Matches in SemOpenAlex for { <https://semopenalex.org/work/W4384644509> ?p ?o ?g. }
- W4384644509 abstract "Modeling discourse -- the linguistic phenomena that go beyond individual sentences, is a fundamental yet challenging aspect of natural language processing (NLP). However, existing evaluation benchmarks primarily focus on the evaluation of inter-sentence properties and overlook critical discourse phenomena that cross sentences. To bridge the gap, we propose Disco-Bench, a benchmark that can evaluate intra-sentence discourse properties across a diverse set of NLP tasks, covering understanding, translation, and generation. Disco-Bench consists of 9 document-level testsets in the literature domain, which contain rich discourse phenomena (e.g. cohesion and coherence) in Chinese and/or English. For linguistic analysis, we also design a diagnostic test suite that can examine whether the target models learn discourse knowledge. We totally evaluate 20 general-, in-domain and commercial models based on Transformer, advanced pretraining architectures and large language models (LLMs). Our results show (1) the challenge and necessity of our evaluation benchmark; (2) fine-grained pretraining based on literary document-level training data consistently improves the modeling of discourse information. We will release the datasets, pretrained models, and leaderboard, which we hope can significantly facilitate research in this field: https://github.com/longyuewangdcu/Disco-Bench." @default.
- W4384644509 created "2023-07-19" @default.
- W4384644509 creator A5003642180 @default.
- W4384644509 creator A5016804474 @default.
- W4384644509 creator A5016927477 @default.
- W4384644509 creator A5031126264 @default.
- W4384644509 creator A5039701470 @default.
- W4384644509 creator A5040122577 @default.
- W4384644509 creator A5061528283 @default.
- W4384644509 creator A5087920747 @default.
- W4384644509 creator A5088191810 @default.
- W4384644509 creator A5089949383 @default.
- W4384644509 date "2023-07-16" @default.
- W4384644509 modified "2023-10-17" @default.
- W4384644509 title "Disco-Bench: A Discourse-Aware Evaluation Benchmark for Language Modelling" @default.
- W4384644509 doi "https://doi.org/10.48550/arxiv.2307.08074" @default.
- W4384644509 hasPublicationYear "2023" @default.
- W4384644509 type Work @default.
- W4384644509 citedByCount "0" @default.
- W4384644509 crossrefType "posted-content" @default.
- W4384644509 hasAuthorship W4384644509A5003642180 @default.
- W4384644509 hasAuthorship W4384644509A5016804474 @default.
- W4384644509 hasAuthorship W4384644509A5016927477 @default.
- W4384644509 hasAuthorship W4384644509A5031126264 @default.
- W4384644509 hasAuthorship W4384644509A5039701470 @default.
- W4384644509 hasAuthorship W4384644509A5040122577 @default.
- W4384644509 hasAuthorship W4384644509A5061528283 @default.
- W4384644509 hasAuthorship W4384644509A5087920747 @default.
- W4384644509 hasAuthorship W4384644509A5088191810 @default.
- W4384644509 hasAuthorship W4384644509A5089949383 @default.
- W4384644509 hasBestOaLocation W43846445091 @default.
- W4384644509 hasConcept C104054115 @default.
- W4384644509 hasConcept C119599485 @default.
- W4384644509 hasConcept C119857082 @default.
- W4384644509 hasConcept C121332964 @default.
- W4384644509 hasConcept C127413603 @default.
- W4384644509 hasConcept C128942645 @default.
- W4384644509 hasConcept C13280743 @default.
- W4384644509 hasConcept C137293760 @default.
- W4384644509 hasConcept C138885662 @default.
- W4384644509 hasConcept C151552104 @default.
- W4384644509 hasConcept C152877465 @default.
- W4384644509 hasConcept C154945302 @default.
- W4384644509 hasConcept C165801399 @default.
- W4384644509 hasConcept C166957645 @default.
- W4384644509 hasConcept C178790620 @default.
- W4384644509 hasConcept C185592680 @default.
- W4384644509 hasConcept C185798385 @default.
- W4384644509 hasConcept C204321447 @default.
- W4384644509 hasConcept C205649164 @default.
- W4384644509 hasConcept C2777530160 @default.
- W4384644509 hasConcept C2781181686 @default.
- W4384644509 hasConcept C41008148 @default.
- W4384644509 hasConcept C41895202 @default.
- W4384644509 hasConcept C62520636 @default.
- W4384644509 hasConcept C66322947 @default.
- W4384644509 hasConcept C79581498 @default.
- W4384644509 hasConcept C95457728 @default.
- W4384644509 hasConceptScore W4384644509C104054115 @default.
- W4384644509 hasConceptScore W4384644509C119599485 @default.
- W4384644509 hasConceptScore W4384644509C119857082 @default.
- W4384644509 hasConceptScore W4384644509C121332964 @default.
- W4384644509 hasConceptScore W4384644509C127413603 @default.
- W4384644509 hasConceptScore W4384644509C128942645 @default.
- W4384644509 hasConceptScore W4384644509C13280743 @default.
- W4384644509 hasConceptScore W4384644509C137293760 @default.
- W4384644509 hasConceptScore W4384644509C138885662 @default.
- W4384644509 hasConceptScore W4384644509C151552104 @default.
- W4384644509 hasConceptScore W4384644509C152877465 @default.
- W4384644509 hasConceptScore W4384644509C154945302 @default.
- W4384644509 hasConceptScore W4384644509C165801399 @default.
- W4384644509 hasConceptScore W4384644509C166957645 @default.
- W4384644509 hasConceptScore W4384644509C178790620 @default.
- W4384644509 hasConceptScore W4384644509C185592680 @default.
- W4384644509 hasConceptScore W4384644509C185798385 @default.
- W4384644509 hasConceptScore W4384644509C204321447 @default.
- W4384644509 hasConceptScore W4384644509C205649164 @default.
- W4384644509 hasConceptScore W4384644509C2777530160 @default.
- W4384644509 hasConceptScore W4384644509C2781181686 @default.
- W4384644509 hasConceptScore W4384644509C41008148 @default.
- W4384644509 hasConceptScore W4384644509C41895202 @default.
- W4384644509 hasConceptScore W4384644509C62520636 @default.
- W4384644509 hasConceptScore W4384644509C66322947 @default.
- W4384644509 hasConceptScore W4384644509C79581498 @default.
- W4384644509 hasConceptScore W4384644509C95457728 @default.
- W4384644509 hasLocation W43846445091 @default.
- W4384644509 hasOpenAccess W4384644509 @default.
- W4384644509 hasPrimaryLocation W43846445091 @default.
- W4384644509 hasRelatedWork W1485630101 @default.
- W4384644509 hasRelatedWork W2028665553 @default.
- W4384644509 hasRelatedWork W2081245617 @default.
- W4384644509 hasRelatedWork W2094555469 @default.
- W4384644509 hasRelatedWork W2491561177 @default.
- W4384644509 hasRelatedWork W2492287160 @default.
- W4384644509 hasRelatedWork W2801693842 @default.
- W4384644509 hasRelatedWork W3204955359 @default.
- W4384644509 hasRelatedWork W4221142186 @default.
- W4384644509 hasRelatedWork W4310844685 @default.
- W4384644509 isParatext "false" @default.
- W4384644509 isRetracted "false" @default.