REA: Robust Cross-lingual Entity Alignment Between Knowledge Graphs

Pei, Shichao; Lu, Yu; Yu, Guoxian; Zhang, Xiangliang

doi:10.1145/3394486.3403268

Cited by 23 publications

(7 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Danh sách các tòa nhà của Ottawa CLISM Input: [CLS] [QUE] of [QUE] Here we present the input sequence in English and Vietnamese. Selected spans in English are from off-the-shelf NER tools like Spacy and the corresponding aligned spans in Vietnamese obtained by GIZA++ toolkit (Pei et al, 2020). starts dominating various cross-lingual understanding tasks including entity recognition (Liang et al, 2021a,b) and question answering (Asai et al, 2018;Chen et al, 2021b).…”

Section: Viestamese(vi)mentioning

confidence: 99%

“…Considering the translation of the same phrase may be different when pushing it into translation systems separately or concat it with a whole sentence. Therefore, we leverage the off-the-shelf alignment tool like GIZA++ (Pei et al, 2020) to align the corresponding words of the selected entity in the target language. After extracting entity spans and alignment 2 , we replace these selected spans with a special token [QUE] in source language sentence s s , and the corresponding alignment spans in s t act as the"ground-truth answer" span for each masked "question" ([QUE]).…”

Section: Cross-lingual Language Informative Span Maskingmentioning

confidence: 99%

See 1 more Smart Citation

Bridging the Gap between Language Models and Cross-Lingual Sequence Labeling

Shou¹,

Gong²,

Pei³

et al. 2022

Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Langua

View full text Add to dashboard Cite

Large-scale cross-lingual pre-trained language models (xPLMs) have shown effectiveness in cross-lingual sequence labeling tasks (xSL), such as cross-lingual machine reading comprehension (xMRC) by transferring knowledge from a high-resource language to low-resource languages. Despite the great success, we draw an empirical observation that there is a training objective gap between pre-training and finetuning stages: e.g., mask language modeling objective requires local understanding of the masked token and the span-extraction objective requires global understanding and reasoning of the input passage/paragraph and question, leading to the discrepancy between pretraining and xMRC. In this paper, we first design a pre-training task tailored for xSL named Cross-lingual Language Informative Span Masking (CLISM) to eliminate the objective gap in a self-supervised manner. Second, we present ContrAstive-Consistency Regularization (CACR), which utilizes contrastive learning to encourage the consistency between representations of input parallel sequences via unsupervised cross-lingual instance-wise training signals during pre-training. By these means, our methods not only bridge the gap between pretrain-finetune, but also enhance PLMs to better capture the alignment between different languages. Extensive experiments prove that our method achieves clearly superior results on multiple xSL benchmarks with limited pretraining data. Our methods also surpass the previous state-of-the-art methods by a large margin in few-shot data settings, where only a few hundred training examples are available.

show abstract

Section: Viestamese(vi)mentioning

confidence: 99%

Section: Cross-lingual Language Informative Span Maskingmentioning

confidence: 99%

Bridging the Gap between Language Models and Cross-Lingual Sequence Labeling

Shou¹,

Gong²,

Pei³

et al. 2022

Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Langua

View full text Add to dashboard Cite

show abstract

“…REA [20] has proposed a framework for robust entity alignment over KGs. The framework consists of two components: noise detection and noise-aware entity alignment.…”

Section: Related Workmentioning

confidence: 99%

Finding Similar Entities Across Knowledge Graphs

Aghaei¹,

Fensel²

2021

Computer Science &Amp; Information Technology (CS &Amp; IT)

View full text Add to dashboard Cite

Finding similar entities among knowledge graphs is an essential research problem for knowledge integration and knowledge graph connection. This paper aims at finding semantically similar entities between two knowledge graphs. This can help end users and search agents more effectively and easily access pertinent information across knowledge graphs. Given a query entity in one knowledge graph, the proposed approach tries to find the most similar entity in another knowledge graph. The main idea is to leverage graph embedding, clustering, regression and sentence embedding. In this approach, RDF2Vec has been employed to generate vector representations of all entities of the second knowledge graph and then the vectors have been clustered based on cosine similarity using K medoids algorithm. Then, an artificial neural network with multilayer perception topology has been used as a regression model to predict the corresponding vector in the second knowledge graph for a given vector from the first knowledge graph. After determining the cluster of the predicated vector, the entities of the detected cluster are ranked through sentence-BERT method and finally the entity with the highest rank is chosen as the most similar one. To evaluate the proposed approach, experiments have been conducted on real-world knowledge graphs. The experimental results demonstrate the effectiveness of the proposed approach.

show abstract

“…Such entity alignment task usually assumes that different KGs are partially aligned and tries to predict more aligned entities. For example, the entity alignment can be based on cross-lingual KGs [8,32,43,52], KGs with multi-view entity-related information [5,7,50], and KGs in similar domains with significant entity overlaps [36,38,47,51]. After joint learning, embeddings for entity alignment are usually aligned in a unified space so the vectors can be used to find nearest entities in other KGs.…”

Section: Knowledge Graph Embeddingmentioning

confidence: 99%

Differentially Private Federated Knowledge Graphs Embedding

Peng,

Li,

Song

et al. 2021

Preprint

View full text Add to dashboard Cite

In this paper, we propose a novel decentralized scalable learning framework, Federated Knowledge Graphs Embedding (FKGE), where embeddings from different knowledge graphs can be learnt in an asynchronous and peer-to-peer manner while being privacypreserving. FKGE exploits adversarial generation between pairs of knowledge graphs to translate identical entities and relations of different domains into near embedding spaces. In order to protect the privacy of the training data, FKGE further implements a privacy-preserving neural network structure to guarantee no raw data leakage. We conduct extensive experiments to evaluate FKGE on 11 knowledge graphs, demonstrating a significant and consistent improvement in model quality with at most 17.85% and 7.90% increases in performance on triple classification and link prediction tasks.

show abstract

REA: Robust Cross-lingual Entity Alignment Between Knowledge Graphs

Cited by 23 publications

References 42 publications

Bridging the Gap between Language Models and Cross-Lingual Sequence Labeling

Bridging the Gap between Language Models and Cross-Lingual Sequence Labeling

Finding Similar Entities Across Knowledge Graphs

Differentially Private Federated Knowledge Graphs Embedding

Contact Info

Product

Resources

About