Hello everyone,
I would request you to guide me how to annotate a large text dataset based on a small and manually annotated dataset. I used self-learning technique but it is giving a poor results after expanding the manually annotated dataset.
Kindly guide me if there are another techniques rather than the crowdsourcing
Have a look at deeply over : https://mlatgt.blog/2019/09/10/overcoming-large-scale-annotation-requirements-for-understanding-videos-in-the-wild/
https://medium.com/pytorch/training-faster-with-large-datasets-using-scale-and-pytorch-946dfe774d8c https://openaccess.thecvf.com/content_CVPR_2019/papers/Niitani_Sampling_Techniques_for_Large-Scale_Object_Detection_From_Sparsely_Annotated_Objects_CVPR_2019_paper.pdf
Progressive Self Label Correction (ProSelfLC) for missing labels (semisupervised learning)
https://xinshaoamoswang.github.io/blogs/2020-06-07-Progressive-self-label-correction/
.
if you have enough time and expert knowledge to code a lot of "labelling fiunctions", you could consider Snorkel
https://www.snorkel.org/
.
I think crowdsourcing will be the best option in today's scenario for collecting datasets.
- Badges
- Science topic