Deepan Das (@dd_deepan) header
Deepan Das (@dd_deepan) profile picture

Can train your dragon if given a nice loss landscape.

Los Angeles, CA
Joined January 2010
Photo by Deepan Das (@dd_deepan)
Deepan Das retweeted
Join the @datachatAI team this August in San Diego at the @TDWI Data and Analytics Conference. Use the code DataCha15 and save 15% of registration. tdwi.org/events/conferences/… #data #analytics #tdwi #conference
This is hands down the greatest talk on Machine Learning ever made. It was a keynote at ICLR and is now public! Thank you @_beenkim for being such an incredible person and always taking your time to articulate your thoughts so well. Watch it at: youtube.com/watch?v=Ub45cGEc… 👀‼️
9
140
5
761
Deepan Das retweeted
Congrats!! @datachatai, @katanagraph, @posthog, @coiledhq, @decodableco, @eversightlabs, @rasgoml have made to Emerging Startups 2022: Top Big Data Analytics Startups! tracxn.com/d/emerging-startu… #cloud #AI #data #emergingtech #startups #soonicorns #unicorn #tech
Deepan Das retweeted
prn.to/3DOA5DH Delighted to announce the closing of our $25M Series A round led by @RedlineCapital @anthoscapital, along with @CelestaCapital, Nepenthe and @WARF_News. HUGE thanks to our customers, and @HazyResearch, #DavidDewitt, @samrmadden for their help and support.
1
6
2
18
Plenty among the world's rich & powerful enjoy careers predicated on a combination of luck and survivorship bias. But even among this crowd, @NateSilver538 stands out as the very archetype of such a beneficiary. The man makes $100k/talk on this false aura.
10
2
1
55
The scikit-learn consortium (@sklearn_inria ) operational budget is approximately 3 times this. It drives me nuts that people are willing to pay that much for talks, but we are having a hard time supporting a core library that powers a good chunk of the data applications.
4
8
1
69
Suffering from overconfident softmax scores? Time to use energy scores! Excited to release our NeurIPS paper on "Energy-based Out-of-distribution Detection", a theoretically motivated framework for OOD detection. 1/n Paper: arxiv.org/abs/2010.03759 (w/ code included)
16
248
10
1,349
What are some recent research developments or directions in Computer Vision you've enjoyed following (or creating 😉) over the last few months/year? I would be happy to read a few papers.
27
24
2
211
At ICML next week, @KDziugaite @roydanroy @mcarbin and I will present Linear Mode Connectivity and the Lottery Ticket Hypothesis. We study the effect of SGD noise (like data order) on neural net optimization. Those results shed new light on lottery tickets arxiv.org/abs/1912.05671
3
50
8
259
Deepan Das retweeted
Our ML Research Engineer Antoine Cordier, will be giving a talk on his work on "active learning with weakly supervised signals" on Wednesday, June 17th! 🚀 Register to the web conference here : meetup.com/fr-FR/Bordeaux-Ma… #deeplearning #qualityintelligence
4
4
Deepan Das retweeted
My validation accuracy is 1.0
You’re unable to view this Post because this account owner limits who can view their Posts. Learn more
9
2
1
76
Looks like all of the ICLR 2020 virtual conference videos (and papers and posters) are now available to everyone: iclr.cc/virtual_2020/papers.…
87
6
305
CNN Explainer is an interactive visualization tool for learning purposes. It runs a pre-tained CNN in the browser and lets you explore the layers and operations: poloclub.github.io/cnn-expla… Video: youtube.com/watch?v=udVN7fPv… Code: github.com/poloclub/cnn-expl… Paper: arxiv.org/abs/2004.15004
4
169
9
608
Wrote a blog on automating the art of data augmentation, featuring latest works on the practice, theory and new direction of data augmentation from @HazyResearch. Check out on @StanfordAILab website: ai.stanford.edu/blog/data-au…
4
78
5
307
Deepan Das retweeted
Last week, our Sales Director @VandecasteleL illustrated during the “Tech’Days” in Lyon how our #QualityIntelligence Solution is designed to tackle the constraints linked to deploying a #deeplearning solution in the factory. Thank you @AllizePlast and @Plastipolis!
3
8
Deepan Das retweeted
When scientists put slime mold over a map of Tokyo, with food used to represent urban areas, and after a day the mold created a network nearly identical to Tokyo's rail network: all this without any brain ow.ly/7CA730o4Sjk
68
1,369
225
4,624
Oh here's a fun one: "Oops! Predicting Unintentional Action in Video" Great project site, check out the graphs showing ground-truth versus predictions for the videos: abs: arxiv.org/abs/1911.11206 pdf: arxiv.org/pdf/1911.11206.pdf site: oops.cs.columbia.edu/
7
77
12
288
Twenty-five years ago I bought a suitcase for a trip to the US, and marked it "GvR Python". Today I am back for six days in Amsterdam with the same suitcase. photos.app.goo.gl/mJePGzkrXi…
19
87
4
913
Deepan Das retweeted
Real-time data is one of the most powerful catalysts enabling greater growth and is the core aspect of growing #manufacturing operations. 81% of all manufacturers say real-time monitoring is improving their business. erpblog.iqms.com/real-time-m… #Industrialization #QualityControl
2
1
Deepan Das retweeted
We're happy to be one of the projects supported in this program. We'll use the grant to fund additional project maintenance and to cross off some of the items on our project roadmap: dev.pandas.io/about/roadmap.…
#Opensource software is essential to science—that’s why we’re supporting 42 open source tools that accelerate biomedical research and serve the larger community. Learn about the grantees bit.ly/34YhzX1
9
58