Sumit (@sumitppai) profile picture
India
Joined May 2009
Today, @RaiaHadsell, @kchonyc and I are happy to announce the creation of a new journal: Transaction on Machine Learning Research (TMLR) Learn more in our post: medium.com/@hugo_larochelle_…
48
647
123
2,417
Today is #WorldPrematurityDay - our partnership with @RotundaCharity and @SpecsaversIE aims to secure a RetCam for premature babies born at @RotundaHospital. Learn more & donate here: idonate.ie/fundraiser/114188…
We recently teamed up with @RotundaCharity and @SpecsaversIE to help secure a RetCam for the Rotunda NICU. With your support we can do this! If you have the means, please donate today to help us achieve our goal! Read more here: rotundafoundation.ie/retcam #RetCam #neonatal #nicu
5
7
Half of my patients in clinic today (former ex premature babies) are on crazy waiting lists to see other specialists as a result of unfilled posts. Vulnerable patients. Scared parents. This is not ok. Time to act @SimonHarrisTD @LeoVaradkar @Paschald #CARECANTWAIT @IrishNeonatal
6
83
4
264
Can we degrade the performance of neural link predictors by poisoning their training set? Tune in today at #ACL2021NLP 23:40 UTC. Peru Bhardwaj will show how she exploited KGE inference patterns to lower MRR of ~20 base points. aclanthology.org/2021.acl-lo… github.com/PeruBhardwaj/Infe…
At this years #ACL2021NLP conference @accenturetech funded @tcddublin PhD student Peru Bhardwaj will present her paper outlining security vulnerabilities of KGE that also help to understand model behaviour. Find more at: github.com/PeruBhardwaj/Infe… @lukostaz @johndkelleher
2
7
— Context — Speaking about the transformer architecture, one may incorrectly talk about an encoder-decoder architecture. But this is *clearly* not the case. The transformer architecture is an example of encoder-predictor-decoder architecture, or a conditional language-model.
I'm mad 😡😡😡 I just finished teaching and… my understanding was NOT matching my slides! Dang! 😠😠 I'm redrawing my deck. Again! Argh! 😤 Summary: new pretty slides on attention and transformers coming out for y'all very soon 😇😇😇
9
30
1
203
🎉 Papers with Code partners with arXiv! Code links are now shown on arXiv articles, and authors can submit code through arXiv. Read more: medium.com/paperswithcode/pa…
42
1,684
282
6,098
Fantastic @TheEconomist podcast with the great @kncukier about the amazing #virus world. It will be amazing when we turn phage virus onto awesome new therapies!
Viruses are everywhere and have evolved to attack every living thing. But they are also engines of evolution. Geoff Carr, our science editor, tells @kncukier how they have shaped the world on our “Babbage” podcast econ.trib.al/VO7oDAZ
1
11
We now have 44 high-quality, short code examples demonstrating common workflows at keras.io/examples/ Most of them are contributed by the community. They cover computer vision, NLP, timeseries, audio, generative deep learning, RL...
5
259
21
1,138
A very simple Kaggle notebook that shows how to use KerasTuner to automatically find a high-performing model: kaggle.com/fchollet/keras-ke… Includes finding the best image augmentation config as part of the parameter search (this is actually one of the most important things to tune)
3
122
3
542
I wrote a minimal/educational GPT training library in PyTorch, am calling it minGPT as it is only around ~300 lines of code: github.com/karpathy/minGPT +demos for addition and character-level language model. (quick weekend project, may contain sharp edges)
36
615
54
3,306
Interested in Anomaly Detection? Check out these two new tutorials on tensorflow.org and keras.io: tensorflow.org/tutorials/gen… keras.io/examples/timeseries… The TF tutorial is inspired by this truly excellent interactive demo by @vykthur anomagram.fastforwardlabs.co…
3
89
4
310
“You’ve always suspected AlexNet to be a bit of a cheat — but what about you, ResNet? Is it too much to ask for a bit more depth of character from someone with 152 layers?” 😂 towardsdatascience.com/are-a…
5
12
Wow, AmpliGraph just won the 2019 "Best Technical Advance in the Field of Data Science" award! Big shout-out to our terrific @accenturelabs team @sumitppai @PedroTabacof @chan_le_van @Rory_M_McGrath @nickpmcc #datsciawards and check out ampligraph.org /cc @AccentureDock
4
11
2
48
And the winner is @accenturelabs 💪 🏆 #DatSciAwards
5
2
10
4/ Tutorials! (Jupyter notebooks): - AmpliGraph basics: training, saving and restoring a model, evaluating a model, discover new links, visualize embeddings. - Link-based clustering and classification docs.ampligraph.org/en/1.1.0… (Also available on Colab colab.research.google.com/dr… )
1
1
3
Teams of high school students built bottle-flipping robots for RoboCon 2018 in Japan
28
796
144
2,463
Hey @BBCWorld .. each time you call #kashmir ‘Indian Occupied Kashmir’ I keep wondering why you refuse to call Northern Ireland ‘British Occupied Ireland’ ?
3,290
27,911
1,729
83,737
Replying to @aviationreg
@aviationreg @MoCA_GoI @jetairways flight from Dub-Paris on 16Apr on schedule but connecting flight from Paris-Mumbai rescheduled by 24hr to 17th April with ONLY 3 days notice. No alternative flights provided as per EU laws. (1/2)
1
@jetairways Forcing us to cancel tickets to avoid EU regulations on compensation. Not confirming if accommodation would be provided at Paris. If we cancel and rebook on other airlines prices are skyhigh for this month. #FeelingHelpless @aviationreg @MoCA_GoI (2/2)
Sumit retweeted
New top story on Hacker News: AmpliGraph: A TensorFlow-Based Library for Knowledge Graph Embeddings ift.tt/2CZY98E