@BicksonDanny

2× Exits | VLDB Test of Time Award | Building & Investing in AI

 Israel
Joined January 2023
  • Tweets8
  • Following5
  • Followers11
1Media
Hot take: Most academic spinouts fail because they inherit the wrong optimization function. Academic research optimizes for what's publishable. Startups optimize for problems customers will pay to solve.
14
Danny Bickson retweeted
Visual Layer@visual_layer11 Aug 2023
Here are some of the quality issues that you may find: ‣ Duplicates ‣ Outliers ‣ Mislabels ‣ Corrupted images ‣ Train/test leakage ‣ Overly bright/dark/blurry images Notebooks: ‣ Kaggle Notebook - lnkd.in/gs8mc6Kx ‣ Colab Notebook - lnkd.in/gTHYJqdE
1
1
1
64
See our paper for a lot more technical details and results. Speaking personally, I have many thoughts on this paper. First, everyone should de-duplicate their data as it reduces memorization. However, we can still extract non-duplicated images in rare cases! [6/9]
6
15
5
458
102,042
Load more