• Martin Thoma
  • Home
  • Categories
  • Tags
  • Archives
  • Support me

New Year 2019

Contents

  • New Year 2019
    • Review of 2018
      • Old year resolutions
      • Old year predictions
    • New year resolutions
    • Finishing things
    • Predictions for 2019
    • See also

It's new years eve and - as always - I try to finish some things and have some plans for next year.

Review of 2018

Not a lot happened about which I want to write publicly, so here are some other mentionable facts:

  • ML-KA is still active.
  • I have 40 583 points on StackOverflow and reached 22.4 million people with my posts (source).
  • I wrote only 48 articles for this blog (including this one, not counting 13 ones which were pulished yesterday) in 2018. A couple of them might be interesting:
    • Nepal
    • Datetime - or even better my short summary What every developer should know about time
    • A couple about details in Python such as packaging, str vs repr, property, testing, style guide
    • A couple of shorter data science related ones, such as an overview, a project guide, the problems of Feature Importance, time series forecasting
    • Only one recipe!
  • I've uploaded 195 files to Wikipedia Commons in 2018 (2164 files in total, source). My uploads to Wikipedia Commons are about 6 GiB in total now. On commons, I made 5333 edits in total.

Interesting changes in my life:

  • I moved within Munich on 01.06.2017
  • Found a couple of new friends 🙂

Overall, I have to say that I spent a lot of time with working. Maybe too much time. I've noticed that I contributed a lot less to Wikipedia, wrote less blog articles of lower quality and spent less time on StackOverflow.

Old year resolutions

  • ✗ Participate in a Flashmob
  • ✔ Visiting a new country - I've been in the Czech Republic

Old year predictions

I forgot to make predictions last year 😨

New year resolutions

The plans for my new year are:

  • ✗ Get comfortable with Orange Bouldering routes (5c-6c) and finish at least 5 harder ones (6c - 7b)… yes, I'm aware that I had exactly that already. I wasn't bouldering for a while. I miss it.
  • [ ] Create a GAN for my HASYv2 dataset
  • ✔ Create at least 10 more blog posts with recipes: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
  • [ ] Create at least 10 high-quality Media files for Wikipedia: 1, 2, 3,

Finishing things

I always do some cleanup before the new year begins. This means especially cleaning my computer stuff of unnecessary trash.

  • ✓ Deleting several hundret E-mails (over 500 unread e-mails, 824 starred e-mails)
  • [ ] Clean up my smartphones TODO list.
  • [ ] Clean up my Desktop / Download folder.
  • ✓ Make Backup of Google+ and Gmail (takeout), WhatsApp (tutorial), Amazon, Banking. I skipped Facebook (tutorial) and Twitter (tutorial) this year.
  • ✓ Push git stuff
  • [ ] Remove things from my YouTube "Watch Later" list

Predictions for 2019

Machine Learning and Technology

  1. Tensorflow will keep being the dominant deep learning framework as measured by (1) StackOverflow questions, (2) Google Trend, (3) Github repositories
  2. There will still not be a DOTA playing AI which can replace a human DOTA player in a human team with the usual flexibility.
  3. We will see fraud cases related to ML (e.g. using speech synthesis for identity theft or tricking a ML system).

About Me

  1. ~ I will come back well from Nepal.
  2. I will write at least 10 summaries of papers on cs.CV.
  3. ✔ I will learn and fully understand how Flask Blueprints work.
  4. I will reach level 5 on the first 15 lessons for Spanish on Duolingo and level 3 on the next 13 ones. (So far: 6 on Level 5, 8 on Level 3)

The bigger Picture: Politics, Nature and the World

  1. ✔ There will be over 250 mass shootings in the US in 2019. I define a mass shooting as anything where at least 4 people get injured by a gun. (List of mass shootings in the United States in 2019)
  2. ✔ The government shutdown will last at least 20 days, but no more than 40 days. ✔
  3. ✔ Trump will still be in office

There are also a couple of predictions which I don't know how to check, e.g.

  • The hype of Machine Learning and data science will go down a bit as companies are faced with problems / the reality. It will still be important, though.
  • Amazing results by GANs. Large Scale Gan Training for High Fidelity Natural Image Synthesis is already pretty awesome. I expect more of this, easier to use for non-ml developers.

See also

  • New Year 2016
  • New Year 2017
  • New Year 2018
  • New Year 2019
  • New Year 2020

Published

Dez 31, 2018
by Martin Thoma

Category

Cyberculture

Tags

  • Cleaning 5
  • New Years Eve 5

Contact

  • Martin Thoma - A blog about Code, the Web and Cyberculture
  • E-mail subscription
  • RSS-Feed
  • Privacy/Datenschutzerklärung
  • Impressum
  • Powered by Pelican. Theme: Elegant by Talha Mansoor