The seminar of the machine intelligence laboratory is more like a small scientific conference, where first the laboratory staff will talk about the topic as a whole, then about their developments, and at the end there will be a General discussion for those interested. The workshop will be useful for anyone who is engaged in machine learning and wants to expand their knowledge.
To attend the seminar, please register!
To attend the seminar, please register!
Schedule of seminars
Wait for announcements of new seminars!
Past seminars
All recordings of seminars are posted on the youtube channel, and presentations on google drive
- 9. Webinar - using the TopicNet library for thematic modeling29.04.2020 19:00 | Youtube
- Purpose and motives for creating the TopicNet library
- Demonstration of using the library to solve the problem of thematic modeling: clustering an unstructured collection of documents (texts)
- Visualization results of modelling
- Answers to your questions
- 8. Demand forecasting: tasks and applicationsSeminar from GoodsForeast and Forecsys together with the laboratory of machine intelligence on predicting demand in retail.
Seminar program:
- Business case studies
- The problem of demand forecasting and approaches to its solution
- International M5 Competition and the opportunity to join the GoodsForecast team to participate in it
- 7. Fundamental AI23.01.2020 | MegaFon office, conference hall, 4th floor (m. Novoslobodskaya)
- The implicit Generative model
- Quantization and compression of neural networks
- Overview of evolutionary search methods for neural network architectures (NAS)
- 6. CV: New image processing methods20.12.2019 17:30| Megafone office
• Video deconvolution using machine learning algorithms
• Perceptual deep depth super-resolution and friends
• Neural networks for improving the quality of photos from a smartphone camera on the example of the problem of removing reflections
• Video super resolution-challenges and solutions - 5. NLP: The challenges of topic modeling06.12.2019 19:00 |113Main Building MIPT (9 Institutskiy Pereulok, Dolgoprudny)
• Application of renormalization theory in machine learning based on the entropy approach
• Balancing themes in the theme model
• Theme Bank - 4. Creating a research group and organizing research.22.11.2019 20:00 | Higher School of Economics
• A flexible process and tools for the implementation of applied ML research
• Automation of research in machine learning
• Model of the Institute of system programming of the Russian Academy of Sciences
• MIL.ScienceClub. Approach to organizing research in the machine intelligence Laboratory - 3. Time Series in accelerometry07.11.2019 19:00 | MIPT Поточная аудитория Физтех.Цифры
• Flexible process and tools for implementing applied ML research
• Time series analysis for manual labor monitoring
• End-to-end approach for human activity detection with IMU time series
• Machine learning in the task of restoring the object trajectory using the IMU system - 2. CV: Optical Character Recognition26.10.2019 | School of Data Science (Yandex)
• Search for glare on a document image to determine whether it is suitable for OCR "Fast Glare Detection in Document Images"
• Methods for detecting and binarizing text on a printed document image as part of the overall pipeline solution to the OCR problem
• DDI-100: Dataset from the machine intelligence Laboratory for training stable character recognition models in images of printed documents - 1. NLP: Topic Modelling11.10.2019 | MIPT
• Topic modeling
• Presentation of the open-source TopicNet library for automated construction of thematic models on collections of any size and subsequent visualization of modeling results
• Review of the best presentations of the RANLP conference and fresh ideas for solving NLP applications
07.11.2019 | MIPT
Time Series в акселерометрии
19:00 | поточная аудитория Физтех.Цифры
19:00 | поточная аудитория Физтех.Цифры
Tamaz Gadaev
Researcher of the group Time Series Analysis Machine Intelligence Laboratory
"Machine learning in the task of restoring an object's trajectory using an IMU system".
Anastasia Motrenko
Lead researcher at Forecsys
"Time series analysis for manual labor monitoring".
Vladimir Tsepulin
engineer Samsung R&D Institute
"End-to-end approach for human activity detection with IMU time series".
Researcher of the group Time Series Analysis Machine Intelligence Laboratory
"Machine learning in the task of restoring an object's trajectory using an IMU system".
Anastasia Motrenko
Lead researcher at Forecsys
"Time series analysis for manual labor monitoring".
Vladimir Tsepulin
engineer Samsung R&D Institute
"End-to-end approach for human activity detection with IMU time series".
22.11.2019 20:00 | HSU R307
Creating a student research group and organizing research
• Andrey Ustyuzhanin (head of the laboratory of big data analysis methods at HSE) - flexible process and tools for implementing applied ML research
• Vadim Strizhov (Professor of the Department of intelligent systems at MIPT) - automation of scientific research in machine learning.
• Denis Turdakov (head of information systems Department ISP RAS) and Jaroslav Naumov (researcher, ISP RAS) model of the Institute of system programming of RAS.
• Alexey Goncharov (head of the MIL MIPT research team) and Filipp Nikitin (research engineer of the Computer Vision group at MIL MIPT) - MIL.ScienceClub. An approach to organizing research in the machine intelligence Laboratory.
• Vadim Strizhov (Professor of the Department of intelligent systems at MIPT) - automation of scientific research in machine learning.
• Denis Turdakov (head of information systems Department ISP RAS) and Jaroslav Naumov (researcher, ISP RAS) model of the Institute of system programming of RAS.
• Alexey Goncharov (head of the MIL MIPT research team) and Filipp Nikitin (research engineer of the Computer Vision group at MIL MIPT) - MIL.ScienceClub. An approach to organizing research in the machine intelligence Laboratory.
06.12.2019 19:00 | 113 Main Building MIPT
The challenges of topic modeling
• Sergey Koltsov (candidate of physical and mathematical Sciences, Deputy head. laboratory for Internet studies, HSE) - application of renormalization theory in machine learning based on the entropy approach
• Evgenia Veselova (NLP group researcher at the machine intelligence laboratory) - Balancing topics in a thematic model
• Vasily Alekseev (NLP group researcher at the machine intelligence laboratory) -
Theme Bank
• Evgenia Veselova (NLP group researcher at the machine intelligence laboratory) - Balancing topics in a thematic model
• Vasily Alekseev (NLP group researcher at the machine intelligence laboratory) -
Theme Bank
20.12.2019 17:30 | MegaFon Office. audience 3.04 (address: 41 Armory lane. entrance from the garden ring. Novoslobodskaya Metro Station)
Current Computer Vision problem statements
• Evgeny Burnaev (Skoltech, Professor, head of. http://adase.group) - "deep Perceptual depth super-resolution and friends";
• Yaroslav Koshelev (researcher in the image and video processing group, Huawei research center) - " video Deconvolution using machine learning algorithms";
• Pavel Ostyakov (lead engineer at Huawei research center) - "Neural networks for improving the quality of photos from a smartphone camera on the example of the problem of removing reflections";
• Ilya Zharikov (research Engineer of the Computer Vision & Time Series Analysis group at the MIPT machine intelligence Laboratory) and Philip Nikitin (research Engineer of the group, MIPT machine intelligence Laboratory) - Computer Vision & Time Series Analysis - "Video super resolution - challenges and solutions".
• Yaroslav Koshelev (researcher in the image and video processing group, Huawei research center) - " video Deconvolution using machine learning algorithms";
• Pavel Ostyakov (lead engineer at Huawei research center) - "Neural networks for improving the quality of photos from a smartphone camera on the example of the problem of removing reflections";
• Ilya Zharikov (research Engineer of the Computer Vision & Time Series Analysis group at the MIPT machine intelligence Laboratory) and Philip Nikitin (research Engineer of the group, MIPT machine intelligence Laboratory) - Computer Vision & Time Series Analysis - "Video super resolution - challenges and solutions".
23.01.2020 19:00 | MegaFon office, auditorium 3.36 (address: 41 Armory lane. entrance from the garden ring. Novoslobodskaya Metro Station)
7. Fundamental AI
Topics and speakers:
• Dmitry Vetrov (head of the Samsung laboratory at the HSE) - implicit generative models
• Mikhail Burtsev (head of the laboratory of neural systems and deep learning at MIPT and head of the DeepPavlov project) - review of methods for evolutionary search of neural network architectures (NAS)
• Tamaz Gadaev (junior researcher in MIL) - quantization of neural networks
• Philip Nikitin (research engineer in MIL) - neural network compression
• Dmitry Vetrov (head of the Samsung laboratory at the HSE) - implicit generative models
• Mikhail Burtsev (head of the laboratory of neural systems and deep learning at MIPT and head of the DeepPavlov project) - review of methods for evolutionary search of neural network architectures (NAS)
• Tamaz Gadaev (junior researcher in MIL) - quantization of neural networks
• Philip Nikitin (research engineer in MIL) - neural network compression
13.03.2020 19:00 | MIPT 107 Phystech.Bio
Demand forecasting: tasks and applications
The organizer of the seminar: GoodsForecast company, which for the last 15 years has been developing and implementing demand forecasting systems for business and everything related to it – logistics management, production planning, etc.among the company's clients are Okey and Verny retail chains, Gazpromneft gas station network, Knauf production companies, Baltika, Philip Morris. GoodsForecast claims that the successful solution of such problems as demand forecasting is always a combination of mathematical culture with business expertise, and we are ready to tell you how this happens in real projects.
Seminar program:
Cases for solving business problems. Says Andrey Lisitsa, CEO of GoodsForecast
The problem of demand forecasting and approaches to its solution. Says Daniel Kanevsky, Director of Analytics GoodsForecast.
International M5 Competition and the opportunity to join the GoodsForecast team to participate in it. Eugene Golubyatnikov, head of research at GoodsForecast, tells the story.
Registration is required to visit the site.
Seminar program:
Cases for solving business problems. Says Andrey Lisitsa, CEO of GoodsForecast
The problem of demand forecasting and approaches to its solution. Says Daniel Kanevsky, Director of Analytics GoodsForecast.
International M5 Competition and the opportunity to join the GoodsForecast team to participate in it. Eugene Golubyatnikov, head of research at GoodsForecast, tells the story.
Registration is required to visit the site.
13.03.2020 19:00 | MIPT 107 Phystech.Bio
Demand forecasting: tasks and applications
The organizer of the seminar: GoodsForecast company, which for the last 15 years has been developing and implementing demand forecasting systems for business and everything related to it – logistics management, production planning, etc.among the company's clients are Okey and Verny retail chains, Gazpromneft gas station network, Knauf production companies, Baltika, Philip Morris. GoodsForecast claims that the successful solution of such problems as demand forecasting is always a combination of mathematical culture with business expertise, and we are ready to tell you how this happens in real projects.
Seminar program:
Cases for solving business problems. Says Andrey Lisitsa, CEO of GoodsForecast
The problem of demand forecasting and approaches to its solution. Says Daniel Kanevsky, Director of Analytics GoodsForecast.
International M5 Competition and the opportunity to join the GoodsForecast team to participate in it. Eugene Golubyatnikov, head of research at GoodsForecast, tells the story.
Registration is required to visit the site.
Seminar program:
Cases for solving business problems. Says Andrey Lisitsa, CEO of GoodsForecast
The problem of demand forecasting and approaches to its solution. Says Daniel Kanevsky, Director of Analytics GoodsForecast.
International M5 Competition and the opportunity to join the GoodsForecast team to participate in it. Eugene Golubyatnikov, head of research at GoodsForecast, tells the story.
Registration is required to visit the site.
26.10.2019 | School of Data Science (Yandex)
CV: OCR
• Alexey Zhuravlev (ABBYY) - search for glare on the image of documents to determine suitability for OCR "Fast Glare Detection in Document Images". Search for glare on an image of documents to determine suitability for OCR "Fast Glare Detection in Document Images"
• Vladimir Dokholyan - methods for detecting and binarizing text on a printed document image as part of a General pipeline solution to the OCR problem.
• Ilya Vasiliev - DDI-100: Dataset from the machine intelligence Laboratory for training stable character recognition models on images of printed documents.
• Vladimir Dokholyan - methods for detecting and binarizing text on a printed document image as part of a General pipeline solution to the OCR problem.
• Ilya Vasiliev - DDI-100: Dataset from the machine intelligence Laboratory for training stable character recognition models on images of printed documents.
11.10.2019
NLP: Topic Modelling
• Konstantin Vorontsov - about topic modeling
• Eugene Yegorov and Evgenia Veselova - Presentation of the open-source TopicNet library for automated construction of thematic models on collections of any size and subsequent visualization of modeling results
• Viktor Bulatov - Review of the best speeches of the RANLP conference and fresh ideas for solving applied NLP problems
• Eugene Yegorov and Evgenia Veselova - Presentation of the open-source TopicNet library for automated construction of thematic models on collections of any size and subsequent visualization of modeling results
• Viktor Bulatov - Review of the best speeches of the RANLP conference and fresh ideas for solving applied NLP problems
Registrations
Нажимая на кнопку, вы даете согласие на обработку персональных данных и соглашаетесь c политикой конфиденциальности
Остались вопросы?
Пишите нам на почту:
Пишите нам на почту:
info@machine-intelligence.ru
© 2017 MIL. MIPT "Machine Intelligence Laboratory"
© 2019 MIL. LLC "GVA"
info@mil-team.com
© 2019 MIL. LLC "GVA"
info@mil-team.com
