Traudl
Re:Figma/Miro for Data – Einblick.ai
on Monday September 20, 2021 @02:03PM (#28596749)
This looks pretty interesting. Like a mix of Figma/Miro with the video bubbles from pitch.com
Re:Learned indexes was integrated Google Bigtable [pdf]
on Wednesday January 27, 2021 @07:52AM (#25927599)
Similar interesting is also the new Learned Index benchmark study: https://vldb.org/pvldb/vol14/p1-marcus.pdf
Re:Learned indexes was integrated Google Bigtable [pdf]
on Wednesday January 27, 2021 @07:50AM (#25927585)
Throughput increased 54% for point lookups and 56% for scans
Learned indexes was integrated Google Bigtable [pdf]
Re:PGM Indexes: Learned indexes that match B-tree performance with 83x...
on Wednesday January 27, 2021 @07:48AM (#25927557)
It looks like it found its way into Google Bigtable though: https://arxiv.org/abs/2012.12501
Re:A Recursive Model learned index structure
on Wednesday January 27, 2021 @07:45AM (#25927529)
Yeah, apparently the results do not hold up under fair evaluation. They used a not well tuned RMI implementation. See: https://vldb.org/pvldb/vol14/p1-marcus.pdf
That said, PGMs are still interesting given their theoretical guarantees.
Data systems that learn to be better
Re:Learned Multi-dimensional Indexes – Result with big impact for key/...
on Thursday December 05, 2019 @02:41PM (#21715699)
This could have huge implications for key/value stores and cloud storage systems as it provides a mechanism to access data by more than one key. The numbers to Amazon's z-order encoding also look very interesting. Of course, updates/inserts might still be a problem, but many of the other techniques (e.g, Amazon) are also static.