Hands-on Workshops

Bring your notebook and take part in one of 12 AI & NLP related workshops to quickly gain hands-on knowledge on current methods and technologies.

State-of-the-art Knowledge

Talk to field experts, scientists and PolEval competition winners to discuss state-of-the-art knowledge in the area of AI and NLP.

Science meets Business

Exchange ideas and talk business with participants both from academia and industry.

12 speakers

4 parallel tracks

150 participants

collocated with PolEval 2019 competition


The schedule of AI & NLP Workshop Day is tentative and may change in the coming weeks.

08:00 - 08:45

Registration and coffee

08:45 - 09:00

Opening and Welcome Speech

09:00 - 09:45

Train, explain, acclaim - Build a good model in three steps

Przemysław Biecek

Associate Professor at Warsaw University of Technology / Principal Data Scientist at Samsung SRPOL

Would you choose a model with high accuracy or a model that is easier to understood and explain? In this talk I will argue that in the long term the latter is more important. Embrace the power of XAI - eXplainable Artificial Intelligence.

09:45 - 10:30

Cyberbullying and Online Violence: from Detection to Reduction

Gniewosz Leliwa

Director of AI research and co-founder of Samurai Labs

Detection of cyberbullying and online violence itself is a very complex and challenging task. Yet it is just the first step. This talk is about the methods we use at Samurai Labs to move from detection to reduction. The author will provide a big-picture view on the problem and try to answer how the ultimate goal of reduction affects the real-life requirements of detection and how Samurai Labs deals with these requirements.

ML/AI Workshop

10:45 - 12:15

Working with images in Azure using Vision API

Kamil Mrzygłód

Senior Software Developer at Demant Technology Centre / Microsoft Azure MVP / Trainer at Sages

Azure Cognitive Services are a set of advanced APIs, which allow you to quickly build intelligent applications hosted in the cloud. During the workshop, you will learn the basics of this particular service using one of its components - Vision API. We will try to perform simple exercises like face detection, emotion detection and handwriting recognition and compare the result with custom models and services.

ML/AI Workshop

12:30 - 14:00

TensorFlow Object Detection API

Karol Majek

Organizer of Warsaw Self-Driving Cars regular meetups

During the workshop you will learn how to prepare a dataset for object detection task, which tools to use to annotate images and videos. I will show you the training and evaluation process. Finally you will run object detection on your laptop using video from webcam or file. The first part will require laptop with Internet browser and the second part will require Python, TensorFlow and OpenCV installed.

Lunch Break

ML/AI Workshop

14:45 - 16:15

Understanding LSTM and GRU networks in PyTorch

Piotr Migdał

Deep Learning Consultant at p.migdal.pl

Long short-term memory (LSTM) and gated recurrent unit (GRU) network are popular network architectures for text processing. During this workshop (held in PyTorch) we will work with them in a low-level way, getting access to memory cells and intermediate states. Targeted at people using LSTMs/GRUs as black boxes OR have a background in other network architectures and would like to understand natural language processing with deep learning.

ML/AI Workshop

16:30 - 18:00

Why does machine learning need adversaries?

Anna Wróblewska

Senior Data Scientist at Applica / Assistant Professor at Warsaw University of Technology

Kamil Grabowski

MSc student in Data Processing and Analysis at WUT / Data Scientist at Applica.ai

Adversaries work as a villain in disguise - a model predictions are incorrect for them. This occurs even in those cases in which the input data are very similar to the data that model can interpret correctly. Sometimes adversaries prepared in a very bright deliberate way make the model totally foolish. We will tell about how to make more robust models, vaccinate them against adversaries or other perturbations.

Big Data Workshop

10:45 - 12:15

Text processing in Apache Spark

Patryk Pilarski

Data Scientist / Trainer at Sages / Ph.D. student at PJATK & SWPS

Text processing at scale - from several books to thousands of volumes. In this workshop we will build a scalable system for book authorship attribution using Apache Spark. In the process we will see how to extract meaningful features from text and how to use them to infer who the author is. All of this using Python, Spark SQL and Spark ML.

Big Data Workshop

12:30 - 14:00

Getting started with BigQuery ML for data analysts

Radosław Stankiewicz

Strategic Cloud Engineer at Google

BigQuery Machine Learning is a new feature in BigQuery where data analysts can create, train, evaluate, and predict with machine learning models with minimal coding. In this lab, you will explore datasets and create machine learning models inside BigQuery. Prerequisites - GCP account with billing set up (credit card need to be provided to GCP but we won’t go beyond free tier).

Lunch Break

Big Data Workshop

14:45 - 16:15

Predicting typed words using n-grams - introduction to NLP in Azure Machine Learning Studio

Katarzyna Pakulska

Łukasz Wójcik

In this workshop we will build an Azure Machine Learning Studio Experiment in which we will train a model using the n-grams extracted from various data sources. Then we will use this model in our text editor to predict what user would want to type. In the first part of the workshop we will present typical blocks and cloud cognitive services related with NLP in Microsoft Azure. In the second part we will present how to build a tip suggester for search from scratch

Big Data Workshop

16:30 - 18:00

Kubeflow and Kubeflow Pipelines

Michał Żyliński

Customer Engineer at Google

In this session, you will learn how to install and use Kubeflow Pipelines to create a full machine learning application on Kubernetes.Starting with an empty environment, you will create a Kubernetes cluster and install Kubeflow from scratch. Then you will build and run a full pipeline that first trains a model using TensorFlow, then serves the model, and finally deploys a web front-end for interacting with the resulting predictions. You will then move into a notebook to build and run your pipeline using the Python SDK.

NLP Workshop

10:45 - 12:15

Creating R Shiny application for drug names comparison

Paweł Ekk-Cierniakowski

Advanced Analytics Expert at Polpharma / Trainer at Sages / PhD Student at Warsaw School of Economics

The workshop will be divided into theoretical and practical part. The first part will be the theory of measuring string distances. The second part will be devoted to the implementation of this theory by using R libraries to build application in R Shiny comparing drug names.

NLP Workshop

12:30 - 14:00

Dependency Parsing

Alina Wróblewska

Assistant professor at ICS PAS

Dependency parsing is a crucial issue in various NLP tasks. The predicate-argument structure transparently encoded in dependency-based syntactic representations may support machine translation, question answering, sentiment analysis, etc. In the workshop, I will present the fundamental ideas of the dependency theory, selected dependency-based resources, and some state-of-the-art dependency parsing systems.

Lunch Break

NLP Workshop

14:45 - 16:15

Neural Network Models of Text Normalization for Speech Applications

Paweł Cyrta

Speech Research Scientist at Rev.ai / AI Consultant at Metamedia Technologies

NLP Workshop

16:30 - 18:00

Solving NLP tasks using CLARIN-PL web services API

Michał Marcińczuk and Maciej Piasecki

Wrocław University of Technology

10:45 - 11:00

PolEval 2019 - introduction

Maciej Ogrodniczuk

11:00 - 11:15

PolEval Task 1 Results

Jan Kocoń

11:15 - 11:30


Marcin Bodziak

11:30 - 11:35

Award for the Task 1 Winners

11:35 - 11:50

PolEval Task 2 Results

Michał Marcińczuk

11:50 - 12:05


Zbigniew Bronk

12:05 - 12:10

Award for the Task 2 Winners

12:10 - 12:25

PolEval Task 3 Results

Aleksander Smywiński-Pohl

12:25 - 12:40

Cheeky Mouse

Szymon Roziewski

12:40 - 12:45

Award for the Task 3 Winners

12:45 - 13:00

PolEval Task 4 Results

Krzysztof Wołk

13:15 - 13:30


Marcin Chochowski

13:30 - 13:45


Łukasz Podlodowski

13:45 - 13:50

Award for the Task 4 Winners

Lunch Break

14:45 - 15:00

PolEval Task 5 Results

Danijel Koržinek

15:00 - 15:15


Marek Lange

15:15 - 15:20

Award for the Task 5 Winners

15:20 - 15:35

PolEval Task 6 Results

Michał Ptaszyński

15:35 - 15:50


Piotr Czapla, Marcin Kardas

15:50 - 16:05


Marcin Ciura

16:05 - 16:10

Award for the Task 6 Winner

16:10 - 16:25


Maciej Biesek

16:25 - 16:40

ensamble spacy + tpot + BERT

Renard Korzeniowski, Rafał Rolczyński, Przymysław Sadownik, Tomasz Korbak, Marcin Możejko

16:40 - 16:45

Award for the Task 6 Winner


We have a great lineup of speakers from science and business. Here we introduce you to some of them.

Book your ticket now


Early bird

  • Participate in all talks and workshops
  • Lunch included
  • Snacks and Coffee break
  • Early bird tickets available until April 15th
499 PLN


Regular pass

  • Participate in all talks and workshops
  • Lunch included
  • Snacks and Coffee break
  • Available until May 31st if tickets remaining
799 PLN
buy ticket

Proud Organizers

Our Sponsors

Media Partners

Additional questions?


  • Contact us


  • The administrator of your personal data is Sages sp. z o.o. with its registered office in Warsaw, Poland at Nowogrodzka 62c Str., registered by the District Court for the Capital City of Warsaw, XIII Commercial Division of the National Court Register under the number KRS: 0000313297, with a share capital of PLN 50 000 paid-in full. Tax ID: 1132737407. REGON: 141560043. Providing personal data is voluntary, but necessary to allow to contact you. Data shall be kept for a period necessary to realize the purpose. The disclosed data may be forwarded to any third party participating in the event's organization. You have the right to access, correct and delete your personal data. In the matter of personal data protection, the Administrator may be contacted in writing by e-mail visible above.