Finlay Maguire comments

Results 31 comments of


                                            Finlay Maguire

For full reproducibility, always report all details of your analyses, including sharing of data, methods, and models

Covered in Tip 1 and Tip 3. However, might want to add specific recommendation of notebooks?

Understand the data

Implicitly mentioned in https://github.com/Benjamin-Lee/deep-rules/blob/master/content/06.know-your-problem.md but might want to make more explicit, especially @brettbj's comment about having a hypothesis for why DL might work for your data.

Have access to adequate infrastructure

Seems best fit for the training complexities

Close issues that have been addressed

So I just went through our extant issues, I've closed any that I thought looked covered and assigned any still open ones to the most relevant project. A lot of...

Discuss computational issues

Tangentially referred to in https://github.com/Benjamin-Lee/deep-rules/blob/master/content/05.dl-complexities.md > Despite improved framework ease-of-use and on-demand cloud computing resources this means DL can be very demanding, often requiring significant infrastructure and patience to achieve...

New paper on DA for scRNA-seq

Assigned to tip mentioned by @rasbt

New paper on DA for scRNA-seq

I was just basing the project assignment for where this paper fitted best based on your comment above and @tbrittoborges WIP.

Don't forget your ML fundamentals/Rules that apply to ML also apply to DL

One area I think would be good to add to this is more explicitly: a trained ML model itself (almost always) isn't the scientifically interesting thing in and of itself....

Make sure you are optimizing the right objective function!

Could do with being more explicitly mentioned in the ML fundamentals tip.

Perform sanity checks and follow good coding practices

Sort of mentioned in tip 3, need to add this reference to it.