Dataset Information

Generating Classical Chinese Poems from Vernacular Chinese.

ABSTRACT: Classical Chinese poetry is a jewel in the treasure house of Chinese culture. Previous poem generation models only allow users to employ keywords to interfere the meaning of generated poems, leaving the dominion of generation to the model. In this paper, we propose a novel task of generating classical Chinese poems from vernacular, which allows users to have more control over the semantic of generated poems. We adapt the approach of unsupervised machine translation (UMT) to our task. We use segmentation-based padding and reinforcement learning to address under-translation and over-translation respectively. According to experiments, our approach significantly improve the perplexity and BLEU compared with typical UMT models. Furthermore, we explored guidelines on how to write the input vernacular to generate better poems. Human evaluation showed our approach can generate high-quality poems which are comparable to amateur poems.

SUBMITTER: Yang Z

PROVIDER: S-EPMC7255431 | biostudies-literature | 2019 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Generating Classical Chinese Poems from Vernacular Chinese.

Yang Zhichao Z Cai Pengshan P Feng Yansong Y Li Fei F Feng Weijiang W Chiu ElenaSuet-Ying EY Yu Hong H

Proceedings of the Conference on Empirical Methods in Natural Language Processing. Conference on Empirical Methods in Natural Language Processing 20191101

Classical Chinese poetry is a jewel in the treasure house of Chinese culture. Previous poem generation models only allow users to employ keywords to interfere the meaning of generated poems, leaving the dominion of generation to the model. In this paper, we propose a novel task of generating classical Chinese poems from vernacular, which allows users to have more control over the semantic of generated poems. We adapt the approach of unsupervised machine translation (UMT) to our task. We use segm ...[more]

PMID: 32467928

Similar Datasets

Project description:Objective. Bone lesion is crucial for diagnosing and management of polyneuropathy, organomegaly, endocrinopathy, monoclonal protein, and skin change (POEMS) syndrome, a rare plasma cell disorder. This study is to compare the effectiveness of X-ray skeletal survey (SS) and computed tomography (CT) for detecting bone lesions in Chinese POEMS syndrome patients, and to investigate the relationship between bone lesion features and serum markers. Methods. SS and chest/abdomen/pelvic CT images of 38 Chinese patients (26 males, 12 females, aged 21-70 years) with POEMS syndrome recruited at our medical center between January 2013 and January 2015 were retrospectively analyzed. Bone lesions identified by CT were further categorized according to the size (<5 mm, 5-10 mm, >10 mm) and appearance (osteosclerotic, lytic, mixed). The percentage of plasma cells in bone marrow smears, type of immunoglobulin, platelet (Plt), and levels of serum bone metabolic markers and inflammatory factors including alkaline phosphatase (ALP), calcium, phosphate, parathyroid hormone (PTH), beta-isomerized C-telopeptide (β-CTx), vascular endothelial growth factor (VEGF), and interleukin (IL)-6 levels were also recorded. Results. Of the 38 POEMS syndrome patients, the immunoglobulin heavy chain isotypes were IgA in 25 patients (65.8%; 25/38) and IgG in 13 patients (34.2%; 13/38), and the light chain isotypes were λ in 35 patients (92.1%; 35/38) and κ in 3 patients (7.9%; 3/38). There were 23 patients with thrombocytosis. More patients with bone lesions were detected by CT than by SS (97.4% vs. 86.8%). The most commonly affected location was the pelvis (89.5%), followed by the spine, clavicle/scapula/sternum/ribs, skull, and long bones. Of the 38 POEMS syndrome patients, 35 (94.6%) had osteosclerotic and 32 (86.5%) had mixed lesions. Osteosclerotic lesions were typically scattered, variable in size, and plaque-like, whereas mixed lesions were pouch-shaped or soup bubble-like with a clear sclerotic margin and were generally larger. Although the majority of bone lesions were small in size, 23 (62.2%) had at least one lesion >10 mm. There was no correlation between serum marker levels and bone lesion patterns after Bonferroni correction (all P > 0.001). Conclusions. CT is more sensitive and accurate than SS in detecting bone lesions in POEMS syndrome.

Dataset Information

Generating Classical Chinese Poems from Vernacular Chinese.

Publications

Generating Classical Chinese Poems from Vernacular Chinese.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets