uhh-lt poincare issues - Githubissues

uhh-lt / poincare

Using sense representations to improve detection of compositional nominal compounds

0 stars 0 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Mixed prediction experiment setup

#19 adhaesitadimo closed 5 years ago
0
Try the original poincare metrics

#18 alexanderpanchenko opened 5 years ago
0
Combine two similarity scores (poincare and word embeddings)

#17 alexanderpanchenko opened 5 years ago
0
Compute the results on Reddy-Nouns

#16 alexanderpanchenko opened 5 years ago
2
New vector embeddings

#15 adhaesitadimo closed 6 years ago
0
SenseGram on the new 100 dimensional

#14 alexanderpanchenko opened 6 years ago
5
Lemmatized ukWaC

#13 alexanderpanchenko closed 6 years ago
0
Try senses with the supervised model

#12 alexanderpanchenko opened 6 years ago
0
Try Poincare embeddings

#11 alexanderpanchenko opened 6 years ago
16
Instead of using addition using PCA

#10 alexanderpanchenko closed 5 years ago
0
Using angular distance instead of cosine distance

#9 alexanderpanchenko closed 5 years ago
0
A dataset for Russian - a tentative shared task on its basis

#8 alexanderpanchenko opened 6 years ago
0
Use both capital and lower senses as candidates (for max)

#7 alexanderpanchenko opened 6 years ago
0
Shared task Chris 2011 in context

#6 alexanderpanchenko opened 6 years ago
0
Examples where it is improved using senses

#5 alexanderpanchenko closed 5 years ago
0
Make an overview of the related work

#4 alexanderpanchenko opened 6 years ago
3
Use more recent sense and context embeddings, e.g. MUSE and ELMo

#3 alexanderpanchenko opened 6 years ago
0
Use supervised learning as the baseline approaches

#2 alexanderpanchenko opened 6 years ago
8
Use the new vector based algorithm for WSI

#1 alexanderpanchenko closed 6 years ago
0