Pradyumna Singh | प्रद्युम्न @pradyu_singh

Data Science/Decision Science/Machine Learning. Tweets: Science, Literature, Art, Arch, Design. English/हिन्दी. SF Bay Area + India (via NL) Joined December 2011

Tweets

7K
Followers

421
Following

290
Likes

16K

Eric Horvitz @erichorvitz

4 years ago

Discussion on causal machine learning with @yudapearl @Susan_Athey Yoshua Bengio bit.ly/3HBh89m Panel at Microsoft Research Summit @MSFTResearch #causality #ML #DNN

7 60 278 0 67

Download Image

Pradyumna Singh | प्रद्युम्न @pradyu_singh

4 years ago

@trengriffin There's a fabulous video of Federico Faggin and Ted Hoff at the Computer History Museum reminiscing about how they (together with Stan Mazor and Masatoshi Shima of Busicom) designed and architected the 4004. youtube.com/watch?v=j00AUL…

1 6 9 0 3

Tren Griffin @trengriffin

4 years ago

"The world changed on Nov. 15, 1971, and hardly anyone noticed. It is the 50th anniversary of the launch of the Intel 4004 microprocessor..." wsj.com/articles/the-c…

6 72 269 0 39

Pradyumna Singh | प्रद्युम्न @pradyu_singh

4 years ago

“... Which is that we are a collection of fragile beings taking care of one another, and that those who do the lion’s share of this care work that keeps us alive are overtaxed, underpaid, and daily humiliated.” A very nice profile, incl. his last days :( nymag.com/intelligencer/…

0 0 0 0 0

Sameer Agarwal @sandwichmaker

4 years ago

Prof. Bradley Efron has written a lovely article about statistics in the age of deep learning. I hope this becomes essential reading for anyone engaged in machine learning. statweb.stanford.edu/~ckirby/brad/p…

10 148 635 0 339

Shubhendu Trivedi @_onionesque

4 years ago

Occam's shaver.

4 81 586 0 34

Download Image

cees dekker - also at @ceesdekker.bsky.social @cees_dekker

4 years ago

Today we publish a paper in @ScienceMagazine that expands nanopore readings to the proteome: a nanopore-based scanner to read off PROTEINS at the single-molecule level! 🤩 Awesome experiments by postdoc Henry Brinkerhoff of our #CDlab, with MD simulations of @aksimentievLab 1/x

61 1K 3K 0 290

Download Image

Pradyumna Singh | प्रद्युम्न @pradyu_singh

4 years ago

Going off Twitter is a bit like dying, except you can actually come back and see for yourself how beautifully the world carries on without you.

0 0 2 0 0

Pradyumna Singh | प्रद्युम्न @pradyu_singh

4 years ago

Stand by! Folks who won't tire of explaining the Halo Effect, will shortly explain Munger's architectural genius :-)

0 1 2 0 0

Pradyumna Singh | प्रद्युम्न @pradyu_singh

4 years ago

Does anyone use MATLAB anymore? (Honest question prompted by genuine nostalgia!)

2 0 1 0 0

Pradyumna Singh | प्रद्युम्न @pradyu_singh

4 years ago

Some nice work on effective design of decision support systems for high-stakes decisions - where increasingly the intuition of domain experts (developed over long years) has to contend with predictions from machine learning models. dl.acm.org/doi/10.1145/34…

0 0 2 0 0

Download Image

Pradyumna Singh | प्रद्युम्न @pradyu_singh

4 years ago

Some great recommendations for Decision Theory in this thread!

Dean Eckles @deaneckles

4 years ago

Some great recommendations for Decision Theory in this thread!

10 14 119 0 89

Download Image

0 0 1 0 1

Pradyumna Singh | प्रद्युम्न @pradyu_singh

4 years ago

Lovely resources (historical + pedagogical) on Braess' Paradox - the idea that adding extra road links between a source and destination can paradoxically *increase* overall travel time for travelers. An important result in transportation, networks, game theory, decision making.

Anna Nagurney @Supernetworks

4 years ago

1 1 7 0 1

0 0 4 0 0

André Spicer @andre_spicer

4 years ago

Forrester Shocks. Non-linear changes which are not created by external dynamics but unintended consequences of decisions made within a system. (Hat tip to @oliverbeige) sloanreview.mit.edu/article/jay-fo…

1 8 23 0 4

Pradyumna Singh | प्रद्युम्न @pradyu_singh

4 years ago

💯. The direct/combative style you're trained for in school isn't a universal virtue; only works if a) others share the same culture/style or b) there's implicit trust with other team members thru other means. Can often be ineffective+countr productive in large engg teams

Jeff Lundeen @LundeenOttawa

4 years ago

14 25 173 0 26

Download Image

0 0 4 0 1

Pradyumna Singh | प्रद्युम्न @pradyu_singh

4 years ago

Pretty good overview on the future of semiconductor scaling - incl. challenges around patterning GAA transistors, interconnects, 3D packaging, 3D SoCs etc. semiengineering.com/whats-next-for…

0 0 1 0 0

Ryan Petersen @typesfast

4 years ago

Yesterday I rented a boat and took the leader of one of Flexport's partners in Long Beach on a 3 hour of the port complex. Here's a thread about what I learned.