Data Fusion 2021

The first conference dedicated to the synergy of Data Fusion data was held on March 30 - 31, 2021 at the Skolkovo Innovation Center
96
SPEAKERS
50+
PRACTICAL CASES
25
SESSIONS

Performance DATA fusion 2021

Science notes «Compositional AI: Fusion of AI/ML Services»
Science notes «Bipartite graph of science and industry»
Science notes «Artificial Intelligence vs fakes and post-truth politics»
Science notes «Prospects to ensemble neural connections»
Science notes «ML-renaissance of time-series forecasting»
Session «Data Fusion and work with customers» featuring Gartner, Tele2, Multicarta, VTB Bank, Russian Post
Session «ML-Renaissance of timeseries forecasting»
Discussion «NLP» featuring Skoltech, DeepPavlov, VTB Bank, Huawei, Skoltech, Rostelecom
Session «The future of data science and machine learning: critical trends you can’t ignore»
Session «Time series in economy»
Session «Graph analytics»
Discussion «Artificial Intelligence, data and government»
Session «Data classification and recommendation systems»
Session «Machines see everything. Computer vision and Data Fusion»
Discussion «Legal issues in Data Fusion concept»
Session «A system approach to teaching Data Science»
Session «Unified Data Fusion»
Discussion «Artificial Intelligence factory»
Analysis of Data Fusion finalists’ solutions
Opening speech. Vadim Kulik, Deputy president - chairman of the board, VTB Bank
Data Fusion 2021 Videos

CONTEST Data Fusion 2021

Data Fusion Contest is a data science contest with an unprecedented set of data from receipts in terms of volume and structure: a unique dataset consisting of 7,000,000 simulated receipts has been formed for the contest. The participants worked on solving problems that help to digitize the economy of consumption.
1000 PARTICIPANTS
18 COUNTRIES
50 CITIES
8 PRIZE-WINNING TEAMS
2 000 000 ₽ PRIZE FUND
NOMINATIONS AND WINNERS
Task No. 1
Award
Winner
GOODSIFICATION
1st place
500 000 ₽
flow
2nd place
250 000 ₽
hype
3rd place
150 000 ₽
keks
For the best public solution
50 000 ₽
exotol
For the speed of solving
50 000 ₽
Niw
Task No. 2
Award
Winner
BRANDEFINE
1st place
500 000 ₽
GIA
2nd place
250 000 ₽
Niw
3rd place
150 000 ₽
Artem Voronov
For the best public solution
50 000 ₽
InfusionTeam
For the speed of solving
50 000 ₽
Artem Voronov

TASKS

Task No. 1
GOODSIFICATION
Classification of data from receipts
From the English: “goods” and “classification”
Classification of data from simulated receipts, creation of an algorithm for categorizing goods on truly big data. Creation of a unique classifier for a wide range of the bank's client services, which has become one of the key parts in solving the tasks of enriching the client's portrait, personifying communication with them, and improving customer journey.
Metric
Weighted F1 score
Format of solution
Code in a docker container
Task No. 2
BRANDEFINE
Product brand definition
From the English: “brand” and “define”
Determining the brand of the product from the product name using current NLP methods. Teaching the brand allocation model to improve the lives of millions of people through the continuous development of the bank's customer service and partner ecosystem. Saving customers' time and providing them only with useful information
Metric
Accuracy
Format of solution
Code in a docker container