• A
  • A
  • A
  • ABC
  • ABC
  • ABC
  • А
  • А
  • А
  • А
  • А
Regular version of the site

International Laboratory of Deep Learning and Bayesian Methods

Publications
Article
Composing Tree Graphical Models with Persistent Homology Features for Clustering Mixed-Type Data In print

Ni X., Quadrianto N., Wang Y. et al.

Proceedings of Machine Learning Research. 2017. Vol. Volume 70: International Conference on Machine Learning. P. 2622-2631.

Book chapter
Structured Bayesian Pruning via Log-Normal Multiplicative Noise

Neklyudov K., Molchanov D., Ashukha A. et al.

In bk.: Advances in Neural Information Processing Systems 30 (NIPS 2017). Montreal: 2017.

Working paper
Scalable Gaussian Processes with Billions of Inducing Inputs via Tensor Train Decomposition

Izmailov P., Novikov A., Kroptov D.

math. arxive. Cornell University, 2017

About the Laboratory

International Laboratory of Deep Learning and Bayesian Methods is established on the basis of Bayesian Methods Research Group. The group is one of the strongest scientific groups in Russia in the area of machine learning and probabilistic modeling. The laboratory researches the neurobayesian models that combine the advantages of the two most successful machine learning approaches, namely neural networks and Bayesian methods.


Second Summer School on Bayesian Methods in Deep Learning announced

During his talk on Sberbank Data Science Journey Dmitry Vetrov announced about second summer school on NeuroBayesian methods to be held in HSE in Aug 2018. 

Dmitry Vetrov Took Part in the Samsung AI Forum

At October 19-20, head of the Deep Learning and Bayesian Methods laboratory Dmitry Vetrov took part in the forum on artificial intelligence at the headquarters of the corporation Samsung.

HSE and Yandex launched a new English specialization on Coursera: Advanced Machine Learning

In this specialization the listeners will complete the courses on deep learning, Bayesian methods, reinforcement learning, natural language processing etc. Alexander Novikov, research fellow of the laboratory, is a lecturer of Bayesian Methods course.

Mini-workshop Stochastic Processes and Probabilistic Models in Machine Learning

On September 12 and 13, a mini-workshop "Stochastic processes and probabilistic models in machine learning" was held at the faculty. Four invited foreign specialists gave lectures about the application of parametric and nonparametric probabilistic methods in machine learning, and representatives of Russian scientific groups told about particular projects where these approaches are used.

Bayesian Methods in Deep Learning Summer School in Moscow

Bayesian Methods in Deep Learning Summer School was held in Moscow fron 26 to 30 August. During these five days 96 participants from 8 countries listened to lectures about Bayesian methods in deep learning and trained neural networks.

A paper on CVPR 2017

Michael Fugurnov's paper written in collaboration with researchers from Google, Carnegie Mellon University and Dmitry Vetrov has been presented on Computer Vision and Pattern Recognition. The conference was held from 21 to 26 July in Honolulu, USA.

A new grant by Russian Science Foundation

A group of 8 researchers including 5 laboratory's staff members received a large grant by Russian Science Foundation. The grant has been received in collaboration with the Laboratory of Computer Graphics and Multimedia (MSU) and Bayesian Methods research group.

Variational dropout sparsifies DNNs paper has been accepted to ICML'17

The paper authored by laboratory's research assistants Dmitry Molchanov and Arsenii Ashukha and head Dmitry Vetrov has been accepted to the International Conference on Machine Learning'2017. In this research a state-of-the-art result in deep neural networks sparsification was achieved using Bayesian framework applied to deep learning.

Collaboration with Samsung Opens New Perspectives for the Laboratory and the Faculty

Dmitry Vetrov, head of the laboratory, held a meeting with Mr. Shi-Hwa Lee, a Vice-President of Samsung, a company the laboratory collaborates with. Interim research results, internship possibilities and collaboration perspectives were discussed.

Around 300 applications are submitted to Bayesian Methods in Deep Learning Summer School

Application to Bayesian Methods in Deep Learning Summer School is now closed. There are 297 applications from citizens of Russia, Ukraine, Belarus, Great Britain, Spain, France, Switzerland, Germany, the USA and Ireland.The school will be held in Moscow in August, 2017.