The Lottery Ticket Hypothesis: Finding Small, Trainable Neural Networks
Report a link problem
Submission timeline
2007–2026One slot for every year since HN launched. Height is that year's peak points; orange marks a 100+ point or 50+ comment breakout. Select a bar to open its strongest thread.
Top comments
Working on DNNs is more similar to gambling on the slot machines then it is to traditional programming. I gather training data, write a python script, and then pay around $1 for GPU time to see if it's a winner. Instead of watching spinning reels I watch loss function graphs. 95% of the time it is a bust, but every once in a while you get a big payoff. I feel very smart when I win, but deeper investigation inevitably…
Article from 2018/19 and this hypothesis remains just that afaik with plenty of evidence going against it
The first top-level comment from each of the four biggest threads, in HN’s own order. Excerpts are shortened; open a comment for full context.
HN stats
- Breakout years
- 2
- Total points
- 261
- Total comments
- 41
100+ points or 50+ comments
reference only — not used in Hall rules or ranking
reference only — not used in Hall rules or ranking
Every submission
| Date | Title as submitted | By | Points | Comments |
|---|---|---|---|---|
| 2018-03-19 | The Lottery Ticket Hypothesis: Training Pruned Neural Networks | jonbaer | 2 | 0 |
| 2018-07-05 | The Lottery Ticket Hypothesis: Finding Small, Trainable Neural NetworksFirst breakout · Best thread | jonbaer | 137 | 14 |
| 2020-10-23 | The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks (2019) | plat12 | 1 | 0 |
| 2026-01-02 | The Lottery Ticket Hypothesis: finding sparse trainable NNs with 90% less paramsLatest 20+ point return | felineflock | 121 | 27 |