Tweetovi
- Tweetovi, trenutna stranica.
- Tweetovi i odgovori
Blokirali ste korisnika/cu @artygadetsky
Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @artygadetsky
-
Artyom Gadetsky proslijedio/la je Tweet
Low-variance Gradient Estimates for the Plackett-Luce Distribution by
@artygadetsky,@k_struminsky and Dmitry Vetrov in collaboration with Christopher Robinson and Novi Quadrianto http://bayesiandeeplearning.org/2019/papers/115.pdf …pic.twitter.com/O5q1Rsz1Os
Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Artyom Gadetsky proslijedio/la je Tweet
Check out our new paper "Low-variance Black-box Gradient Estimates for the Plackett-Luce Distribution", accepted as an oral to
#AAAI2020 (@RealAAAI) on how to reduce the variance of gradients when optimizing w.r.t. a distribution over permutations. Paper: https://arxiv.org/abs/1911.10036 pic.twitter.com/p0hVK6ptcD
Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Artyom Gadetsky proslijedio/la je Tweet
Our team received the "Outstanding Educator Award" for the
@coursera "Advanced Machine Learning" specialization we created! Very proud to be a part of itpic.twitter.com/mN3IoVA5Hr
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Artyom Gadetsky proslijedio/la je Tweet
We are hiring a postdoc on deep reinforcement learning. Come work with us! Deadline 31 June 2019. Call for applications:https://bayesgroup.ru/news/call-for-postdoc-on-deep-rl/ …
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Artyom Gadetsky proslijedio/la je Tweet
The Deep Weight Prior by
@andrew_atanov,@senya_ashuha,@k_struminsky, Dmitriy Vetrov in collaboration with@wellingmax: an empirical prior for conv. layers in Bayesian Neural Nets that improves learning on small datasets#ICLR2019 https://openreview.net/forum?id=ByGuynAct7 … https://youtu.be/yA8k_da_co0Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Artyom Gadetsky proslijedio/la je Tweet
Variance Networks: When Expectation Does Not Meet Your Expectations by
@k_neklyudov,@dmolch111,@senya_ashuha, Dmitry Vetrov: stochastic neural networks with N(0,σ²) weights where only variances are learned work surprisingly well. https://openreview.net/forum?id=B1GAUs0cKQ …https://youtu.be/KwfED-brvj8Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Artyom Gadetsky proslijedio/la je Tweet
Variational Autoencoder with Arbitrary Conditioning by Oleg Ivanov,
@mfigurnov, Dmitry Vetrov: an extension of conditional VAE allowing to condition on an arbitrary subset of the features and sample the remaining ones#ICLR2019 https://openreview.net/forum?id=SyxtJh0qYm …https://youtu.be/pyJ6JVGb27kHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Artyom Gadetsky proslijedio/la je Tweet
We've got 3 papers accepted to
#ICLR2019https://bayesgroup.ru/news/iclr-2019/Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Artyom Gadetsky proslijedio/la je Tweet
Check out the short video summarizing our
#NeurIPS2018 paper "Loss Surfaces, Mode Connectivity, and Fast Ensembling of DNNs". https://www.youtube.com/watch?v=UwKdNzb_Q6U … Come to our presentation (spotlight at 04:20 PM, poster #162, Tue Dec 4th). We would be happy to talk to you at the conference!pic.twitter.com/moZ8AoeN65Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Artyom Gadetsky proslijedio/la je Tweet
We released a pre-print on the Deep Weight Prior - a flexible prior distribution for Bayesian CNNs via generative modeling of convolutional kernels. Thanks to
@bayesgroup and my awesome co-authors@andrew_atanov Kirill Struminsky Dmitry Vetrov@wellingmax http://arxiv.org/abs/1810.06943 pic.twitter.com/070AVKgdJ4
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Artyom Gadetsky proslijedio/la je Tweet
At
@acl2018? Come visit our poster on "Conditional Generators of Words Definitions" by@artygadetsky and@silkbottle. Today (July 17), poster session 2D, poster no. D8 https://acl2018.org/paper/1397 https://twitter.com/bayesgroup/status/1011945583454584834 …
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Artyom Gadetsky proslijedio/la je Tweet
We have significantly updated our preprint on Variance Networks! Training only variances of the weights w_ij ~ N(0, sigma_ij^2) leads to state-of-the-art performance. Now more explanation, practical and theoretical results. https://arxiv.org/abs/1803.03764 https://twitter.com/SashaVNovikov/status/973826922726805509 …
Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi -
Artyom Gadetsky proslijedio/la je Tweet
Check out our new paper "Conditional Generators of Words Definitions" (
@acl2018) by@artygadetsky,@silkbottle and Dmitry Vetrov on how to generate word definitions in face of ambiguity by virtue of a usage example. Paper: https://arxiv.org/abs/1806.10090 Code: https://github.com/agadetsky/pytorch-definitions …pic.twitter.com/uSw0dbWs7v
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.