aiwabdn

3 ms·

1. 1.▲
   by aiwabdn 6y ago

   Possibly some scenarios in RL where online learning is involved
2. 2.▲
   by aiwabdn 6y ago

   Main reason for now is that the cosine similarity used as context function is not suitable for natural images like those in Imagenet. We are experimenting some extensions based on convolutions for such use cases.
3. 3.▲
   Intuition and Implementations for Gated Linear Networks(aiwabdn.github.io)

   3 points by aiwabdn 6y ago|5 comments
4. 4.▲
   by aiwabdn 6y ago

   This blog post details our interpretation of Gated Linear Networks, a new family of neural networks proposed by researchers at DeepMind. Implementations in JAX, NumPy, PyTorch and Tensorflow are here: https://github.com/aiwa
5. 5.▲
   by aiwabdn 6y ago

   - We have come up with implementations of GLN from the paper in NumPy, PyTorch, TensorFlow and JAX. Check it out here: https://github.com/aiwabdn/pygln - Comments, feedback, pointers, use-case suggestions are all very
6. 6.▲
   DeepMind's Gated Linear Networks in NumPy, PyTorch, TensorFlow and JAX(github.com)

   6 points by aiwabdn 6y ago|1 comments
7. 7.▲
   by aiwabdn 8y ago

   I moved to London from India to do a masters and then found a job. The market definitely tries to capitalise on the visa requirement. They will try to get you at lower salaries, £40-50k. That being said, comparing it with the US market is q