AI & BigData Day is bi-annual conference for Data Science Community in Ukraine.
Conference are organizing from 2017.

About speaker:


Presentation topic:


More info about the topic:

About speaker:
Head of the Data Science architecture at joint-stock company "First Ukrainian International Bank" and member of the ambitious and successful Data Science team.

Presentation topic:
MLOps: your way from nonsense to valuable effect (approaches, cases, tools) (UA)

More info about the topic:
Let's talk about how to build a process that will help the Data Sсience team quickly, successfully and painlessly place models in a productive environment, how to implement processes that will make the team work easier and faster, look at the tools, approaches and cases that allow you to achieve this.

Presentation topic:
Challanges in real-world CV product

More info about the topic:
Зазвичай всі доповіді про нові типи нейронок, про деталі навчання їх і який крутий результат вийшов.

Проте мало хто розказує про те, звідки для реального проєкту взяти дані для навчання моделі, як багато їх потрібно в еру глибинного навчання. І про проблему синтетичних даних, як з нею боротись і т.д.

Ми з колегою зробимо доповідь по реальному проєкту, що недавно запустили в продакшин, поділимося своїм досвідом роботи в реальному житті.

Під час проєкту як раз ми зіштовхнулись з проблемою даних, і що на синтетичних даних нічого нормально не працювало і розкажемо, як важливі реальні дані і т.д.

Ось посилання про наш проєкт:
https://itc.ua/news/fozzy-group-pochala-klientske-testuvannya-shtuchnogo-intelektu-kissa-ai-yakij-zdaten-obslugovuvati-gostej-na-kasi-bez-dopomogi-lyudini/

https://www.youtube.com/watch?v=60FmYfO5_8c

https://rau.ua/ru/news/budushhee-segodnja-silpo/

About speaker:
Finalsit of Vodafone Hackatone, organizer of Data Science meetings in Lean Coffee format, passionate about data.

Presentation topic:
Working with BERT

More info about the topic:

• preBert history;
• Why BERT;
• Different flavours of BERT;
• Making BERT your own.

About speaker:
Senior Deep Learning Researcher and Machine Learning Architect with over 7 years of commercial experience in applying Deep Learning Models. Strong skills in Google's and Facebook's Deep Learning Frameworks - TensorFlow, Pytorch and Keras with Python. Industrial experience in processing large amounts of data of various types. Developing complex deep learning architectures.
Researches in AI, machine learning, deep learning, computer vision, natural language processing and generative models. Rich experience in the development and successful training of models at the intersection of CV and NLP using state-of-the-art methods of deep convolutional architectures and neural machine translators.

Presentation topic:
How Neural Networks Think and What They See

More info about the topic:
1. Modern neural networks logic: from MLP to ViT.
2. Convolution vs Attention vs Mixer.
3. Level of abstractions in neural networks.
4. Visualization techniques.
5. beyond neural visualization: Interpretation is All You Need.

Presentation topic:
Асоціативна пам'ять, мережі Хопфілда і не тільки
від механізму Attention до біологічних нейронів.

More info about the topic:
В цій доповіді ми спочатку зробимо огляд сучасних мереж Хопфілда. Такі мережі мають дуже велику ємність та можуть навчатися градієнтним методом.
Ми покажемо що ці мережі можуть поводитися еквівалентно до механізму Attention, який лежить в основі мереж-трансформерів. Далі ми познайомимося з ядерною асоціативною пам'яттю , що функціонує дуже подібно до сучасних хопіфлдівських мереж, та має в основі ядерне перетворення (kernel trick)

В заключній частині доповіді ми зробимо огляд свіжих результатів з відтворення поведінки складних моделей біологічних нейронів за допомогую глибокої нейронної мережі
а також покажемо як біологічні нейрони вирішують задачі для яких в машинному навчанні потрібна велика "інженерна" (глибока) нейромережа. Під час доповіді ми зробимо демонстрацію роботи згаданих моделей і алгоритмів.

About speaker:
Brian Gillikin is an Project Lead and Data Scientist at IBM, where he works with U.S. Government clients on AI/ML, cloud, and information management projects. He is also an open source contributor to Egeria, a Linux Foundation AI & Data project.

Presentation topic:
Scaling the AI Enterprise Ecosystem

More info about the topic:
1. As organizations scale, data, data architecture, and ability to utilize AI as a tool can struggle.
2. Enterprise-level decisions effect how and how well data science can be done.
3. To scale a healthy AI enterprise ecosystem:
a. Love your metadata.
b. Embrace domain-driven data as a product.
c. Decide how you and your AI ought to be (i.e., AI Ethics).

About speaker:
I am in the second year of my Ph.D. study at CTU, CIIRC. Currently, we work on an automated approach that aims to improve the scalability of imitation learning in robotics. We work on replacing expert demonstrations in a robot environment with information extracted from videos featuring humans performing the tool manipulation task.

Presentation topic:
Attention: Visual Transformers

More info about the topic:
RNN is dead long live the Transformers?
No! ConvNets are dead long live Visual Transformers!

• Why CNN are good for Computer Vision?
• The Dawn of Convolution from LeNet to EfficientNet.
• Giant Steps: AlexNet, VGG, Inception, ResNet, MobileNet, NAS.
• CNN for classification, detection, segmentation and other tasks (e.g. image captioning, etc).
• BiT: Big Transfer.
• Why CNN is bad for Computer Vision?
• What is receptive field?
• Attention is all you need!
• From BiT to ViT.
• An Image is Worth 16x16 Words.
• Future of ViT.
• From BYOL to CLIP.
• Swin Transformer.
• MLP, Convolution, Attention, Mixer.. What next?

About speaker:
Senior Machine Learning Engineer with over 8 years of research and commercial experience in applying Deep Learning models. Has extensive experience in various computer vision tasks such as feature engineering, object detection/segmentation, and face recognition.

About speaker:
Стартап-консультант та підприємець. Протягом часу роботи в Lviv Startup Club долучився до розвитку декількох сотень українських стартапів. Активно досліджує нові моделі успіху в GameDev, Digital Marketing, Project Management, Product Management, Data Science, IT Service Business Management.

About speaker:
I am co-founder and CEO of DevRain, a software development company. We are good at creating web, cloud, mobile and Artificial Intelligence solutions. In 2013 company was awarded as "Lead partner in creating mobile solutions on Windows Phone 8 platform" by Microsoft.

Our product, @Chatty, is the conversational AI platform for employees recruiting, interviewing, onboarding, and learning processes automation. Our mission is to help companies to hire the best talents, and to help job seekers constantly improve their skills to be able to find a job of their dream.

I am co-founder and CTO of award-winning social startup DonorUA - automated blood donors recruitment and management system. We won several competitions such as Stockholm Startup Weekend: Social Innovation, IDCEE Hackathon, "Ukraine is looking for startups" and others.

I am PhD. in Computer Science. My work was connected with Data Mining, information search, NLP, AI/ML.

Presentation topic:
Чому створення data strategy для компаній – це першочергове завдання?

More info about the topic:
Переважна більшість компаній хоче впроваджувати ті чи інші елементи штучного інтелекту. Попри це, багато ініціатив закінчуються достроково через брак даних, відсутність централізованого сховища даних тощо. Тому для компаній є критично важливим розробка стратегії даних (data strategy), що дозволить значно пришвидшити реалізацію проєктів у сфері штучного інтелекту?

Невеличкий спойлер: чи є випадки, коли вам не потрібна стратегія даних? Звісно. Якщо у вас немає даних :-)

About speaker:
Oleksandr Gurbych is a Research Scientist at SoftServe. Oleksandr holds a PhD in Theoretical Chemistry and has over eight years of academic and industry research experience in applied data science and machine learning in chemistry, biotechnology, and physics.

About speaker:
Microsoft cloud technologies expert. More than 15 years' experience in IT systems management and software development. During the last 13 years has been working at Microsoft on different roles in product marketing, technical evangelism and software engineering.

About speaker:
Andy has 25 years experience in software development. He built his first Artificial Neural Network in 1986, worked on various IT projects related to big data and data analytics, from 2016 switched on data science projects and set up a company MindCraft fully dedicated to creating data-driven solutions to bring new ideas from data insights and reduce staff costs.

Presentation topic:
3D modeling using Differentiable Programming

More info about the topic:
How enhance CNN with complex activation function written using Differentiable Programming that helps the model to recognize patterns in the 3D world. Practical lecture with real data examples and explanations deep in the linear algebra and vectorization.

Presentation topic:
How to make data useful for Bank.

More info about the topic:
Specific of data science products & integration with Bank data landscape

Presentation topic:
Kubeflow for end2end machine learning lifecycle (RU)

More info about the topic:
We are going to cover the machine learning model lifecycle, see what tools are used for each step of this journery. After that, we are going to explore some real-world examples and will deep dive into one of them with Kubeflow as a base.

About speaker:
Dmitry is a Microsoft veteran, working for more than 15 years. He started as a Technical Evangelist, and in this role presented on numerous conferences, including twice being on stage with Steve Ballmer. He then worked for 2 years as Senior Software Engineer, helping big European companies to start pilot digital transformation projects based on in AI and ML. As Cloud Developer Advocate, Dmitry focuses on creating educational content and working with academic and research institutions. He is also an Associate Professor at MIPT, HSE and MAI in Moscow, a big fan of functional programming and F#, and a maintainer/primary developer of mPyPl library. In his spare time, Dmitry explores Science Art and Technological Magic, as well as performs Chinese tea ceremonies. He can be reached at http://soshnikov.com.

Presentation topic:
Using Cloud and Text Analytics to Gain Insights from COVID-19 Papers Corpus

More info about the topic:
In this session, we show how to leverage CORD dataset, containing more than 400000 scientific papers on COVID and related topics, and recent advances in natural language processing and other AI techniques to generate new insights in support of the ongoing fight against this infectious disease.

The idea explored in our talk is to apply modern NLP methods, such and named entity recognition (NER) and relation extraction to article's abstracts (and, possibly, full text), to extract some meaningful insights from the text, and to enable semantically rich search over the paper corpus. We first investigate how to train NER model using Medical NER dataset from Kaggle, and specialized version of BERT (PubMedBERT) as a feature extractor, to allow automatic extraction of such entities as medical condition names, medicine names and pathogens. Entity extraction alone can provide us with some interesting findings, such as how approaches to COVID treatment evolved with time, in terms of mentioned medicines. We demonstrate how to use Azure Machine Learning for training the model.

To take this investigation one step further, we also investigate the usage of pre-trained medical models, available as Text Analytics for Health service on the Microsoft Azure cloud. In addition to many entity types, it can also extract relations (such as the dosage of medicine provisioned), entity negation, and entity mapping to some well-known medical ontologies. We investigate the best way to use Azure ML at scale to score large paper collection, and to store the results.