Yesterday, I gave my first-ever machine learning talk at a local meetup in Munich. For those who have been following my path to becoming a data scientist, this seems like something I couldn’t even imagine doing about a year ago.

Talking about black-and-white photo colorization using neural nets

Yet, there I was, sweating nervously, yet, giving my best, trying to explain how colorizing black-and-white imagery using neural networks works. Judging by the reaction of the audience, the presentation must have been good, because a few people caught me aside afterwards, wanting to know more details about my project, as well as ways they could contribute. I found this reaction a sort of a confirmation that my work has not been in vain, but just the opposite. As my motto goes, learn and share. I think that with this project, I have somehow managed to do both.

I am really looking forward to an opportunity to present the next steps of my research. If you are reading this and know of a meetup, workshop, or a conference that would be glad to have me speak, please, don’t hesitate to drop me a line on Twitter.

As for the project, there will be a series of articles describing in detail different aspects of the colorization process. So far, I can share a link to the GitHub repository:

It is a little disorganized, but I am hoping that with a little more time and push from the community, I will manage to clean up my thoughts and code. Feel free to send me issues, suggestions, or pull requests. I’ll be more than happy to check them out.

I have also put the slides from my talk online.

About the event

Hacking Machine Learning is a monthly meetup event in Munich, where ML professionals and enthusiasts gather to share their ideas with the community, or simply grab a beer and chat with like-minded friends. Anyone with an interesting idea can tweet them and get their idea validated in front of a welcoming audience.

Leave a comment