Abhishek Roy
408 posts

Abhishek Roy retweetledi

@MarcJBrooker There has been a lot of work on this problem! Check out @carlo_curino's work on Schism dspace.mit.edu/handle/1721.1/… + follow-up papers by different groups.
English

@conor_power23 Hmm, not sure. In the industry, ultimately customer has to pay. Difficult to fool someone who is paying money for a sustained period.
English

@AlekhJindal On the other hand, we can do 1000/60/30 only on int64 😀
English

Interesting:
“while it’s possible to represent a timestamp with an int64, it’s more advantageous to use the native Arrow Timestamp type”
ApacheArrow@ApacheArrow
Using Apache Arrow to store structured open telemetry data efficiently and compactly: arrow.apache.org/blog/2023/04/1…
English

@polylogblog b) with a little bit of prioritization of recently overdue small tasks? Along the lines of Nudge arxiv.org/abs/2106.01492
English

@elonmusk I changed my mind and would like Parag to come back.
English

@emollick And we can listen to songs at CHM archives --
computerhistory.org/collections/ca…
English

Corporate culture used to be far weirder. Look no further than corporate songs of the 30s. The IBM songbook has pages of songs in praise of the president of IBM (and the corporate treasurer of IBM!).
Here is a fascinating academic article on the songbook: journals.sagepub.com/doi/pdf/10.117…



English
Abhishek Roy retweetledi
Abhishek Roy retweetledi

@TaliaRinger So, what went well? The first wave definitely brought with it worthwhile evaluation practices. I think there is also a lot of scientific interest in the interplay between manual data analysis (theory, annotation) and pattern recognition applied to large datasets.
4/
English

Delighted to be part of three publications at VLDB 2021 (SparkCruise, AutoExecutor, Phoebe) and one best demo award 🏆!
Torsten Grust@Teggy
🥁 The @vldb2021 Best Demo Award goes to "AutoExecutor: Predictive Parallelism for Spark SQL Queries" (R. Sen, A. Roy, A. Jindal, R. Fang, J. Zheng, X. Liu, R. Li) from Microsoft. vldb.org/pvldb/vol14/p2… Congratulations! Aaaand that's a wrap for the VLDB 2021 Demo Track.
English

🥁 The @vldb2021 Best Demo Award goes to "AutoExecutor: Predictive Parallelism for Spark SQL Queries" (R. Sen, A. Roy, A. Jindal, R. Fang, J. Zheng, X. Liu, R. Li) from Microsoft.
vldb.org/pvldb/vol14/p2…
Congratulations!
Aaaand that's a wrap for the VLDB 2021 Demo Track.
Rottenburg am Neckar, Deutschland 🇩🇪 English

Congratulations @AlekhJindal, @virtuabhi!!
Torsten Grust@Teggy
🥁 The @vldb2021 Best Demo Award goes to "AutoExecutor: Predictive Parallelism for Spark SQL Queries" (R. Sen, A. Roy, A. Jindal, R. Fang, J. Zheng, X. Liu, R. Li) from Microsoft. vldb.org/pvldb/vol14/p2… Congratulations! Aaaand that's a wrap for the VLDB 2021 Demo Track.
English
Abhishek Roy retweetledi

Happy 75th Independence Day, India! Wishing everyone a safe and happy celebration:)
Many thanks to our guest artist Sayan Mukherjee for this year's beautiful doodle! google.com/doodles/india-…

English

@GidMK What do you think about the quality of analysis in this paper?
It seems to have lots of correction factors, using data from 2018, comparing to IFRs from other countries without basis, multiplying numbers without considering temporal/spatial patterns...
English

Interesting new paper from @giridar100 arguing that India's COVID-19 undercount was *always* likely because of systematic issues with death reporting
Estimates that the true death toll in India is likely to be 5.5-11x higher
frontiersin.org/articles/10.33…
English











