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

Anniversary workshop "Data recovery clustering"

Event ended

On December 12, 2022 an anniversary workshop will be held on School of Data Analysis and Artificial Intelligence, dedicated to the 80th anniversary of the Tenured Professor, Professor at School of Data Analysis and Artificial Intelligence, Doctor of Technical Science Boris G. Mirkin.

The workshop is dedicated to the anniversary of Tenured Professor Boris G. Mirkin will take place on Monday, December 12th 2022 at 15.00.

Topic: "Data recovery clustering".

Abstract:  The talk reviews the author's work in clustering. It starts from approximation problems for square similarity matrices. Resulting cluster structures may represent conventional structures such as partitions and single clusters, as well as rather unusual structures such as structured partitions (aka block-models) and chain orders. Most recent work with author's participation include (a) least squares community detection in attribute-rich networks (2020-21)  and (b) core-shell clusters in the analysis of oceanic upwellings (2022-2023). Then I move on to the analysis of k-means clustering in this perspective. I present a Pythagorean decomposition of the data scatter, a comlementary criterion, and applications to (c) anomalous clustering and (d) geometric interpretation of chi-squared. Some  examples of application are given. Joint work with I. Muchnik, T. Smith, E. Koonin, Y. Wolf, and T. Fenner.

Room: G 303

PROGRAM

time speaker topic 
15.00-15.15 Sergei Kuznetzov  Opening speech
15.15-16.00 Boris Mirkin Data recovery clustering
16.00-16.30 Fuad Aleskerov Модели интервального выбора
16.30-17.00 Dmitriy Ilvovskiy  Attention in NLP: grabbing for good and for bad
     

COFFEE-BREAK (17.30-18.00)

18.00-18.30 Boris Goldengorin Mirkin's walk from prebase (in automata theory) to abstractions, generalizations, and interpretations

18.30-19.00

Vasilii Gromov

Spot the bot 

19.00-19.30

Sergei Kuznetzov

Delta-Closure Structure for Studying Data Distribution

Registration