ramybaly/article-bias-prediction

article-bias-prediction's Issues

Dataset increased from EMNLP paper

Hi,
Thank you for your dataset. I just wanted to say that this dataset is slightly different from the one described in your paper.
Probably it's because you have been adding newer articles, but it would be nice to state this information in the README file.

I was trying to reproduce the experiments and, for comparability, it would be nice to be able to do this on the same exact data.

	EMNLP	GitHub
total size	34737	37554
train (media)	22,969	26,590
validation (media)	5,098	2,356
test (media)	1,200	1,300
train (random)	26,828	27,978
validation (random)	6,709	6,996
test (random)	1,200	1,300
Left	...
Center	...
Right	...

Best,
Martino

Source code for crawling and annotation

Could you please share source codes for crawling and annotation @ramybaly

Will BERT models be published on HuggingFace model hub?

Thanks for the great article!
Have you considered publishing your BERT Twitter bios+Article with TLP on the HuggingFace model hub? Would be great to have a pre-trained model for detecting political orientation of articles available for the community and can imagine that people would make good use of it. (Would also be good advertisement for your model and research :) )

The time of releasing code

Hi,

Thank you so much for your awesome research work.

I am wondering about the time of releasing the code.

Thanks in advanced and regards,
Qin

timeline for dataset and code

Hi,
I was curious to take a look at the data and code. Is there any predicted day it will be uploaded?
Thank you.

another paper that never released the data

It's a common trend in CS research where a paper says that the data or the code will be released, and an empty repo is created to support that conclusion. This helps authors get favorable reviews, and to get accepted in prestigious conferences. Then the code or data are never released, and the authors pretend as they have never promised anything. Unfortunately, it looks like the authors of this paper might be doing just that.

Many more news sources than reported in the paper

Hi Ramy,

Thank u for publishing this amazing paper.

It seems in this published dataset there are 192 news sources, many more than 73 reported in paper. After manually merging some similar sources there are still more than 170 sources. I wonder if you could release the exact same dataset that is reported in the paper. Thanks a lot!

Best,
Zihao