CLARIN-PL
Not a member yet
504 research outputs found
Sort by
plWordNet 4.5
No full textPLWordNet ver. 4.5 is a lexico-semantic network that reflects the lexical system of the Polish language with projection to the English language. Słowosieć, Princeton Wordnet, EnWordnet together. It is now the largest wordnet in the world and is still growing
Big data language model with part of speech tags stemmed in ARPA format
No full textBig data language model with part of speech tags stemmed in ARPA forma
Neural Language Models vs Wordnet-based Semantically Enriched Representation in CST Relation Recognition
Full text linkNeural language models, including transformer-based models, that are pretrained on very large corpora became a common way to represent text in various tasks, including recognition of textual semantic relations, e.g. Cross-document Structure Theory. Pretrained models are usually fine tuned to downstream tasks and the obtained vectors are used as an input for deep neural classifiers. No linguistic knowledge obtained from resources and tools is utilised. In this paper we compare such universal approaches with a combination of rich graph-based linguistically motivated sentence representation and a typical neural network classifier applied to a task of recognition of CST relation in Polish. The representation describes selected levels of the sentence structure including description of lexical meanings on the basis of the wordnet (plWordNet) synsets and connected SUMO concepts. The obtained results show that in the case of difficult relations and medium size training corpus semantically enriched text representation leads to significantly better results
Big data language model stemmed with BPE in RAW format
No full textBig data language model stemmed with BPE in RAW forma
PoLitBert_v50k_linear_50k - Polish RoBERTa model
No full textPolish RoBERTa model trained on Polish Wikipedia, Polish literature and Oscar
PoLitBert_v32k_tri_50k - Polish RoBERTa model
No full textPolish RoBERTa model trained on Polish Wikipedia, Polish literature and Oscar
Cleaned Polish Oscar corpus (128M lines)
No full textCleaned Polish Oscar corpus (part: 128M lines, 3.53 GB). Data was prepared with a few cleaning heuristics:
- remove sentences shorter than
- remove non-polish sentences
- remove ungrammatical sentences
- perform sentence tokenization and save each sentence in a new line, after each document the new line was adde
The procedure of the correction of plWordNet (ver. 1)
No full textThe pdf entails the specificationof tipical errors of lexicographic description of lexical units and synsets in plWordNet, and the procedure of them manual correction
