SSL4MIS
uda
SSL4MIS | uda | |
---|---|---|
2 | 2 | |
2,014 | 2,153 | |
2.6% | 0.0% | |
6.2 | 0.0 | |
10 months ago | over 2 years ago | |
Python | Python | |
MIT License | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
SSL4MIS
-
Researchers at Oxford University Propose a Machine Learning Framework Called ‘TriSegNet’ Based on Triple-View Feature Learning for Medical Image Segmentation
Continue reading | Check out the paper and github link
-
How to get image dataset annotated? Any idea?
Otherwise, you may be able to look into semi-supervised learning. Basically, you label a subset of your data, and use semi-supervised techniques to extrapolate and label the rest. This, of course, is a challenge in itself, but luckily this particular challenge has been researched a lot, so you may find something to get started with.
uda
-
BERT models: how resilient are they to typos?
Another thought is to do some data augmentation using back-translation, a la https://arxiv.org/abs/1904.12848
-
A Visual Survey of Data Augmentation in NLP
The words that replaces the original word are chosen by calculating TF-IDF scores of words over the whole document and taking the lowest ones. You can refer to the code implementation for this in the original paper here.
What are some alternatives?
cleanlab - The standard data-centric AI package for data quality and machine learning with messy, real-world data and labels.
transformers - 🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.
awesome-data-labeling - A curated list of awesome data labeling tools
nlpaug - Data augmentation for NLP
alibi-detect - Algorithms for outlier, adversarial and drift detection
clip-as-service - 🏄 Scalable embedding, reasoning, ranking for images and sentences with CLIP
cleanlab - The standard package for machine learning with noisy labels and finding mislabeled data. Works with most datasets and models. [Moved to: https://github.com/cleanlab/cleanlab]
contractions - Fixes contractions such as `you're` to `you are`
squirrel-datasets-core - Squirrel dataset hub
bert - TensorFlow code and pre-trained models for BERT
squirrel-core - A Python library that enables ML teams to share, load, and transform data in a collaborative, flexible, and efficient way :chestnut: