There is no correct answer to How to Publish your first Deep Learning Paper?
Here is how I'd do this, if I were to start now:
- Get comfortable with the Deep Learning basics ideas and tools.
- What sort of ideas? You should understand basic usage of the concepts such as AdamOptimizer, Backpropagation, LSTM/RNNs, Sequence to Sequence, CNNs and preferably even Generative Adversarial Networks (GANs).
- What sort of tools? I mostly mean programming tools and heuristics. E.g. Tensorflow and PyTorch and probably good starting points. If you are a complete noob, I recommend Tensorflow. E.g. Anyone in GANs should have looked at this from a FAIR guy: soumith/ganhacks: starter from "How to Train a GAN?" at NIPS2016