Hang Dong

Hi anantzoid, Thanks for sharing. The current link for the report seems not working. Would it be possible to update the link? Best wishes, Acadtags

Are uw and us global weights? just to conform.

1

Thank you ematvey for this paper. I wonder the uw and us are two vectors as global weights, or there are different uw(s) for each sentence, and different us(s) for...

Is the embedding initialized with a pre-trained one?

2

From the code it seems the embedding is not initialized with a pre-trained embedding (i.e. word2vec), although in the paper it says so. Am I right or I missed something?...

I had a general look at the BERT code. If I am not wrong, the BERT model in this folder is simply trained from the data itself, rather than transferred...

The change will allow the wrapper to construct java command that contains both MEKA and WEKA parameters. the below function will be able to work, which can specify SMO parameters,...

Wondered if we can fully reproduce the log_reg.py? It seems that something missing here. Thanks! Best, A

Number of classes dropped after the transition

4

Thanks for publishing the tool. It seems that the number of classes drops after the transition to an owl ontology? For example for `SNOMEDCT_US_20170301`, `332210` classes were kept after getting...

Biencoder with GPU RuntimeError: Expected object of device type cuda but got device type cpu for argument #3 'index' in call to _th_index_select

3

Hi, thanks for your code! When I set the `no_cuda` in `biencoder_wiki_large.json` as True, then I run `python blink/run_benchmark.py`, it returns an error, please see below, is there anything that...

about candidate generation

2

Hi Lihu, this is very good work. I have some questions when I try to adapt it to another dataset. The candidate generation seems not very straightforward, like how to...