FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313342836> ?p ?o ?g. }

• W4313342836 abstract "Knowledge Base is a semantic data repos-itory made available over the web. Knowledge bases are multi-relational graphs consisting of entities and relations representing facts about the world. These facts follow a controlled vocabulary that drives the knowledge base, often called ontology. Entities are the graph nodes, while relations are the edges connecting the nodes. While knowledge base attention prolifer-ates, extracting information from these resources is challenging. One of the promising approaches is Question Answering systems. The Question Answering over Knowledge Base(KB-QA) system aims to provide a natural language interface to pose a query to the KB. QA systems use embeddings to map the posed Question to the facts stored in the knowledge bases. QA systems can be effectively improved using large transformer models. These transformer models are computationally expensive, which prevents their use in real-world applications. To solve this, we started our study with the use-case of Knowledge Base Question Answering systems. In KB-QA systems, representing the entities and relationships play a pivotal role in extracting the answer to the posed query. Traditional KB embedding techniques represent entities/relations as vectors, mapping them in different semantic spaces and ignoring the subtle cor-relation. A pre-training language model for encoding the KB facts will preserve the semantic and contextual correlation. While learning linguistic knowledge, these models also store relational knowledge in the training data. These Language Models are often trained on a massive corpus addressing the Out-Of- Vocabulary problem. In this paper, we incorporate a co-embedding model for knowledge base embedding, which learns low-dimensional representations of entities and relations in the same semantic space. We propose a Variational Auto-Encoder, an efficient transformer architecture that represents knowledge base representations as Gaussian distributions to address the issue of neglecting uncertainty. In addition, our method has high quality and interpretability advantages compared with previous methods. Our experimental results show the effectiveness and the superiority of the VAE approach for question-answering systems on knowledge bases over other well-known pre-trained embedding methods." @default.
• W4313342836 created "2023-01-06" @default.
• W4313342836 creator A5078942833 @default.
• W4313342836 creator A5090802966 @default.
• W4313342836 date "2022-12-01" @default.
• W4313342836 modified "2023-09-29" @default.
• W4313342836 title "Question Answering over Knowledge Base with Variational Auto-Encoder" @default.
• W4313342836 cites W1552847225 @default.
• W4313342836 cites W1566256432 @default.
• W4313342836 cites W1801721664 @default.
• W4313342836 cites W1832693441 @default.
• W4313342836 cites W1891272071 @default.
• W4313342836 cites W1894439495 @default.
• W4313342836 cites W1979263599 @default.
• W4313342836 cites W2022166150 @default.
• W4313342836 cites W2068470708 @default.
• W4313342836 cites W2094728533 @default.
• W4313342836 cites W2121269107 @default.
• W4313342836 cites W2132839294 @default.
• W4313342836 cites W2142898321 @default.
• W4313342836 cites W2148721079 @default.
• W4313342836 cites W2251079237 @default.
• W4313342836 cites W2251287417 @default.
• W4313342836 cites W2251289180 @default.
• W4313342836 cites W2511149293 @default.
• W4313342836 cites W2739716023 @default.
• W4313342836 cites W2913318911 @default.
• W4313342836 cites W2962739339 @default.
• W4313342836 cites W2964212344 @default.
• W4313342836 cites W2981120761 @default.
• W4313342836 cites W3046843033 @default.
• W4313342836 cites W3090028958 @default.
• W4313342836 cites W3099235767 @default.
• W4313342836 cites W3102844651 @default.
• W4313342836 cites W4205247775 @default.
• W4313342836 cites W80422862 @default.
• W4313342836 doi "https://doi.org/10.1109/bigmm55396.2022.00012" @default.
• W4313342836 hasPublicationYear "2022" @default.
• W4313342836 type Work @default.
• W4313342836 citedByCount "0" @default.
• W4313342836 crossrefType "proceedings-article" @default.
• W4313342836 hasAuthorship W4313342836A5078942833 @default.
• W4313342836 hasAuthorship W4313342836A5090802966 @default.
• W4313342836 hasConcept C111472728 @default.
• W4313342836 hasConcept C115925183 @default.
• W4313342836 hasConcept C121332964 @default.
• W4313342836 hasConcept C124101348 @default.
• W4313342836 hasConcept C138885662 @default.
• W4313342836 hasConcept C154945302 @default.
• W4313342836 hasConcept C165801399 @default.
• W4313342836 hasConcept C195324797 @default.
• W4313342836 hasConcept C204321447 @default.
• W4313342836 hasConcept C2129575 @default.
• W4313342836 hasConcept C23123220 @default.
• W4313342836 hasConcept C25343380 @default.
• W4313342836 hasConcept C25810664 @default.
• W4313342836 hasConcept C2777601683 @default.
• W4313342836 hasConcept C41008148 @default.
• W4313342836 hasConcept C41608201 @default.
• W4313342836 hasConcept C41895202 @default.
• W4313342836 hasConcept C44291984 @default.
• W4313342836 hasConcept C4554734 @default.
• W4313342836 hasConcept C5655090 @default.
• W4313342836 hasConcept C62520636 @default.
• W4313342836 hasConcept C66322947 @default.
• W4313342836 hasConcept C96711827 @default.
• W4313342836 hasConceptScore W4313342836C111472728 @default.
• W4313342836 hasConceptScore W4313342836C115925183 @default.
• W4313342836 hasConceptScore W4313342836C121332964 @default.
• W4313342836 hasConceptScore W4313342836C124101348 @default.
• W4313342836 hasConceptScore W4313342836C138885662 @default.
• W4313342836 hasConceptScore W4313342836C154945302 @default.
• W4313342836 hasConceptScore W4313342836C165801399 @default.
• W4313342836 hasConceptScore W4313342836C195324797 @default.
• W4313342836 hasConceptScore W4313342836C204321447 @default.
• W4313342836 hasConceptScore W4313342836C2129575 @default.
• W4313342836 hasConceptScore W4313342836C23123220 @default.
• W4313342836 hasConceptScore W4313342836C25343380 @default.
• W4313342836 hasConceptScore W4313342836C25810664 @default.
• W4313342836 hasConceptScore W4313342836C2777601683 @default.
• W4313342836 hasConceptScore W4313342836C41008148 @default.
• W4313342836 hasConceptScore W4313342836C41608201 @default.
• W4313342836 hasConceptScore W4313342836C41895202 @default.
• W4313342836 hasConceptScore W4313342836C44291984 @default.
• W4313342836 hasConceptScore W4313342836C4554734 @default.
• W4313342836 hasConceptScore W4313342836C5655090 @default.
• W4313342836 hasConceptScore W4313342836C62520636 @default.
• W4313342836 hasConceptScore W4313342836C66322947 @default.
• W4313342836 hasConceptScore W4313342836C96711827 @default.
• W4313342836 hasLocation W43133428361 @default.
• W4313342836 hasOpenAccess W4313342836 @default.
• W4313342836 hasPrimaryLocation W43133428361 @default.
• W4313342836 hasRelatedWork W2087943365 @default.
• W4313342836 hasRelatedWork W2121333271 @default.
• W4313342836 hasRelatedWork W2133976672 @default.
• W4313342836 hasRelatedWork W2150908465 @default.
• W4313342836 hasRelatedWork W2573309543 @default.
• W4313342836 hasRelatedWork W2613467614 @default.
• W4313342836 hasRelatedWork W3090028958 @default.
• W4313342836 hasRelatedWork W3134565026 @default.

• next