Meta-paths have been popularly used to provide explainability in recommendations. Although long/complicated meta-paths could represent complex user-item connectivity, they are not easy interpret. This work tackles this problem by introducing a meta-path translation task. The objective is translate its comparable explainable that perform similarly terms of recommendation but higher compared the given one. We propose definition determine and grammar allows be formed similar way as sentences...

SemEval-2024 Task 8 introduces the challenge of identifying machine-generated texts from diverse Large Language Models (LLMs) in various languages and domains. The task comprises three subtasks: binary classification monolingual multilingual (Subtask A), multi-class B), mixed text detection C). This paper focuses on Subtask A & B. Each subtask is supported by datasets for training, development, testing. To tackle this task, two methods: 1) using traditional machine learning (ML) with natural...

In observation of land information using satellite images, clouds are one the most serious obstacles due to their opacity property which can block visibility ground objects and also be blended with underlying details. Hence, retrieval actual covered by is frequently necessary. this paper, we propose a novel method remove taking an advantage HSI color space instead directly removing in RGB space. The proposed uses concept dark channel prior estimate cloud appearance called scattering light...

Recommender systems contain rich relation information. The multiple relations in a recommender system form heterogeneous information network. How to efficiently find similar users and items based on hop-n networks is one significant challenge develop scalable the era of big data. Hashing has been popularly used for dimensionality reduction data size reduction. Current hashing techniques mainly focus directly related (i.e. hop-1) features. This paper proposes relation-aware bridge this gap....

Legal documents tend to be large in size. In this paper, we provide an experiment with attention-based approaches complemented by certain document processing techniques for judgment prediction. For the prediction of explanation, consider as extractive text summarization problem based on output (1) CNN attention mechanism and (2) self-attention language models. Our extensive experiments show that treating endings at first results a 2.1% improvement across all Additional content peeling from...

SemEval Task 4 Commonsense Validation and Explanation Challenge is to validate whether a system can differentiate natural language statements that make sense from those do not sense. Two subtasks, A B, are focused in this work, i.e., detecting against-common-sense selecting explanations of why they false the given options. Intuitively, commonsense validation requires additional knowledge beyond statements. Therefore, we propose utilising pre-trained sentence transformer models based on BERT,...

Crowdsourcing has been ubiquitously used for annotating enormous collections of data. However, the major obstacles to using crowd-sourced labels are noise and errors from non-expert annotations. In this work, two approaches dealing with in proposed. The first approach uses Sharpness-Aware Minimization (SAM), an optimization technique robust noisy labels. other leverages a neural network layer called softmax-Crowdlayer specifically designed learn According results, proposed can improve...

Explainability of recommender systems has become essential to ensure users' trust and satisfaction. Various types explainable have been proposed including graph-based systems. This review paper discusses state-of-the-art approaches these categorizes them based on three aspects: learning methods, explaining explanation types. It also explores the commonly used datasets, explainability evaluation future directions this research area. Compared with existing papers, focuses graphs covers topics...

Knowledge graphs (KGs) have been popularly used in recommender systems to leverage high-order connections between users and items. Typically, KGs are constructed based on semantic information derived from metadata. However, item images also highly useful, especially for those domains where visual factors influential such as fashion In this paper, we propose an approach augment extracted by image feature extraction methods into KGs. Specifically, introduce visually-augmented the is integrated...

Thanet Markchom, Huizhi Liang, Joyce Gitau, Zehao Liu, Varun Ojha, Lee Taylor, Jake Bonnici, Abdullah Alshadadi. Proceedings of the The 17th International Workshop on Semantic Evaluation (SemEval-2023). 2023.

The complexity of stacked imaging and the massive number radiographs make writing radiology reports complex inefficient. Even highly experienced radiologists struggle to maintain accuracy consistency in interpreting under prolonged high-intensity work. To address these issues, this work proposes CRRG-CLIP Model (Chest Radiology Report Generation Radiograph Classification Model), an end-to-end model for automated report generation radiograph classification. consists two modules: module...

In human languages, there are many presuppositional constructions that impose a constrain on the taxonomic relations between two nouns depending their order. These create challenge in validating real-world contexts. SemEval2022-Task3 Presupposed Taxonomies: Evaluating Neural Network Semantics (PreTENS), organizers introduced task regarding within variety of constructions. This is divided into subtasks: classification and regression. Each subtask contains three datasets multiple i.e.,...

Most question answering tasks focuses on predicting concrete answers, e.g., named entities. These can be normally achieved by understanding the contexts without additional information required. In Reading Comprehension of Abstract Meaning (ReCAM) task, abstract answers are introduced. To understand meanings in context, knowledge is essential. this paper, we propose an approach that leverages pre-trained BERT Token embeddings as a prior resource. According to results, our using outperformed...