ಟ್ವೀಟ್ಗಳು
- ಟ್ವೀಟ್ಗಳು
- ಟ್ವೀಟ್ಗಳು & ಪ್ರತಿಕ್ರಿಯೆಗಳು
- ಮಾಧ್ಯಮ
@joeri_hermans ತಡೆಹಿಡಿಯಲಾಗಿದೆ
ನೀವು ಖಚಿತವಾಗಿಯೂ ಈ ಟ್ವೀಟ್ಗಳನ್ನು ನೋಡಲು ಬಯಸುವಿರಾ? ಟ್ವೀಟ್ಗಳನ್ನು ನೋಡುವುದು @joeri_hermans ಅವರನ್ನು ತಡೆತೆರವುಗೊಳಿಸುವುದಿಲ್ಲ.
-
When using DOWNPOUR in a data parallel setting, divide the gradient residual by the staleness parameter. https://github.com/cerndb/dist-keras/blob/master/examples/mnist.ipynb … (ADAG)
#ml -
-
New version of "Direct Feedback Alignment" paper is out. Thanks to Colin Raffel for feedback on theorem and proof http://arxiv.org/abs/1609.01596
-
UN rights expert urges States not to sign the ‘flawed’ CETA treaty and put it to referendumhttps://dezayasalfred.wordpress.com/2016/10/28/un-rights-expert-urges-states-not-to-sign-the-flawed-ceta-treaty-and-put-it-to-referendum …
-
Will be presenting my progress on Distributed Deep Learning https://github.com/JoeriHermans/dist-keras/blob/master/examples/workflow.ipynb … at the CERN IML Working Group. https://indico.cern.ch/event/571102/
-
Noam Chomsky on "free trade agreements". https://www.youtube.com/watch?v=4abS3S0PzpQ …
#ceta -
Linux kernel push confirms, Intel to add dedicated neural network hardware ops into their future processors https://lkml.org/lkml/2016/10/12/530 …pic.twitter.com/FeavULow1q
-
Finally public! Molecular autoencoder lets us interpolate and do gradient-based optimization of compounds: https://arxiv.org/pdf/1610.02415.pdf …pic.twitter.com/VW7kiL2qM3
-
Nice post on generative adversarial networks from a dynamical systems perspectivehttp://www.araya.org/archives/1183
-
Deep Reinforcement Learning (DDPG) for 3D Car Racing Simulation TORCS with
#Keras#NeedForSpeed https://yanpanlau.github.io/2016/10/11/Torcs-Keras.html …pic.twitter.com/RxPnrRfAHA -
Releasing code to enable reproducibility (even partially) is important. I feel researchers who share code are more confident in their work.https://twitter.com/AndrewYNg/status/784182224954335233 …
-
Subpixel CNN's huh? I guess that infinite zooming isn't just a cool scene in CSI anymore. -https://github.com/Tetrachrome/subpixel …
-
My paper on [1609.09106] HyperNetworks is out. I wrote a new blog post about it, and released the HyperLSTM code. http://blog.otoro.net/2016/09/28/hyper-networks/ …pic.twitter.com/mfspJt2RNG – Googleplex - 1965 ರಲ್ಲಿ
-
-
Preliminary results using https://github.com/JoeriHermans/dist-keras … Keras with Apache Spark: Single: 6 minutes F1 = 0.723, EASGD: 2.5 minutes F1 = 0.738
-
Keras 1.1.0 is out. Note that TensorFlow is now the default backend for new installs.
-
Clinton's CTR campaign starts "TrumpLeaks" offering cash for info on Trump... despite still trying to prosecute us. http://correctrecord.org/trumpleaks
-
Averaging gradients on different workers with different data actually kinda works, pretty cool. Let's see if we can do better.
-
Joeri Hermans ಹಿಂಬಾಲಿಸಿದ್ದಾರೆ Albin Stjerna, Ian Goodfellow, fastml extra and 15 others
-
@goodfellow_ian
Research Scientist at OpenAI. Lead author of http://www.deeplearningbook.org Co-author of ML security blog http://www.cleverhans.io
ಲೋಡಿಂಗ್ ಸಮಯ ಸ್ವಲ್ಪ ತೆಗೆದುಕೊಳ್ಳುತ್ತಿರುವಂತೆನಿಸುತ್ತದೆ.
Twitter ಸಾಮರ್ಥ್ಯ ಮೀರಿರಬಹುದು ಅಥವಾ ಕ್ಷಣಿಕವಾದ ತೊಂದರೆಯನ್ನು ಅನುಭವಿಸುತ್ತಿರಬಹುದು. ಮತ್ತೆ ಪ್ರಯತ್ನಿಸಿ ಅಥವಾ ಹೆಚ್ಚಿನ ಮಾಹಿತಿಗೆ Twitter ಸ್ಥಿತಿಗೆ ಭೇಟಿ ನೀಡಿ.
Joeri Hermans
NEOMECHANICA
Arild Nøkland
Alfred de Zayas
Soumith Chintala
David Duvenaud
Mikael Huss
Ben Lau
hardmaru
SpaceX
François Chollet
WikiLeaks