Machine Learning Conference

See you in
The place where engineering meets business and art. Applied machine learning & real use cases.

This year's leitmotif is "STATE OF THE ART", so be ready for the most advanced designs and ML implementations with a strong focus on practice and real-world use cases.

Meet experts who implement cutting-edge machine learning solutions in areas such as
medicine and healthcare, finance, industrial processes, transport and logistics, conversations and communication, design and art.


It's very important for us to deliver you valuable content without any marketing "games". We've given up cooperation with sponsors.
Each speaker that appears was carefully selected. We paid attention to practical experience and unique achievements. We chose from several hundred proposals, over 10 countries.

Join if you are...
ML engineer
Data Scientist
Head of Innovation
CxO who (want to) runs a business based on ML
Developer who wants to start working with ML
Geek-artist who uses ML
Start-up founder
You're hungry for knowledge of applied ML
You touch ML topic in another way
Be the first to know!
In order to be the first to receive the latest news, ticket sales and much more.
Subscribe to DWCC 2019 newsletter.

By sending the message I give my consent to process my personal data within the DataWorkshop project. The consent above is given to Biznes Myśli – Uladzimir Aliakseichanka, Grodzka 42/1 street, 31-044 Kraków, Tax Identification Number (NIP): 6751364881. You can withdraw your consent at any time.
What to expect
Practical ML knowledge
& use cases
The presentations will contain:
- a use case,
- a classic way of solving the presented problem,
- the latest production solutions (state of the art in ML),
- success metrics,
- useful technical and business advice.

Each presentation is a 30-minutes intense dose of practical knowledge of Machine Learning and applied AI.

Additionally, you will get from us:
  • recordings of all presentations
  • summary with practical tips and links to tools, libraries, etc. written by experts, so that you can explore the topic at home or at work.

Expert experience & motivating stories
Among our speakers, you can meet experts from small and large companies/organizations, who are actively working on new models and show the most advanced designs and ML implementations that solve real problems.

Don't miss the opportunity and join AMA sessions with Speakers!

AMA is a less formal form of Q&A and has one rule: ask me anything! So take a cup of coffee, go to the AMA zone just in time, sit comfortably next to the speaker and ask anything you want to know.
Effective game networking
If you want to go somewhere, it is best to find someone who has already been there (R.Kiyosaki).

We designed Game Networking to optimize the search for such people. During the hour you will meet from 25 to 30 people who have similar interests and different experience. You will not have to talk about the weather, don't worry. The moderator will keep his finger on the pulse and will take care of the topic of conversation and a good atmosphere.
Online transmission
Are you going to get married on September 28? :)

Don't worry. If you can't join DWCC for some reason, watch online transmission and keep your finger on the pulse of ML's innovations.

You can watch the live broadcast, or watch recorded lectures at a convenient time.

By joining the conference live, you have the chance to talk with participants and ask questions to experts via Slack. You will also receive materials after the conference: summary with practical tips and links to tools, libraries, etc. written by experts.

Subscribe to the list. You will receive an email when tickets for online transmission will be available.

Brainstorm & Discussion zones
How to avoid losing money and time for unsuccessful ML projects? What are common and less common mistakes in ML I should avoid?How to build a team of ML specialists?

If you have these and much more doubts related to the business side or your ML career choices, then don't miss this zone.

As Eleanor Roosevelt said: "Learn from the mistakes of others. You can't live long enough to make them all yourself". Experienced by failures and success in ML field people will help you save a few years of wandering.
Do you want to improve your technical skills?

If so, subscribe to the list. We will inform you about our workshops soon.
Creative technology & art zone
When you take technology and mix it with art, you always come up with something innovative, because often the art challenges the technology.

At the conference, you will meet a geek artist and see projects that combine both areas.

Open your eyes and mind for non-standard ways working with intelligent systems.
Unique ML community from different countries
Become a member of the ML community, learn from others and support others. Meet people with different ML experience and varied background.

Join our Slack and meet the conference participants before the event and stay in touch for a long time. You will receive an invitation to Slack a few days before the conference.

During the conference, you can ask questions to experts also via Slack.

We also regularly send practical doses of knowledge and news about ML (also after the conference) to help you always have your finger on the pulse. The world is changing faster than you think.

DWCC2019 in numbers
Meet our speakers
Tomas Mikolov
Research Scientist / Facebook
He is a research scientist at Facebook AI Research since May 2014. Previously he has been a member of Google Brain team, where he developed and implemented efficient algorithms for computing distributed representations of words (word2vec project). He has obtained PhD from Brno University of Technology (Czech Republic) for his work on recurrent neural network based language models (RNNLM). His long term research goal is to develop intelligent machines capable of learning and communication with people using natural language.

Machine learning, and natural language processing.

Emeli Dral
Chief Data Scientist / Mechanica AI
Emeli Dral leads the data science team at Mechanica AI. She is responsible for the development of core products and technologies for the application of artificial intelligence in industrial processes.
Prior to co-founding Mechanica AI, she served as the Chief Data Scientist at Yandex Data Factory. She led a team of accomplished data scientists and oversaw the development of machine learning solutions for various industries - from banking to manufacturing. Emeli is a lecturer at the Yandex School of Data Analysis and Harbour.Space University, where she teaches courses on machine learning and data analysis tools. In addition, she is a co-author of the "Machine Learning & Data Analysis" and "Big Data for Data Engineers" curriculums at Coursera.

Janusz Marecki
Senior Research Scientist / DeepMind
Janusz Marecki is currently a lead scientist at Google DeepMind and a co-founder of AI Machina. Prior to joining DeepMind, Janusz was a senior research scientist in the cognitive computing division of IBM T.J. Watson Research, a research assistant at the European Laboratory for Nuclear Research (CERN), a research associate at the Ukrainian Academy of Sciences and a full professor at the Academy of Computer Sciences in Poland. An author of over 100 refereed publications and 7 patents and a recipient of a commendation from the Los Angeles Airport Police, a commendation from the Department of Homeland Security and an Invention Award from an IBM CEO, Janusz's current research interests are in reinforcement learning, deep neural networks, and artificial brain computing.

Mikhail Burtsev
Founder and Leader /
Mikhail Burtsev is head of Neural Networks and Deep Learning Laboratory at Moscow Institute of Physics and Technology. In 2005, he received the Ph.D. degree from Keldysh Institute of Applied Mathematics of Russian Academy of Sciences. From 2011 to 2016 he was head of the lab of Department of Neuroscience at Kurchatov NBIC Centre. Now he is one of the organizers of NIPS 2017 Conversational Intelligence Challenge, and a head of iPavlov Project.

His research interests are in fields of Natural Language Processing, Machine Learning, Artificial Intelligence, and Complex Systems.

Aditya Guglani
Data Scientist / Uber
He is currently a data scientist at Uber where he is using deep learning and natural language processing to provide magical support experiences to customers and driver partners. He has a background in engineering and a graduate degree in data science, which enables him to solve some hard problems using data and machine learning. His interests currently include neural networks, document semantics, and language understanding.

Marta Markiewicz
Head of Data Science / Objectivity
Born in 1988, a graduate in Mathematical Statistics at the Wrocław University of Technology. Her master's thesis entitled "Stable feature selection methods" won the second place in the XLVI competition of the Polish Mathematical Society for the Best Student Work in Probability Theory and Statistics Applications. Professionally, over almost 8 years, she has been discovering the potential of data in various business domains - from medical data, through retail, HR, finance, aviation, real estate, ... She deeply believes in the power of data in every area of life.

Valentin Malykh
Applied Research Scientist / Huawei
Valentin has hand-on experience working with Information Retrieval, Natural Language Processing and even Self-Driving tasks. He worked as an engineer in Sputnik & Yandex, as an applied research scientist in iPavlov & VK. Now he is employed as a research scientist at Huawei Noah's Ark Lab.

Recently, Valentin defended a PhD on robustness in NLP tasks. His research interests are robustness to noise, dialog systems, summarisation. Valentin co-organized NeurIPS ConvAI challenges for Conversational AI in 2017 & 2018.

Łukasz Kidziński
Researcher at Stanford
Łukasz Kidziński is a researcher in the Mobilize Center at Stanford, working on the intersection of computer science, statistics, and biomechanics. He is a co-founder of -- a startup focusing on improving efficacy of clinical trials using artificial intelligence. Previously, Łukasz was a researcher in the CHILI group, Computer-Human Interaction in Learning and Instruction, at the EPFL. He obtained his Ph.D. at Université Libre de Bruxelles in mathematical statistics, working on time series analysis and functional data. He obtained two master degrees in mathematics and computer science, from the University of Warsaw. His research interests include time series analysis, large scale statistics, alternative data, and applications of data science in biological and medical sciences.

Łukasz Dziekan
Chief Technology Officer / FinAI
For the 10 years he is dealing with advanced analytics and production deployed machine learning systems. He started in Deloitte where he did complex risk analysis systems in credit risk and stress tests. Then he did 2 years in New York Unicorn Zocdoc where he was analytics tech lead and infrastructure senior developer. However he missed home and came back and built 50 person machine learning advisory team in PwC CEE, It was first such case in Poland. At the end of 2016 he found his entrepreneurship vibe and founded and became CTO. In FinAi he deals with simplifying retail finance – using Machine learning methods – in a way where they are designed to help rather than obstruct. They do things like fraud detection and face recognition and credit risk. And it's interesting, but very hard to do right.

Bartek Skorulski
Senior Data Scientist at Research Team / Alpha
Bartek Skorulski is Senior Data Scientist in Research Team in Alpha (Telefonica), a team that designs and builds machine learning solutions for improving people mental health. He has extensive experience in both industry and academic word. On one hand he was working in Messaging Team in Schibsted (company knows for portals like Fotocasa, Leboncoin, Infojobs, Segundamano) helping building experimentation driven products like, for example, smart replies in chats. In SCRM (digital hub of Lidl) he was setting up data driven environment, building recommender systems, forecasting demand and AB Test tool. In King (Activision/Blizzard) using data he was helping build best games. On the other hand, he was also working as Associate Professor doing research in Dynamical Systems and teaching graduate and undergraduate courses. Last few years he has been also teaching courses on machine learning, deep learning and data management.

Anna Anzulewicz
Head of Research / Harimata
Anna is a cognitive scientist with a background in psychology and data science, and a great interest in using new technologies in diagnosis, therapy, and education. She loves working in multidisciplinary projects and developing technology for social good. She combines her academic work at several universities with business.

For six years Anna worked as head of research at Harimata, a startup company that combines psychology and technology to support assessment and treatment of disorders in children and adults. During her time in Harimata, she developed AI-based solutions supporting early detection of autism spectrum disorder in children as well as applications helping adults who suffer from mood disorders.

Tomasz Jadczyk
CTO and Co-founder / Techmo
Tomasz Jadczyk is a co-founder and CTO of Techmo – a Voice Technologies company, where he is responsible for management and development of speech recognition, speech synthesis, voice biometrics, and voice analytics systems.
During his career, he moved from 'I implement this myself' through the implementation of State-of-the-Art research systems to the production environments, where ASR handles more than 60 000 calls every day. His research interests focus on moving voice processing systems from laboratory to real-life usage.

Daniel Baade
Daniel Baade is the founder and current Chairman of Art World Analytics. He holds a Masters and Ph.D. from Hanover University (GER) with further education at INSEAD (FRA) and Stanford (USA). He's holds senior roles in the art and finance world, including Ex-Global Strategy Director at Christie's - the world's largest art business and CEO of Soprano Mergers & Acquisitions.

Discussion Zone
Przemek Maciołek
VP of R&D / LogSense
Przemek's professional career marks over 15 years of extracting business value out of data. During that time he had his own Data Science consulting company, created a Big Data startup (failed one, bummer!), worked on exploring natural resources and built teams for companies such as Base CRM (currently Zendesk Sell) and Toptal, in the meantime he earning a PhD in NLP/ML field. Since 2015 he works for LogSense, building a data platform that can process millions of events for seconds, finding anomalies in the data. He's a big fan of cycling, you can frequently spot him around Kraków.

Anaig Marechal
AI Specialist / Microsoft
Anaig started her career at Microsoft in 2017, working on children's dyslexia risk evaluation using Machine Learning. Now AI specialist in Geneva, she's supporting the top Swiss companies to develop their AI strategy and build their Data Science projects with Microsoft cloud technologies. She is still very engaged in the AI for social good actions, and is currently working on a smart assistant for people under autism.

Bartłomiej Rozkrut
CTO & co-founder /
He started his adventure with the IT industry in 1998 as a network administrator, then he worked as a web application programmer. Based on the experience gained, he developed my competences by running IT projects. From 2005, he was responsible for the technological development of Empathy Internet Software House as CTO, and then from November 2012 in the Management Board of the Unity Group and as the Director of e-commerce development for the development of the proprietary Unity.Commerce platform and other proprietary products. Currently, he is the CTO at, implementing projects in the field of artificial intelligence in business.

Tomasz Trzciński
Chief Scientist / Tooploox
Adiunkt / Politechnika Warszawska
Tomasz Trzciński is a Chief Scientist and co-owner at Tooploox and an Assistant Professor at Warsaw University of Technology, where he leads a team of students behind He obtained his Ph.D. in Computer Vision at École Polytechnique Fédérale de Lausanne in 2014. He has (co)-authored several papers in top-tier computer science conferences and high impact factor journals. His professional appointments include work with Google in 2013, Qualcomm Corporate R&D in 2012 and Telefónica R&D in 2010. In 2017, he was appointed a Visiting Scholar at Stanford University. In 2016, he was named New Europe 100 Innovator as one of 100 outstanding challengers who are leading world-class innovation from Central and Eastern Europe.

Aleksandra Możejko
Machine Learning Engineer / Sigmoidal
Machine Learning Engineer at Sigmoidal with over 2,5 years of experience in Machine Learning, co-organizer of the PL in ML conference. Her main field of interest is Natural Language Processing. At her day-to-day work, she designs and develops AI systems for business clients from all over the world, in fields such as e-commerce, finance, and compliance.

Matthew Opala
Machine Learning Tech Lead / Netguru
Matthew is an entrepreneur, software engineer and machine learning practitioner. He's been working as a machine learning engineer since graduation from AGH University of Science and Technology. Before he moved to Netguru, he worked at Siemens Corporate Research in Princeton and machine learning software house Craftinity, where he led development of various machine learning projects including Deep Learning based solutions for image recognition and Natural Language Processing. Now he's responsible for building and managing the whole Machine Learning department at Netguru.

Łukasz Siatka
CTO / Lonsley
Lukas connects the dots to solve problems and simplify processes. He combines the experience in building Machine Learning products on scale with deep knowledge about the number of domains. He enjoys diving into product management, audio engineering and electronics engineering, as well as teaching others how to build answers to the questions and problems they bring in. He is a huge fan of Quantum Computing, XAI and generative art. Previously at two FAANG companies, currently at Lonsley. INTJ.

Krzysztof Sopyła
Ermlab Software CEO and scientist in the field of machine learning with 12 years of experience. I believe that artificial intelligence is a way to create superhuman people and will accelerate the evolution of our species. I co-create startups using the latest NLP achievements, currently working on polish grammarly and book recommendation engine. I write my thoughts on the blog.

Monika Koprowska
Value Proposition Consultant / Lingaro
AI & Data Science enthusiast with a passion for finding interesting stories and AI applications in the world driven by data. Gained professional experience bridging IT and business with Accenture, P&G, an AI startup, and others. Currently continuing this path at Lingaro as Value Proposition Consultant focusing on the business benefits of data solutions. Embraces the philosophy of "pay it forward."

Bartosz Ziółko
Co-founder and CEO / Techmo
Bartosz is a co-founder and CEO of Techmo - a technology company providing solutions in the field of speech recognition and synthesis. He studied Electronics and Telecommunication at AGH. Then he did PhD in Computer Science at the University of York. In 2017 he received a postdoctoral degree. Over 100 scientific papers, two patents granted by the USPTO and one by the EPO. Author of the book "Speech Processing".His work interests include automatic speech recognition, natural language processing, speaker recognition, and spatial sound simulation. He participated in several national and European research projects. His research and development activities are currently focused on the development of products and technologies within Techmo, mainly in the areas of customer care and telemedicine.

"Learn from the mistakes of others. You can't live long enough to make them all yourself" as Eleanor Roosevelt said.
Experienced by success and failures in ML field people can help you save probably few years of wandering.

Don't miss this Zone and join the selected discussion.
11:10 – 12:10
11:10 – 12:10
How to build an effective team of Machine Learning experts?
Experts: Anaig Marechal, Przemysław Maciołek, Tomasz Trzciński, Łukasz Siatka, Krzysztof Sopyła Moderator: Monika Koprowska
Development versus Production - Turning machine learning models into real products and services
Experts: Bartłomiej Rozkrut, Aleksandra Możejko, Mateusz Opala, Bartosz Ziółko, Łukasz Dziekan
Moderator: Przemysław Maciołek
Art Zone
Open your eyes and mind for non-standard ways working with intelligent systems. Meet creative coders who use machine learning in an unusual way to explore new areas of art and interaction.
Koki Ibukuro
Koki Ibukuro is a creative developer/artist who born in Tokyo and based in Berlin since 2018. He obtained a degree in electronic music and a master's degree in product design at Tokyo University of Arts, and now produces expressions that are not only visual media but from sounds, smells and touch. In addition to artworks, He produces many IoT products and permanent digital signage that is installed in Tokyo. His awards include Cannes Lions International Festival of Creativity, SXSW Innovative Technologies and more. He explores the possibilities of the use of ML in the CG pipeline process and creates Simulacra-objects using GAN.
Emilie Zawadzki
Emilie Zawadzki is an IT Developer, musician, and digital artist. She currently works at IRCAM as head of web development and teaches Web Audio / Web Midi to master students. She grew up with two passions : computer science and music. This led her up to a master of Multimedia Design Production at Gobelins School (Paris). She looks for lastest technologies trends which she likes to try out in artistic activities. In parallel, she is part of several music projects between electronic and improvised music via records and gigs.

Philippe Esling
Philippe Esling received a B.Sc in mathematics and computer science in 2007, a M.Sc in acoustics and signal processing in 2009 and a PhD on data mining and machine learning in 2012. He was a post-doctoral fellow in the Department of Genetics and Evolution at the University of Geneva in 2012. He is now an associate professor with tenure at Ircam laboratory and Sorbonne Université since 2013. In this short time span, he authored and co-authored over 20 peer-reviewed journal papers in prestigious journals. He received a young researcher award for his work in audio querying in 2011, a PhD award for his work in multiobjective time series data mining. In applied research, he developed and released the first computer-aided orchestration software called Orchids, commercialized at fall 2014, which already has a worldwide community of thousands users and led to musical pieces from renowned composers played at international venues. He is the lead investigator of machine learning applied to music generation and orchestration, and directs the recently created Artificial Creative Intelligence and Data Science (ACIDS) team at IRCAM.
We've also created a special space for you, where you will be able to enjoy art created thanks to ML and data. Find out more about a few amazing creators who agreed to show their work during DWCC 2019.
Michael Tyka
Mike Tyka joined Google in 2012 and worked on creating a neuron-level map of fly and mouse brain tissue using computer vision and machine learning. Mike also works with artificial neural networks as an artistic medium and tool. In 2015 created some of the first large-scale artworks using Iterative DeepDream and co-founded the Artists and Machine Intelligence program at Google.

Refik Anadol
Refik Anadol - is a media artist and director born in Istanbul, based in Los Angeles. He is working in the fields of site-specific public art with parametric data sculpture approach and live audio/visual performance with immersive installation approach, particularly his works explore the space among digital and physical entities by creating a hybrid relationship between architecture and media arts with machine intelligence.
Scott Eaton
Scott Eaton is an American artist, designer, and anatomy instructor residing in London, UK. His work explores the representation of the human figure through various mediums – drawing, digital sculpture, photography, and most recently generative AI.
Meet experts and projects from...
DataWorkshop Alumni Day
On September 27 at the Sound Garden Hotel will take place
the DataWorkshop Alumni Day as well.

This meeting is dedicated to all of DataWorkshop's courses graduates. This is a unique opportunity to exchange experiences, meet other graduates and learn how to successfully start a career in ML.
More about
Workshop: Analysis of e-commerce
During the workshop, you will learn the basic concepts related to e-commerce analysis and gain skills related to data processing, visualization and the selection of the right model to predict the future. You'll also see how to verify which model gives the best results. We will work on real data and solve a specific problem.

The workshop will be held in Polish and last all day onSeptember 27
More about

Conference Program
08:15 - 08:59
Morning Networking with coffee & cakes
09:00 - 09:20
Think Outside the Box
Vladimir Alekseichenko / Founder & CEO / DataWorkshop
09:20 - 09:50
09:50 - 10:20
Towards the Future of AI
Janusz Marecki / Senior Research Scientist / DeepMind
Recent years have seen a rising interest in developing AI algorithms for real world big data domains ranging from autonomous cars to personalized assistants. At the core of these algorithms are architectures that combine deep neural networks, for approximating the underlying multidimensional state-spaces, with reinforcement learning, for controlling agents that learn to operate in said state-spaces towards achieving a given objective.

The talk will first outline notable past and future efforts in deep reinforcement learning as well as identify fundamental problems that this technology has been struggling to overcome. Towards mitigating these problems (and open up an alternative path to general artificial intelligence), I will then summarize a brain computing model of intelligence, rooted in the latest findings in neuroscience. The talk will conclude with an overview of the recent research efforts in the field of multi-agent systems, to provide the future teams of humans and agents with the necessary tools that allow them to safely co-exist.

#deeplearning #reinforcementlearning #multiagent_systems #brain_computing #catastrophic_iInterference

10:20 - 10:40
Coffee break & Ask Me Anything with...
Vladimir Alekseichenko / Founder & CEO/ DataWorkshop
Dariusz Gross / DATAsculptor /
Janusz Marecki / Senior Research Scientist / DeepMind
10:40 - 11:10
Efficient algorithms for processing natural language
Tomas Mikolov / Research Scientist / Facebook
Machine learning algorithms are increasingly popular and useful in a wide variety of tasks. However, it is often confusing to understand what type of algorithm is the most suitable for what type of problem. In this talk, I will focus on text classification and representational learning and will describe some efficient solutions based on models optimized for computational efficiency, such as word2vec and fasttext.

#deeplearning #nlp #word2vec #fasttext #wordembeddings

11:10 - 11:40
Optimising industrial production with machine learning: electrolysis example
Emeli Dral / Chief Data Scientist / Mechanica AI
Using accumulated historical data, Mechanica AI developed a machine learning system able to predict which electrolytic cells are likely to underperform in the near future. The goal was to address the periodic efficiency decrease of electrolytic cells that leads to lowered production output and thus lost revenue.

Pilot results demonstrated that the machine learning system correctly identified at least twice as many underperforming cells compared to the current expert-led approach. Early alerts on unnoticed technical problems allowed plant experts to treat them in time, thus avoiding yield loss. The predictive system is now being deployed at the first chosen plant. The gains will result from the timely prevention of cell inefficiencies and a corresponding increase of production levels by up to 0.7%.

#industrialml #industrialai #sensor_data #gradient_boosting #lightGBM #interpretable_ai #shap_values

11:40 - 12:10
Running it on prod - how to do deployment for machine learning and not be scared
Łukasz Dziekan / Chief Technology Officer / FinAI
They say a data scientist will either bloom or wilt after he/she deploys her first model to production where real decisions will be made, without human intervention. This presentation will be all about dirty, not so glamorous, getting the job done, but still state of the art part of machine learning engineer job - managing the whole process ending on production. Combining our 2-year experience of running extremely important biometrics model on production for the banking sector I will explain that moving beyond Jupyter, getting this 1% more on AUC is really way more important - because this differentiates between making a real impact and affecting people's lives - and staying in the research part. FinAi has built it's platform over 2 years - and while it is by no means finished, will tell you some stories from the trenches, hoping you will learn, maybe be scared a little (and it's healthy) and move you to the next step - having a model working online.

#reproducible_results #model_versioning #model_monitoring #model_deployment #best_practices

12:10 - 12:40
Coffee break & Ask Me Anything with...
Tomas Mikolov / Research Scientist / Facebook
Emeli Dral / Chief Data Scientist / Mechanica AI
Łukasz Dziekan / Chief Technology Officer / FinAI
12:40 - 13:40
Game Networking
An effective way to meet new people with similar interests and needs.
13:40 - 14:40
14:40 - 15:10
Track 1 - Symphony
Track 2 - Opera
Data augmentation in speech recognition
Tomasz Jadczyk / CTO and Co-founder / Techmo
In order to deliver an Automatic Speech Recognition system capable of handling 50 000 speakers each day, one needs a few thousand hours of transcribed recordings, with the variety of speech types and styles covered. Since transcribing an hour of audio takes at least 8 hours, production of training data is expensive.
The solution comes with the data augmentation, which makes it possible to build reliable models based on the limited data resources. By using it, we have managed to decrease word error rate in production from 30% to stable 5-10%, depending on the case.
Since in machine learning every solution brings with it brand new problems, limitations of the presented methods and compatibility issues in the production environment will also be discussed.

#nlp #data_augmentation #speech_recognition #dnn #overfitting
Artificial Intelligence in Mental Health
Bartek Skorulski / Senior Data Scientist at Research Team / Alpha
One out of four adults experience mental health problems at a certain point in life, including anxiety, depression, sleep disturbance, phobias, etc. Limited access to healthcare services results in that most cases remains undiagnosed and untreated. At Telefonica Alpha (Health Moonshot) we try to solve this problem by building mobile apps that provide personalised coaching, therapy, monitoring, and crisis prediction. These apps collect passive data (from mobile sensors) and active data (through questionnaires). Then, on one hand, have built could-based infrastructure with tools like airflow, spark, pytorch which allows us to deploy and test our ML algorithms like recommender systems and behavioural models. On the other hand, we collaborate with neuroscientists or psychologists from institutions like MGH, UCL, who help us in establishing a hypothesis, interpreting the developed models and testing our products with clinical trials.

#deeplearning #recommender_system #activelearning #mentalhealth #digital_phenotyping

15:10 - 15:40
Track 1 - Symphony
Track 2 - Opera
Zero- and Few-Shot Learning for Text Classification Valentin Malykh / Applied Research Scientist / Huawei
Here, at Huawei, we are solving different real-world tasks. Our team spends most of our time on dialog systems.
One of the problems we are solving is how to choose the right skill in each situation. And more interestingly, what if we have (almost) no training set? Traditionally, for such tasks, TF-IDF and word2vec are used. We use currently more and more popular BERT and its successors. We are integrating our developed solution into large-scale dialog system, which arises some new challenges. In my talk, I'll try to describe to you our approach and demonstrate its results on open data. On it, we are achieving 80% vs 40% for traditional approaches. Hopefully, you could get these ideas and apply them to your problem solution.

#deeplearning #nlp #text_embeddings #fewshot_classification
Deep learning for clinical trials in movement-related diseases
Łukasz Kidziński / CTO at , Researcher at Stanford
Every second person in the world will be affected by symptomatic movement disorders at some point in their life. Proper treatment of these disorders requires very precise measurements of body mechanics. Collecting data requires sophisticated motion capture hardware and a few hours of work of highly trained medical personnel, bringing the cost to more than $2k per visit. We propose a deep learning approach where expensive hardware and trained personnel is replaced by mobile phone and a neural network. We are now recruiting patients to track their movement at their homes. Measurements derived using our method correlate at r = 0.7 - 0.8 with most of the tested metrics, enabling their clinical use. In this talk, I will present all the steps I had to take to execute this medical data science project, from data sharing agreement with a hospital until deployment. I'll discuss the impact of our solution on drug development.

#deeplearning #reinforcementlearning #computer_vision #drugs #x-ray #mri #healthcare #bioengineering

15:40 - 16:10
Track 1 - Symphony
Track 2 - Opera
Developing conversational AI systems with DeepPavlov open-source framework
Mikhail Burtsev / Founder and Leader /
An open-source DeepPavlov library is specifically tailored for development of dialogue systems. The library prioritizes efficiency, modularity, and extensibility with the goal to make it easier to create dialogue systems from scratch with limited data available. It supports modular as well as end-to-end approaches to implementation of conversational agents. In DeepPavlov framework an agent consists of skills and every skill can be decomposed into components. Components are usually trainable models which solve typical NLP tasks such as intent classification, named entity recognition, sentiment analysis or pre-trained encoders for word or sentence level embeddings. Sequence-to-sequence chit-chat, question answering or task-oriented skills can be assembled from components provided in the library. ML models implemented in DeepPavlov have a performance on par with the current state of the art in the field.

#deeplearning #nlp #chatbots #open-source #deeppavlov

Play.Care: bridging psychology and technology for better autism assessment
Anna Anzulewicz / Head of Research / Harimata
Play.Care, developed by Harimata, analyses movement patterns that children make while using a tablet device and assesses the risk of autism spectrum disorders. The solution uses state of the art CNN architectures to analyse users' behaviour during the gameplay. Additionally, features that characterise movement kinematic characteristics are classified with boosted trees and RNN-s. This new, automated approach to detection of developmental disorders can make a difference in the way in which we learn about children's health and make the initial assessment faster and widely available. The results of three major studies (including 520 participants) conducted so far, has shown that the solution has an overall sensitivity 80% and specificity of 89%. Currently, the solution is undergoing clinical study, which will further assess its effectiveness.

#xgboost #deeplearning #cnn #rnn #healthcare #diagnosis

16:10 - 16:40
Is prediction STILL difficult (especially if it's about the future)?
Marta Markiewicz / Head of Data Science / Objectivity
Who wouldn't like to know what future holds? Humans were intrigued with future mysteries since the beginning of time, and so are we at Objectivity. Forecasting revenue is of especially high importance for us, so over last few years, we've tried several approaches. The first model was productionized in 2016 and it evolved nicely over the years, with various outcomes, coming closer and closer to the actual values. Currently achieved MAPE is around 3%. In the speech I would like to share the pros and cons of different methods we've tried (classical time series forecasting, Monte Carlo, ML), to facilitate this journey for others.

#forecasting_revenue #time_series #monte_carlo #stacked_ensemble
Ask Me Anything with...
Tomasz Jadczyk / CTO and Co-founder / Techmo
Bartek Skorulski / Senior Data Scientist at Research Team / Alpha
Valentin Malykh / Applied Research Scientist / Huawei
Anna Anzulewicz / Head of Research / Harimata
16:40 - 17:00
Coffee break & Ask Me Anything with...
Aditya Guglani / Data Scientist / Uber
Mikhail Burtsev / Founder and Leader /
Łukasz Kidziński / CTO at, Researcher at Stanford
17:00 - 17:30
Can you use Machine Learning to predict the value of fine art?
Daniel Baade / CEO / Soprano Mergers & Acquisitions and Chairman / Art World Insights
Art is Passion, Culture, Inspiration…But art is also an investment and an asset class. Some pictures sell for $100 others sell for $100million. How do you determine the value of an asset class which doesn't produce cashflows and where the range of values is greater than 1,000,000X? Using time-series analytics might be the obvious answer. However, as the average time between two sales of the same paintings can be more than 30 years, this is not a practical solution. We use a form of data decomposition technology to solve this problem. Combining market knowledge, data analytics and machine learning we are able to value art accurate, fast and based on statistical evidence.

#art #artdata #arttech #awi #artasinvestment #data_decomposition
17:30 - 18:00
COTA: Improving the customer support experience using Deep Learning
Aditya Guglani / Data Scientist / Uber
As the leading ridesharing platform, Uber receives thousands of support tickets daily. The tickets are in different global languages and range from missing item from an EATS order in India to fare adjustment request in North America. To scale up the ticket resolution process, we built COTA (Customer Obsession Ticket Assistant) an intelligent system based on machine learning (ML) and natural language processing (NLP) techniques that are integrated with Uber's customer support platform. It provides customer support representatives, suggestions for the ticket type, appropriate reply and relevant actions to take based on ticket text and additional context such as trip information. As a result, we were able to reduce the average number of re-routings by 15% and the time an agent spends on a ticket by 5% on average, thus making the support experience a little more magical.

#deeplearning #nlp #embeddings #ai_assistant #customer_care #customer_satisfaction
Ask Me Anything with...
Daniel Baade / CEO / Soprano Mergers & Acquisitions and Chairman / Art World Insights
Marta Markiewicz / Head of Data Science / Objectivity
18:00 - 18:10
Closing Conference
18:10 - 21:00
Music & Networking
Get the Last Ticket
Don't worry!
If you can't join DWCC for some reason, watch online transmission and keep your finger on the pulse of ML's innovations.

You can watch the live broadcast, or watch recorded lectures at a convenient time.
By joining the conference live, you have the chance to talk with participants and ask questions to experts via #Slack. You will also receive materials after the conference: summary with practical tips and links to tools, libraries, etc. written by experts.
Go online
What you'll get after the registration
Access to all sessions, including AMA, discussion and art zones.
Expert knowlege
You can ask experts anything during the AMA session and discuss your problem with experts in a dedicated zone.
All day food
Stay hungry for knowledge! We'll take care of the rest.
Video pack after the conference
Few weeks after the conference you'll get a pack with all recorded talks.
Post-conference publication
After the conference, you will also get a summary of all lectures with extra practical tips and links to tools, libraries, etc. written by experts.
Including access to dedicated Slack before and after the conference, game-networking session.
Follow DataWorkshop
September 28, 2019

Żwirki i Wigury 18,
02-092 Warsaw
If you need accommodation and want to be as close as possible to DWCC 2019, it will definitely be the Sound Garden Hotel, where the conference takes place.

We have a 15% discount for conference participants. Use password: DataWorkshopClubConf during booking on the hotel's website step by step or save time and click here.

How to get to the DWCC2019

From Chopin Airport - 2 km distance, the fastest way is to take a taxi, there are also two bus
lines 175 or 188. Start your journey at bus stop "Lotnisko Chopina", after 5 minutes you will departure to the bus stop "Novotel".

From City center - 6 km distance, you can take taxi or a bus number 175. Start your journey at bus stop "Centrum", after 30 minutes you will departure to the bus stop «Novotel".

From Warsaw Central Railway Station - 5 km distance, you can take taxi or a bus number 175 as well as 504. Start your journey at bus stop "Dworzec Centralny", after 25 minutes you will departure to the bus stop "Novotel".

Plan your trip to Warsaw

Find things to do during your stay in Warsaw for #DWCC2019 including days out, attractions and sightseeing, what's on, theatre, tours, restaurants and hotels in the city.
Plan your trip to Warsaw with useful traveller information. Visit Warsaw's official city guide to find out more.
Call for Experts
"Call for Speakers" is closed. Thank you for every application. It's time to make a decision and it's not easy... We received very interesting applications from many countries such as USA, Italy, Australia, Norway, Switzerland, the Netherlands and of course Poland :)

Acceptance Decision: July, 31

After July 31 the list of speakers will be closed, but we are still open to cooperation with people who actively explore the field of machine learning in various ways. Don't hesitate to write to us as well if you are a geek artist who creates works using ML.

If you have any questions, please contact us:

Rest of the DWCC2019
Don't Miss a Thing!
Join it Right Now
Meanwhile look how it was in 2018
DataWorkshop Club Conf 2018
Code of Conduct
All conference participants are equal. Forget during the event about your daily roles (chairman, mother, father, programmer, artist, student, etc.). Be open to new experience and respect others and their experience. Be human and treat others equally.

All attendees, speakers, and volunteers at our conference are requested to agree with the Conference Code of Conduct throughout the event. We expect cooperation from all participants to help ensure a safe and pleasant environment for everybody.

We do not tolerate harassment of conference participants in any form. Conference participants violating these rules may be sanctioned or expelled from the conference without a refund at the discretion of the conference organizers.

If you are being harassed, notice that someone else is being harassed, or have any other concerns, please contact immediately a member of the conference team.


Filipa Eisenberga 11/16,
31-523 Kraków, Polska
By sending the message I give my consent to process my personal data within the DataWorkshop project. The consent above is given to Biznes Myśli – Uladzimir Aliakseichanka, Filipa Eisenberga 11/16, 31-523 Kraków, Polska. Tax Identification Number (NIP): 6751364881. You can withdraw your consent at any time.
Interested in Cooperation the #DWCC2019 Conference? Contact us at
This site uses cookie and other technologies. By using it you agree to their use.
I understand and accept :)