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

109028, Moscow,
11, Pokrovsky boulevard

Phone: +7 (495) 531-00-00 *27254

Email: computerscience@hse.ru

 

Administration
First Deputy Dean Tamara Voznesenskaya
Deputy Dean for Research and International Relations Sergei Obiedkov
Deputy Dean for Methodical and Educational Work Ilya Samonenko
Deputy Dean for Development, Finance and Administration Irina Plisetskaya

Colloquium

For a researcher in a diverse and quickly developing area of knowledge such as computer science, it is important to maintain a broad perspective and strive to understand what colleagues in related fields are studying. This requires a platform where specialists can meet and tell each other about their latest findings in a common language. Such a platform is the Colloquium of HSE's Faculty of Computer Science. This platform is a faculty-wide academic seminar designed for teachers and research staff, graduate and undergraduate students, as well as those who are interested in computer science.

Colloquium meetings are held on Tuesdays in the Faculty of Computer Science building at Kochnovsky Proezd, 3, lecture hall 205, 2nd floor.

NB: a somewhat more detailed web page is available in Russian here.

Registration

2020 – 2021

October 26, 16:20

Speaker: 

Max Welling University of Amsterdam

Title: Is the next deep learning disruption in the physical sciences?

Abstract: A number of fields, most prominently speech, vision and NLP have been disrupted by deep learning technology. A natural question is: "which application areas will follow next?".  My prediction is that the physical sciences will experience an unprecedented acceleration by combining the tools of simulation on HPC clusters with the tools of deep learning to improve and accelerate this process. Together, they form a virtuous cycle where simulations create data that feeds into deep learning models which in turn improves the simulations. In a way, this is like building a self-learning computational microscope for the physical sciences. In this talk I will illustrate this using two recent pieces of work from my lab: molecular simulation and PDE solving. In molecular simulation we try to predict molecular properties or digitally synthesize molecules with prescribed properties. We have built a number of equivariant graph neural networks to achieve this. Partial differential equations (PDEs) are the most used mathematical model in natural sciences to describe physical processes. Intriguingly, we find that PDE solvers can be learned from data using graph neural networks as well, which has the added benefit that we can learn a solver that can generalize across PDEs and different boundary conditions. Moreover, it may open the door to ab initio learning of PDEs directly from data.

Register


September 28, 18:10

Speaker: 

Tatiana Likhomanenko (Apple)

Title: Positional Embedding in Transformer-based Models

Abstract: Transformers have been shown to be highly effective on problems involving sequential modeling, such as in machine translation (MT) and natural language processing (NLP). Following its success on these tasks, the Transformer architecture raised immediate interest in other domains: automatic speech recognition (ASR), music generation, object detection, and finally image recognition and video understanding. Two major components of the Transformer are the attention mechanism and the positional encoding. Without the latter, vanilla attention Transformers are invariant with respect to input tokens permutations (making "cat eats fish" and "fish eats cat" identical to the model). In this talk we will discuss different approaches on how to encode positional information, their pros and cons: absolute and relative, fixed and learnable, 1D and multidimensional, additive and multiplicative, continuous and augmented positional embeddings. We will also focus on how well different positional embeddings generalize to unseen positions for both interpolation and extrapolation tasks.
Register 

Colloquium (PDF, 452 Kb) 

 

June 1, 16:20 – 17:40

Speaker:

Konrad Schindler, ETH Zurich

Title: Computer Vision and Machine Learning for Environmental Monitoring

Abstract: I will give an overview of our work that uses machine learning to map and monitor environmental variables (like canopy height, ice cover etc.) from satellite and ground-based images, and discuss engineering challenges that arise in environmental applications.

Colloquium (PDF, 1.62 Mb) 


May 18, 16:20 – 17:40

Speaker

Mikiya Masuda, Osaka City University Advanced Mathematical Institute, HSE University

Topic: Bruhat interval polytopes which are cubes

Abstract

For a pair of permutations with in the Bruhat order, the Bruhat interval polytope is defined as the convex hull of points associated with permutations for . It lies in a permutohedron and is an example of a Coxeter matroid polytope.

The Bruhat interval polytope is the moment polytope of some subvariety of a flag variety called a Richardson variety and it is known that the Richardson variety is a smooth toric variety if and only if is combinatorially equivalent to a cube.

In this talk, I will explain that a certain family of Bruhat interval polytopes, which are particularly combinatorially equivalent to a cube, determines triangulations of a polygon. It turns out that the Wedderburn-Etherington numbers which count \emph{unordered} binary trees appear in their classification. If time permits, I will discuss another family of Bruhat interval polytopes and their classification, where directed paths, more generally directed Dynkin diagrams appear.

This talk is based on recent joint work with Eunjeong Lee (IBS-CGP) and Seonjeong Park (Jeonju Univ.).

Colloquium (PDF, 1.61 Mb) 


April 13, 18:10 – 19:30

Speaker

Alan Herbert, InsideOutBio (President and Founder), HSE University (Academic Supervisor of the International Laboratory of Bioinformatics)

Topic: Genetic computers

Abstract

 • Biologists and Engineers view Programming Problems very Differently
 • Encoding of Genetic Information is by both Nucleic Acid Structure (flipons) and Sequence (codons)
 • Encoding Genetic Programs by Structure is More Adaptive than by Sequence
 • I will use the Biology of Left-handed Z-DNA and Z-RNA to Exemplify these Principles
 • I will also Describe the Potential Use of Other Flipon Classes in Genetic Programming

Colloquium (PDF, 235 Kb) 

 

March 2, 16:20 – 17:40

Speaker
Alexander Panchenko, Skoltech

Topic: Neural entity linking

Abstarct

In this talk, I will provide a brief survey of recent neural entity linking (EL) systems developed since 2015 as a result of the "deep learning revolution" in NLP. I will try to systemize design features of neural entity linking systems and compare their performances to the best classic methods on the common benchmarks distilling generic architectural components of a neural EL system, like candidate generation and entity ranking summarizing the prominent methods for each of them, such as approaches to mention encoding based on the self-attention architecture.
Besides, various modifications of this general neural entity linking architecture can be grouped by several common themes: joint entity recognition and linking, models for global linking, domain-independent techniques including zero-shot and distant supervision methods, and cross-lingual approaches. Since many neural models take advantage of pre-trained entity embeddings to improve their generalization capabilities, I will also briefly discuss several types of entity embeddings. Finally, we briefly discuss classic applications of entity linking, focusing on the recently emerged use-case of enhancing deep pre-trained masked language models such as BERT. The materials are based on the following survey: Özge Sevgili, Artem Shelmanov, Mikhail Arkhipov, Alexander Panchenko, and Chris Biemann (2021): Neural Entity Linking: A Survey of Models based on Deep Learning. CoRR abs/2006.00575 

Colloquium (PDF, 172 Kb) 

Presentation (PDF, 2.59 Mb) 



February 18, 18:10 – 19:30

Pokrovsky blvd 11, room R406

Speaker
Laurent Beaudou (HSE)

Topic: Of Points and Lines: Graphs, Metrics, and Betweenness

Abstarct

Given n points in the Euclidean plane, they are either all collinear or define at least n distinct lines. This result is a corollary of the Sylvester-Gallai theorem. Its combinatorial generalization was proven by de Bruijn and Erdös in the forties. In 2008, Chen and Chvátal described a generalization of the notion of a line to any metric space and conjectured that the same result remains true in that framework. Since then, a growing community of researchers has been investigating this question. It remains open for metric spaces and even for those specific metric spaces generated by graphs. In this talk, we shall see a broad overview of the state of research on the matter: results and (many!) remaining open questions.

Colloquium



December 15, 16:20 – 17:40

Speaker 
Nikita Zhivotovskiy, Google research

Topic: On robust mean estimation and k-means clustering

Abstarct

In this talk we consider the robust algorithms for the k-means clustering problem where a quantizer is constructed based on N independent observations. We start with an overview of the methods of robust statistics. First, we discuss the median-of-means estimator. This simple estimator allows us to evaluate the mean of some heavy-tailed distribution as if this distribution was Gaussian. We discuss some extensions to the multivariate case. In the context of clustering, we present the median of means based non-asymptotic distortion bounds that hold under the two bounded moments assumption. In particular, our results extend the renowned asymptotic result of Pollard who showed that the existence of two moments is dsufficient for strong consistency of an empirically optimal quantizer in R. In a special case of clustering in Rd, under two bounded moments, we show matching non-asymptotic upper and lower bounds on the distortion, which depend on the probability mass of the lightest cluster of an optimal quantizer. This talk is based mainly on the joint work with Y. Klochkov and A. Kroshnin (https://arxiv.org/abs/2002.02339to appear in Annals of Statistics).

Colloquium (PDF, 441 Kb) 



November 12, 16:40 – 18:00
Pokrovsky blvd 11, room R205

Jie-Hong Roland Jiang (National Taiwan University)

Some Adventures in Boolean Satisfiability and Its Logic Synthesis Applications

Boolean satisfiability (SAT) is a fundamental NP-complete problem. Its general and simple formulation makes it an ideal problem to tackle. Although SAT is intractable, many efficient solvers have been engineered and widely applied in industries. This talk will introduce some of the key enabling techniques in SAT solving and showcase some applications in logic synthesis. 

Colloquium
 Registration


September 10, 16:40 – 18:00
Pokrovsky blvd 11, room R503

Ulrich Furbach, University of Koblenz and wizAI GmbH

From Theorem Proving to Cognitive Reasoning

Starting from a depiction of the state of the art in predicate logic theorem proving we address problems which occur if provers are applied in the wild. In particular we discuss how automated reasoning systems can be used for natural language question answering. Our approach to take common sense reasoning benchmarks within the Corg project is presented and we demonstrate how word embeddings can help with the problem of axiom selection.

Colloquium

Registration

 



2018 – 2019



March 26, 18:10 – 19:30 
Kochnovskii proezd, 3, room 205

Konstantin Yakovlev (Federal Research Center «Computer Science and Control» / HSE)

From Start to Goal: How Mobile Robots Plan Their Paths (Methods and Algorithms)

Games on graphs and on trees have been used in the fields of semantics and verification. Usually, they are defined as sequential games, where a play is a sequence of moves by the players.

The ability to move is the key feature of mobile intelligent agents such as mobile robots. No wonder that the problem of autonomous naviga-tion has been extensively studied in robotics and artificial intelligence. One of the prominent approaches is to decompose the navigation task into a series of subtasks such as localization, mapping, path planning, trajectory following, etc. In this talk, we will focus on one of such tasks, namely, on path planning (aka path finding), and will explore modern methods and algorithms tailored to solve it.

Colloquium

Registration



October 16, 18:10 – 19:30 
Kochnovskii proezd, 3, room 205

Luca Bernardinello (University of Milano-Bicocca)

Asynchronous games for Petri nets

Games on graphs and on trees have been used in the fields of semantics and verification. Usually, they are defined as sequential games, where a play is a sequence of moves by the players.

However, when synthesizing or analyzing distributed systems, in which events happen concurrently and the global state is not observable, this approach is not always appropriate, since concurrency is hidden in the interleaving of events. Therefore, several kinds of games in which the players can move asynchronously have been proposed in recent years. I will present an attempt to define such a game, originally conceived in order to tackle the problem of “observable liveness”, in which an agent tries to control a Petri net so that a given transition will fire over and over, assuming that only a subset of the transitions is directly controllable.

Colloquium

Registration



September 11, 18:10 – 19:30
Kochnovskii proezd, 3, room 205

Joseph MacInnes

Head of vision modelling lab / HSE

Computational cognitive neuroscience: A brief primer

Computational models in psychology and neuroscience share many algorithms with machine learning, machine vision and artificial intelligence, but the focus of the research is different. Where applied fields try to create algorithms that solve or automate a specific problem, computational modelling uses these algorithms to better understand fundamental workings of human brain and cognition. Rather than optimizing a new process, we try to simulate and understand an existing process. While computational modelling is still a growing field, there have emerged a number of contenders that perform very well in simulating various neural and cognitive processes. Diffusion models of decision making, salience models of vision and more recently deep learning models of object classification have all shown promise on their respective tasks. This talk will give an overview of a number of these models and discuss possible points of overlap with computer science and cognitive psychology.

Colloquium

Registration


Сolloquium in 2017/2018


Сolloquium in 2016/2017


Сolloquium in 2015/2016