspects. To tackle this problem, this paper proposes a method to incorporate Abstract Meaning Representation (AMR), which describes semantic representation of a sentence as a directed graph, into a text generation model. Furthermore, two regularizers are designed to guide cross attention weights allocation over AMR graphs. One is the identical regularizer that constrains attention weights of aligned nodes, the other is the entropy regularizer that helps the decoder generate tokens by heavily considering only a few related nodes in the AMR graph. Experimental results on three datasets show that the proposed method outperforms state-of-the-art methods, proving the effectiveness of our model.

In product reviews, user and product aspects are useful in sentiment analysis. Nevertheless, previous studies mainly focus on modeling user and product aspects without considering the relationship between users and products. The relationship between users and products is typically helpful in estimating the bias of a user toward a product. In this paper, we, therefore, introduce the Graph Neural Network-based model with the pre-trained Language Model (GNNLM), where the relationship between users and products is incorporated. We conducted experiments on three well-known benchmarks for sentiment classification with the user and product information. The experimental results show that the relationship between users and products improves the performance of sentiment analysis. Furthermore, GNNLM achieves state-of-the-art results on yelp-2013 and yelp-2014 datasets.

Shortcut reasoning is an irrational process of inference, which degrades the robustness of an NLP model. While a number of previous work has tackled the identification of shortcut reasoning, there are still two major limitations: (i) a method for quantifying the severity of the discovered shortcut reasoning is not provided; (ii) certain types of shortcut reasoning may be missed. To address these issues, we propose a novel method for identifying shortcut reasoning. The proposed method quantifies the severity of the shortcut reasoning by leveraging out-of-distribution data and does not make any assumptions about the type of tokens triggering the shortcut reasoning. Our experiments on Natural Language Inference and Sentiment Analysis demonstrate that our framework successfully discovers known and unknown shortcut reasoning in the previous work.

This paper proposes a method to develop a machine translation (MT) system from Myanmar Sign Language (MSL) to Myanmar Written Language (MWL) and vice versa for the deaf community. Translation of MSL is a difficult task since only a small amount of a parallel corpus between MSL and MWL is available. To address the challenge for MT of the low-resource language, transfer learning is applied. An MT model is trained first for a high-resource language pair, American Sign Language (ASL) and English, then it is used as an initial model to train an MT model between MSL and MWL. The mT5 model is used as a base MT model in this transfer learning. Additionally, a self-training technique is applied to generate synthetic translation pairs of MSL and MWL from a large monolingual MWL corpus. Furthermore, since the segmentation of a sentence is required as preprocessing of MT for the Myanmar language, several segmentation schemes are empirically compared. Results of experiments show that both transfer learning and self-training can enhance the performance of the translation between MSL and MWL compared with a baseline model fine-tuned from a small MSL-MWL parallel corpus only.

The emergence of pre-trained language models has taken story generation, which is the task of automatically generating a comprehensible story from limited information, to a new stage. Although generated stories from the language models are fluent and grammatically correct, the lack of coherence affects their quality. We propose a knowledge-based multi-stage model that incorporates the schema, a kind of structured knowledge, to guide coherent story generation. Our framework includes a schema acquisition module, a plot generation module, and a surface realization module. In the schema acquisition module, high-relevant structured knowledge pieces are selected as a schema. In the plot generation module, a coherent plot plan is navigated by the schema. In the surface realization module, conditioned by the generated plot, a story is generated. Evaluations show that our methods can generate more comprehensible stories than strong baselines, especially with higher global coherence and less repetition.

Aspect-based sentiment analysis (ABSA) is a task that involves classifying the polarity of aspects of the products or services described in users’ reviews. Most previous work on ABSA has focused on explicit aspects, which appear as explicit words or phrases in the sentences of the review. However, users often express their opinions toward the aspects indirectly or implicitly, in which case the specific name of an aspect does not appear in the review. The current datasets used for ABSA are mainly annotated with explicit aspects. This paper proposes a novel method for constructing a corpus that is automatically annotated with implicit aspects. The main idea is that sentences containing explicit and implicit aspects share a similar context. First, labeled sentences with explicit aspects and unlabeled sentences that include implicit aspects are collected. Next, clustering is performed on these sentences so that similar sentences are merged into the same cluster. Finally, the explicit aspects are propagated to the unlabeled sentences in the same cluster, in order to construct a labeled dataset containing implicit aspects. The results of our experiments on mobile phone reviews show that our method of identifying the labels of implicit aspects achieves a maximum accuracy of 82%.

Complaint classification aims at using information to deliver greater insights to enhance user experience after purchasing the products or services. Categorized information can help us quickly collect emerging problems in order to provide a support needed. Indeed, the response to the complaint without the delay will grant users highest satisfaction. In this paper, we aim to deliver a novel approach which can clarify the complaints precisely with the aim to classify each complaint into nine predefined classes i.e. acces-sibility, company brand, competitors, facilities, process, product feature, staff quality, timing respec-tively and others. Given the idea that one word usually conveys ambiguity and it has to be interpreted by its context, the word embedding technique is used to provide word features while applying deep learning techniques for classifying a type of complaints. The dataset we use contains 8,439 complaints of one company.

Corpus-based approaches and statistical approaches have been the main stream of natural language processing research for the past two decades. Language resources play a key role in such approaches, but there is an insufficient amount of language resources in many Asian languages. In this situation, standardisation of language resources would be of great help in developing resources in new languages. This paper presents the latest development efforts of our project which aims at creating a common standard for Asian language resources that is compatible with an international standard. In particular, the paper focuses on i) lexical specification and data categories relevant for building multilingual lexical resources for Asian languages; ii) a core upper-layer ontology needed for ensuring multilingual interoperability and iii) the evaluation platform used to test the entire architectural framework.

In Japanese constructions of the form [N1 no Adj N2], the adjective Adj modifies either N1 or N2. Determing the semantic dependencies of adjective in such phrase is an important task for machine translation. This paper describes a method for determining the adjective dependency in such constructions using decision lists, and inducing decision lists from training contexts with correct semantic dependencies and without. Based on evaluation, our method is able to determine adjective dependency with an precision of about 94%. We further analyze rules in the induced decision lists and examine effective features to determine the semantic dependencies of adjectives.