Agenda

Machine Learning Week Europe
14-18 June, 2021

Coffee Roundtables – grab your real coffee and share experiences virtually with your peers to explore the new and old challenges. Just like pre-show breakfast in a regular conference you’ll join a “round table” with fellow attendees and see where the conversation takes you.

Kick-starter: Share the impact the pandemic has had on your working environment, interaction with colleagues, management of projects and processes. Do you see digital transformation in your organization being accelerated as a result and what lasting effects do you think it will have on your career and working environment once the pandemic is over?

1. Deep Learning for Computer Vision in the Manufacturing Industry – Insights from Johnson & Johnson (Vaibhav Verdhan)

Deep learning and Computer Vision are changing the way defects can be identified in manufacturing industry. Johnson and Johnson has used cutting edge neural network architectures to identify the source of problems in products. It improves the product quality and enhances the customer experience. The models are deployed into production and are generating fantastic results. It will be surely a great attraction for the visitors who want to analyse the significance of deep learning, identify the process and challenges. It will be a first hand information for them which will prove to be really useful and will have far-reaching results.

2. Bringing Research to Production: Modular and Highly Adjustable Computer Vision Pipelines (Thomas Wollmann)

Robust detection, segmentation, and classification of objects are crucial for computer vision applications in production. However, the challenges in real world use cases slow down fast adoption of new research findings. Academical efforts are seldom mature enough and ignore the context beyond a proof of concept. In this talk, Thomas presents how SemiAutoML and their modular and highly adjustable computer vision pipelines accelerate bringing research to production and solve real world use cases.

3. Generative Deep Learning for Image Anomaly Detection (Luca Marchesotti)

Supply chains are becoming extraordinarily efficient with the help of sophisticated machinery. These systems are designed to perform repetitive tasks with near-zero efficiency loss, but they are still unable to deal with rare and potentially harmful situations – long-tail events. In this talk, we will discuss how you can set up a fault detection system in your supply chain leveraging computer vision solutions to drastically boost your quality control performance.

1. Introduction to Climate/Sustainable Finance & Its Link with AI (Pedro Baiz)

   The Financial Stability Board (FSB) after working on regulations to address the 2008/9 financial crisis, moved to the next biggest risk: Climate Change. The talk will provide a comprehensive overview of what constitutes Climate/Sustainable Finance. Standards and potential upcoming regulations, such as TCFD (Task Force on Climate related Financial Disclosures) among many others (e.g. SASB, GRI, IRFS, etc) will be covered. Finally, the role of AI and on this emerging field will be discussed.
   
   

2. Predictive Analytics for Climate Risk Assessment (Christian Spindler)

   Climate change is a systemic risk which impacts on all business sectors. It increases uncertainty and investment risk and endangers entire business models. Professional investors and asset managers are taking climate change more and more into account. Corporates are starting to quantify climate risk in their mid- and long-term strategies. Predictive analytics turns out to be key in making quantified assessments in mostly unexplored terrain: How are extreme weather risks impacting on production sites and physical assets of the firm? How is the upcoming carbon taxation impacting on the company now and in future? And how vulnerable is the global supply chain of the company against business interruption risks? This presentation explores methods and tools for climate risk quantification.
   
   

3. Sustainable Federated Learning for Predicting Climate Changes (Sharmistha Chatterjee) 

   The talk unveils the art of building sustainable federated machine learning models by considering different aspects of ethical AI. The talk will highlight various techniques of incorporating fairness in private federated learning while ensuring the sustainability of future smart ecosystems. By the end of the talk, the audience will get the know-how of sustainable federated learning, how to build a scalable distributed architecture, and important KPIs to focus on when designing such a system.

1. Information Extraction: Key Learnings from Real NLP Projects (Jona Welsch)

Information Extraction and Natural Language Processing (NLP) have a large variety of business applications, ranging from the digitalisation of invoices or purchase orders to the automatic interpretation of free-form text. This talk is about key learnings and insights from projects for industry clients and public institutions, which were successfully brought into production. Learn about the pitfalls when processing PDF-Documents or where to effectively use Graph Neural Networks.

2. Boosting Search with Latest NLP – Transformers, Dense Vector Similarity and More (Timo Möller)

Nowadays almost every English Google search result is powered by a Transformer – a Language Model that is hard to scale in production. In this talk, we will dive into some of these modern search methods, show how to improve document retrieval via dense encoders, return exact answers via Question Answering and scale those pipelines to production workloads. Everything will be illustrated with code from the open-source framework Haystack.

3. Deep Learning for Natural Language Processing in Real Life Product (Chongko Snitwong Na Aryuttaya)

Implementing data science to product is hard. Why? We only know machine learning is cool. Everyone wants to have it but when data scientists try to fit machine learning to the business product, business people just wants proof of concept. At the end of the day, the results would be only on .csv file, nobody touches anymore. That’s why it is essential to know data product how it starts, how it ends. As a data science product manager, these things can combined!

1. Turning Data Science into CO2 savings (Daniel Rohr)

“Can I drive more efficiently?” was the question to be answered at the beginning of all data science efforts at Tracks. This is a typical real-life question that is difficult to solve with ML methods. The answer to this question has sparked a number of follow up business questions that we are tackling with Tracks’ complex AI system. In this session you will learn how to turn ML into saved CO2.

2. Finding Sensitive Intervention Points by Mapping the Global Fossil Fuel Supply Chain (Jannes Klaas)

Over 84% of the worlds energy comes from fossil fuels. But oil, gas and coal are also a major transport good. By identifying the networked structure of the global fossil fuel supply chain we can identify sensitive intervention points. We present our work in creating an asset level network of the global fossil fuel supply chain. We will highlight some early research avenues enabled by this dataset, including the use of new modelling techniques.

3. Applying Machine Learning to Solve the Ocean-Bound Plastic Crisis (Bosse Rothe)

11 million tonnes of plastic are estimated to end up in our oceans every year. Cleanhub developed a digital solution that can scale plastic collection in high-impact countries. To deliver digital evidence about how much plastic was collected, all collection partners use CLEANHUB’s software to track and trace the entire recovery process. Detecting anomalies is key to mitigate fraudulent behaviours. We will discuss the problem, opportunities and the role of machine learning.

1. Data Efficient Deep Learning Using Self-Supervised Learning (Igor Susmelj)

The availability of large datasets is one of the main drivers for machine learning. However, common approaches such as supervised learning from labels have become a bottleneck. Self-supervised learning is a new approach where we train models with proxy tasks that allow models to learn without explicit supervision and helps with downstream performance on tasks of interest. In this workshop we will look at modern approaches commonly used in self-supervised learning such as contrastive learning and train our own models. You will learn that we can directly influence learned feature invariance by introducing different data augmentation strategies.

2. Machine Learning on Graphs – Hands-on Approach & Current Challenges (Samuel Lopez Santamaria)

Graphs are mathematical objects used to model from social networks and financial transactions to interactions in quantum chemistry. In recent years, the interest in graph representation learning has grown intensely. Today, data scientists find a plethora of algorithms and tools at their disposal. Samuel provides a hands-on approach, surveying applications of graph learning, reviewing current challenges in the field and demonstrating which Python libraries are best suited to get started.

3. Clustering of COVID-19 Clinical Articles Using Graph Community Detection and Bio-BERT Embedding (Debasmita Das)

A lot of scientific attention got directed towards understanding causes of COVID-19 & impacts of the virus, resulting in large amount of research articles being published everyday. It is, thus, of particular importance to have a method to efficiently categorize such documents with minimum effort for smooth navigation. I will present a graphical network-based unsupervised clustering method on a corpus of scientific articles. Our approach is much simpler and is also computationally efficient.

1. Data Can Feed the World. But Do We Have the Right Data? (Sharavani Basu and Sébastien Foucaud)

Amid the current health and economic crisis, another chronic but fast deteriorating one is the food crisis! Not to sound alarmist, but shifting climatic patterns, trade & geopolitical instability, rising population among others has brought millions to the brink of poverty and starvation. As the productivity gains through mechanization and chemistry-based solutions in an intensive mode is fast plateauing or even reversing due to resistance build-ups, Digital Agriculture & Precision Farming, powered with advanced Analytics and Data Science seems to hold much promise in building sustainable systems with reduced environmental impact. Challenges to use data in agriculture is an experience shared across industries: aggregating data at large scales can be difficult, but identifying the right data is where the real challenge lies! We will present different use cases of novel pest outbreak forecasts, where despite abundant data availability, new approach for data acquisition is urgently needed.

2. Challenges & Best Practices: How to Handle Weather Data in Forecasting Models with Success (Christian Schneider)

Weather happens all the time with an undeniable influence on behavior, such as consumption, and therefore also on commercial success. Though one can´t change the weather, there are many options to utilize weather for one’s advantage. However, weather is not just weather. What are the special characteristics of weather compared to other data? Which machine learning models are most suitable? Christian Schneider talks about challenges and opportunities using weather in scalable forecast solutions.

 3. ESG in the News: Text Mining for Sustainability Signals (Gerhard Rolletschek)

Many current ESG rankings lead to absurd results. Greenwashing occurs when companies can rank high in ESG scores by streamlining their reports and figures instead of truly reducing their ecological footprint. Using text mining based on predicate-argument-structures and using a few-shot learning             approach, we show how ESG-relevant signals in multiple dimensions can be extracted from unstructured news sources and how this can lead to a more balanced and fair assessment of ESG activities.

Meet the speakers, fellow attendees, sponsors, moderators – randomly for a quick chat, just like in real life. If you are a match you can exchange contact details with one click. If not, you simply move on to the next contact.

1. Introduction to Explainable AI Using a Real World Demo (Julia Brosig)

Explainable AI helps to understand decisions of black box models and thus improves confidence in machine learning models. XAI methods provide global or local insights into the black box model’s decisions. For this presentation, various global and local interpretable machine learning methods were applied to a Munich rent index based use case. The winning methods were implemented and visualized in a dashboard, understandable even for non-statisticians.

2. The State of Secure and Trusted AI 2021 (Alex Polyakov)

To raise security awareness in the field of Trusted AI, we started more than a year ago a project to analyze the past decade of academic, industry, and governmental progress. The eye-opening results reveal an exponential growth of interest in testing AI systems for security and bias and the absence of adequate defenses. This research aims to make companies aware of insecure and malicious AI by sharing insights formed based on almost 2000 research papers.

3. Responsible AI: A Cross-Industry Overview (Aishwarya Srinivasan)

With the accelerated use of Machine Learning and AI technologies, having a comprehensive view of the use of these technologies, the data involved, and how the technology interacts with the users have become diagnostically complicated. The presentation shows varied areas where industries are building user-facing AI applications, understand the responsibility of the organization to govern and follow certain regulations around building these.

1. ESA Earth Observation Data Generation, Management and Access (Nicolaus Hanowski)
ESA is operating the most productive and sophisticated fleet of Earth Observation satellites in the world. Understanding the data content, the way it is processed and made accessible to everyone should facilitate new applications. New missions are relied for generating entirely new Earth Observation data sets and provide new opportunities for new applications. The presentation will illustrate easy data access options. The potential of “predictive” Earth Observation will be explained.

2. Why Open EO Data Should be Accessible for Everyone and How They Could Help Solve Our Global Issues (Niklas Jordan)
Earth observation data from public space agencies, such as the ESA, is available to everyone free of charge, but not everyone can access it. Technical and professional requirements make it challenging to access this treasure trove of data. In my presentation, I will discuss how, in addition to science and professional users, civil society, such as journalists, teachers, NGOs, and almost all citizens, can benefit from this data to solve our time’s major global problems.

3. Convolutional Neural Networks for Damage Assessment and Disaster Relief (Thomas Chen)
Natural disasters ravage the world’s cities, valleys, and shores on a monthly basis. Having precise and efficient mechanisms for assessing infrastructure damage is essential to channel resources and minimize the loss of life. Using a dataset that includes labeled pre- and post- disaster satellite imagery, we have conducted research training multiple convolutional neural networks to assess building damage on a per-building basis. In order to investigate how to best classify building damage, we present a highly interpretable deep-learning methodology that seeks to explicitly convey the most useful information required to train an accurate classification model. Participants in this session will learn about why interpretability is important and the ramifications of AI for disaster management. Our research seeks to computationally contribute to aiding in this ongoing and growing humanitarian crisis, heightened by climate change.

Choose the topic that is relevant to you. In this session we offer focused topics to dig in deeper into specific content segments. So this is the perfect opportunity to share your story, your specific question and get first hand advice from your peers and experts.  All attendees can participate in the roundtable of their choice and it’s a camera on event.

Current Challenges for Securing AI Applications (Alex Polyakov & Eugene Neelou)
In April, the EU released requirements for AI. The requirements ensure that AI applications are trusted and can be used in highly critical environments. One of them is ensuring that AI applications are secure and safe. We will discuss who should be responsible for securing AI applications in an organization, what the first steps to operationalize AI security are, as well as what the differences for securing AI in various environments and the latest approaches practically applicable.

Barriers for Machine Learning Applications in Manufacturing and How to Overcome Them (Nikolaus Jäger-Grassl)
In production environments, across all industries, data is collected in a massive amount and over a long period of time. Furthermore the question we should ask ourself is: Why is it so hard to convince even base customers to use this data in machine learning applications? In this table discussion we’ll look into main reasons for that and how to convince customers to trust you in building applications and processes that helps to achieve business goals and increase security.

More Roundtable Topics will be announced soon

Europe mastered the corona crisis, but a bigger one is already on the list: the climate crisis. The EU set ambitious goals for their climate change programme, but still the question remains how to reach them. Hopes in ClimateTech are thus huge and AI is to expected to play a key role. However, AI is also a driver to CO2 emissions: it consumes a lot of energy and it is also used e.g. to detect new oil deposits. How helpful is AI really and is Europe prepared to lead in and with AI the climate movement? Finally, what political and legal changes, technical infrastructure, economic conditions and so on are required to strengthen the climate tech community?

Predictive maintenance allows equipment operators and manufacturers to assess the condition of machines, diagnose faults, and estimate time to failure. Because machines are increasingly complex and generate large amounts of data, many engineers are exploring deep learning approaches to achieve the best predictive results.  In this talk, you will discover how to use deep learning for a condition monitoring of an air compressor using an audio-based fault classifier.

“Watching paint dry is faster than training my deep learning model.” / “If only I had ten more GPUs, I could train my model in time.” / “I want to run my model on a cheap smartphone, but it’s probably too heavy and slow.” – If this sounds like you, then you might like this talk. Exploring the landscape of training and inference, we cover a myriad of tricks that step-by-step improve the efficiency of most deep learning pipelines, reduce wasted hardware cycles, and make them cost-effective. We identify and fix inefficiencies across different parts of the pipeline, including data preparation, reading and augmentation, training, and inference. With a data-driven approach and easy-to-replicate TensorFlow examples, finely tune the knobs of your deep learning pipeline to get the best out of your hardware. And with the money you save, demand a raise!

Keynote 1: Why It’s Time to Quit Your Big Tech Job and Work on Climate (Eugene Kirpichov)

Most climate-concerned engineers are still unaware that they can directly work on climate solutions, because of the outdated idea that climate is about activism and non-profits. This may have been true a few years ago, but could not be further from the truth today. I quit Google AI last year to work on climate, and I will try to convince you, too. I’ll give an overview of the commercial climate tech ecosystem, the role ML plays in it, and how to find a climate job as an ML practitioner.

Keynote 2: Tackling Climate Change with Machine Learning (Nikola Milojevic-Dupont) 

Collaborations between researchers and academics are essential to scale up solutions and enable GHG emissions reduction at scale. I will give an overview of existing academic research at the intersection of climate change and machine learning. I will also present Climate Change AI, an organization of volunteers facilitating impactful applications of machine learning to climate change, highlighting resources that can be helpful for companies aiming to work in this space.

Join the moderator of the day for a wrap up of content, discussions, findings and an outlook.

Join the moderators of the day for a wrap up of content, discussions, findings and an outlook.

Coffee Roundtables – grab your real coffee and share experiences virtually with your peers to explore the new and old challenges. Just like pre-show breakfast in a regular conference you’ll join a “round table” with fellow attendees and see where the conversation takes you.

Kick-starter: Share the impact the pandemic has had on your working environment, interaction with colleagues, management of projects and processes. Do you see digital transformation in your organization being accelerated as a result and what lasting effects do you think it will have on your career and working environment once the pandemic is over?

Many manufacturing companies are not yet very data-driven but willing to use their data to optimize production. In that case, one of the most challenging tasks of data experts is to establish an understanding from top management to all employees about circumstances that are prerequisites when working with data. Managers often tend to think in terms of tools – a well-designed front-end delivers the analytics solution – ignoring that proven data quality and data reliability are key to this. In this session, some actions are explained which have triggered a steady change in thinking processes at REHAU – thinking in solutions instead of tools and set up a sustainable data strategy. However, developing an useful data-driven solution for the manufacturing industry requires strong collaboration of industry, technical and data experts. Right from the beginning, all important stakeholders have to be involved and highly motivated even if they have their daily business. Sandra will explain what challenges they faced within the interdisciplinary teams and how they successfully integrated the different experts and created efficient communication channels to avoid misunderstandings.

1. AutoML in the Factory: How to Empower Engineers to Adopt AI by Weidmüller (Markus Köster)

Developing industrial analytics solutions usually requires specific know-how in the data science domain. In the engineering domain, experience in data science is sparse, which prevents unleashing the power of artificial intelligence and machine learning on the factory shop floor. This talk highlights challenges and our experience in implementing AI analytics models for real-world machinery applications, by addressing why a pure data driven approach does not lead to satisfying models. To summarize, Markus names five factors for successfully adopting AI and ML in the engineering domain.

2. Controlling Manufacturing Processes with Causal Models (Maksim Greiner)

Modelling manufacturing processes often suffers from limited training data and spurious correlations. With Bayesian networks we can combine data and process knowledge to amplify the data and reduce spurious correlations, the most important part being the causal structure of a process. This session will present several examples for the manufacturing industry (e.g., from Festo and BMW), explain how causal structure relate to Bayesian networks and how they can be used to predict the outcomes and calculate optimal parameters.

1. Automatic Detection of Railway Track Defects at Goldschmidt Group (Artur Suchwalko and Marcin Kowalski)

Ensuring railway track safety is crucial. Currently, the track (rail, sleeper, fastener) defect detection relies foremost on human visual inspection. Achieving full automation of the track inspection (its defect detection) is important for ensuring track safety and to reduce maintenance cost. For this purpose, GRAW (Goldschmidt Group) with the help of QuantUp developed a custom software product. Rail track images used for analysis were acquired from a linear camera. Computer Vision, Machine Learning and optimisation methods were applied to detect sleepers being cracked, broken, or covered by ballast. This session start off by presenting the solution. Next, it will describe the challenges met along the way and how they were managed in business and analytical dimension

2. Predictive Maintenance: The Curse of Little Failure Data (Marcus Groß)

A major focus of Industry 4.0 is the optimization of manufacturing by analyzing the impact of production parameters on quality. The central conflict of this domain is avoiding errors first of all while deriving insights about failure causations from the very limited amount of failure data. Framed with a case study, we show how the application of suitable AI algorithms can solve this special type of “explore-or-exploit” dilemma and identify complex failure patterns in the production process.

3. Predictive Maintenance in Action: Leakage Localization (Arvin Arora / Nils Funke)

Air leaks lead to high energy losses in pneumatic systems. The result is unplanned downtime & high costs. Early detection & precise location of leaks can reduce maintenance costs & unplanned downtime. The challenges of a method suitable for practical use here are: few signals, heterogeneity, complexity, ltd. training data, nonlinear behavior, localization. We will discuss how you can successfully meet these challenges, using the example of an already implemented use case with Emerson (Aventics).

Meet the speakers, fellow attendees, sponsors, moderators – randomly for a quick chat, just like in real life. If you are a match you can exchange contact details with one click. If not, you simply move on to the next contact.

1. Automated Demand Forecasting in Production at Continental (Amit Tyagi and Lars Schleithoff)

In the Tire Division of Continental, demand planning is crucial as an input for the supply chain and previously involved mainly manual forecasting of almost 100 business experts. In this talk we will reveal how the Continental Advanced Analytics team, together with Informationsfabrik, created a machine learning framework that – deployed on a state-of-the-art infrastructure – today automates large parts of this tedious task. In particular, we will give insight into the problem complexity, concrete improvements achieved as well as the technical and – importantly – organizational challenges that arise from automating manual processes.

2. Deli Salad Sales Forecast at Kugler GmbH (Antje Dittmer and Maria-Alexandra Cimpeanu)

Kugler Feinkost GmbH manufactures a large variety of deli salads and distributes them to their customers – beer gardens, restaurants, and shops. With Machine Learning techniques a production forecast resulting in 100% deliverability and only 0.3% waste is achievable with suitable features e.g. ‘WeekdayAverage’, but also ‘IsGoodBeerGardenWeather’. It will be discussed how predictive analytics can help make informed decisions in a low-tech environment using the knowledge of domain experts.

3. AI Algorithms for Forecasting Container Traffic – Is More Complexity Always Better? (Sven F. Crone)

AI/ML researchers are preoccupied with the most advanced and complex algorihtms, notably deepnets and xgboost, but fail to deliver POC results. In this talk we will explore 30+ forecasting algorithms in AI, ML and statistics for a logistics case study of real-world container traffic forecasting at Hamburg harbour. We develop a landscape of algorithm accuracy versus complexity, runtime, and robustness which shows that simple algorithms designed with domain expert know-how improve performance.

Robert Weber & Peter Seeberg publish a highly recognized podcast on AI in industry. In this session they will moderate a live podcast panel discussion with Nicole Büttner, Co-Founder & CEO Merantix Labs and Carina Mieth, Advisor to the Managing Director Research & Development at TRUMPF about Machine Learning in Europe. They will discuss such topics as maturity level, markets, education, technology, the EU proposal for harmonized rules on AI and where Europe is standing relative to the US and Asia.

1. Predictive Maintenance in Data Centers with Self-Learning AI for NTT (Britta Hilt and Kentaro Ono)

NTT FACILITIES provides critical facility management services for more than 7,000 data centers worldwide. Cost-expensive action is taken to minimize risk of air con failure, like replacing critical components prior to their actual life time end. Activities were started to use self-learning multi-layer AI with the objective to maximize compressor run time, to decrease replacement and to avoid additional backups. This Japanese project was executed with the support of the German company IS Predict. The reliability of the solution was beyond NTT FACILITIES´ expectations who had already executed similar AI project. Accuracy in failure prediction for air condition system compressors of 98% was realized.

2. Artificial Intelligence for Beyond5G Network Management – Challenges and Opportunities (Rachana Desai & Edwin Yaqub)

Global rise in demand has exacerbated the limitations of current 4G networks. Therefore, 5G and Beyond (B5G) networks are being designed to deliver the performance levels expected by the next generation of applications. In the EU research project Ariadne, the goal is to employ Machine Learning and Artificial Intelligence techniques in various B5G deployment scenarios, with the objective to reconfigure network resources as required to ensure continuous reliable high-bandwidth connectivity.

3. Using Machine Learning for Root Issue Analysis in Telco’s Networks (Silvia Veronese)

Root issue identification in large scale network is a key component to proactive fault analysis. In this presentation we describe a methodology which integrates multiple data domains for telco networks driven by big data sets. We aim to automate the intelligence required for operations, networking, services and care/support. This starts with predicting service impacting failures (broken devices, sub-optimal performance, changes in utilization, etc.) primarily in real-time. With this capability for the NOC or SOC, we strive to deliver reduced MTTR ( Mean time to repair) as a business value but in an actionable (and later in an automated) manner. These analytics capabilities extend to the subscriber who can be enabled to solve their own device/service issues. This self-care is a holy grail for CSPs (Cable Service Providers) as they endeavor to dramatically reduce their support costs.

1. Using Machine Learning to Optimize Power Plant Operations at Uniper (Tobias Mathur)

This session shows how Uniper used machine learning to increase the efficiency of a waste-to-energy power plant resulting in lower emissions and higher revenues in terms of heat, power, and gate fees. The presentation will focus on experiences and results from the field. We will furthermore discuss the relevance of this technology and how it can disrupt the power generation industry in the era of sustainability, energy transition, increasing competition, and ongoing global health crisis.

2. Predictive Maintenance by using Digital Twin in the Hydropower Plant of tomorrow for Verbund Hydro Power (Teresa Alberts)
The hydropower plant of the future will be a transparent power plant. What that means is that comprehensive, detailed information about the current and expected state of the plant will be available on a continuous basis, ideally precise to the second. With this, Verbund Hydro Power can identify the sweet spot in operation between performance and availability, balancing revenue and maintenance cost. Verbund Hydro Power GmbH – one of the major European hydropower companies – aims to achieve this goal using digital twins and virtual sensors. The implementation of digital twins is shown based on the turbine of a hydroelectric power plant at the example of the Austrian hydropower plant in Rabenstein. A technical set-up as well as lessons-learned and challenges during the implementation will be discussed.

3. Predictive Modelling of the Aerocondensor Performance of the Luminus CCGT Power Plant (Wout van Alphen)
In order to produce electricity and fulfill its grid stability services to the grid operator, Luminus needs to make accurate forecasts of the maximum production capacity of its CCGT power plant in Gent. Predicting the performance of the plant is a challenging problem, as it depends on a multitude of internal and external (e.g. weather) conditions. In this talk, we will briefly illustrate the AI solution that was implemented, and highlight the challenges and key takeaways of the solution process.

Join the moderator of the day for a wrap up of content, discussions, findings and an outlook.

Coffee Roundtables – grab your real coffee and share experiences virtually with your peers to explore the new and old challenges. Just like pre-show breakfast in a regular conference you’ll join a “round table” with fellow attendees and see where the conversation takes you.

Kick-starter: Share the impact the pandemic has had on your working environment, interaction with colleagues, management of projects and processes. Do you see digital transformation in your organization being accelerated as a result and what lasting effects do you think it will have on your career and working environment once the pandemic is over?

ML, Data, and AI were destined to be THE game changers in the financial services industry. Data scientists came up with approaches that improved the business rather than just claiming big (or creating regulatory risks); they learned to manoeuvre around big data, bad data, legacy technology, and even GDPR. We will share several successful real-world examples that are productive today and revisit them in the light of GDPR as well as the coming EU Artificial Intelligence Act’s proposal.

Time series are everywhere in science and business, and the ability to forecast them accurately and efficiently can provide decisive advantages. For much of its history, time series forecasting has mostly been relying on “classical” statistical methods such as ARIMA. These methods work very well in many cases, but they are not appropriate for capturing patterns in large quantities of data. Very recently, deep learning techniques have been proposed as a way to build very advanced and accurate models from large quantities of time series data. Darts by Unit8 is a Python open source library that provides ready-to-use implementations of all sorts of forecasting models. It puts emphasis on reducing the experiment cycle duration and improving the ease of using, comparing and combining different models. In this talk, we will present how darts can easily be used to solve real business problems and show the simplicity to switch from classical statistical regression models to deep learning techniques.

– Learn more about darts – https://medium.com/unit8-machine-learning-publication/darts-time-series-made-easy-in-python-5ac2947a8878

– Access to the code – https://github.com/unit8co/darts

1. How to Scale AI in Financial Services Institutions (Michael Soucek)

Financial services institutions were data driven long before big data became a buzz-word. So why are many organizations struggling to get full potential of data at core of their business? Our point of view talks on three building blocks successful data driven organizations have in common. First, an operating model allowing to focus on the right processes. Second, technical stack allowing to scale data products. Third, data culture which is different in historically data heavy organizations.

2. Causality in Banking and Financial Use Cases (Maksim Sipos)

Typical machine learning models fit historic data based on correlations. These models fail when environments change (for example when a macroeconomic regime shift occurs). Causal models are more robust and generalize better. However, inferring causality from data is hard. In this talk, we will describe what techniques we use to build causal models and the benefits we have observed in consumer lending and portfolio risk use cases.

3. Real-Time Feedback in Algorithmic Trading (Andreas Petrides)

Algorithmic trading revolves around optimal scheduling, when and how to trade each asset of a portfolio. The schedules produced, however, should not be static; being able to optimally adapt to changing intraday market conditions and analysing effectively other exogenous information is of paramount importance. In this talk, we focus on how machine learning methods can enrich the toolbox available for real-time algorithmic trading.

1. B2B2C Fraud Prediction with Granular Accounting Data as a Valuable Predictor at CRIFBÜRGEL (Sebastian Schnelle and Andreas Kulpa)

Companies need machine learning models that predict B2C and B2B payment defaults with higher selectivity than traditional scoring methods. Get an insight about the results of the CRIFBÜRGEL model using gradient boosting technique and the requirements to implement non-parametric algorithms. Applied to the prediction of B2B payment defaults, due to a lack of data no increased selectivity could be detected. Sebastian and Andreas will show you which new data source for accounting data will fix this problem.

2. Predictive Modeling in Quote & Buy – From Idea to Production (Nina Meinel)

Quote & buy processes in finance/insurance are heavily trying to maximise conversion by managing leads. Integrating predictive analytics into the flow increases the conversion rate and therefore influences companies targets. The journey starting with a rough idea, getting accepted as data science and implementing a productive model is shown during the presentation. A few deep dives are taken into areas as model evaluation, and key take-aways from MVP to a productive model with focus on data.

3. Leveraging Unstructured Data in Insurance (Raymond van Es)

Description: Insurance companies are used to work with structured data, for instance to build their pricing and underwriting models. In analyzing their unstructured data they are just scratching the surface. There is a big potential in this area because insurers have a lot of unstructured data at their disposal. They have vast amounts of textual data (policies, policy conditions, clauses, customer correspondence), image data (pictures of insured objects or claims) and speech data (call center conversations). In this session I will go into some of the use cases and methods for analyzing these unstructured data with a focus on text and speech.

Meet the speakers, fellow attendees, sponsors, moderators – randomly for a quick chat, just like in real life. If you are a match you can exchange contact details with one click. If not, you simply move on to the next contact.

1. Image Forgery Detection System for Digital Assets (Vijay Maharajan)

Digital Assets are getting a great amount of traction in recent times. NFTs – Non Fungible Tokens, being one such example, has a lot of Digital Assets in its marketplace. But there are people who does manipulate a digital asset and try to sell it on another NFT marketplace. To prevent this, this session presents the development and demonstration of an Image Forgery Detection System.

2. The Role of AI in Assessing Climate Related Risks (TCFD) and ESG Investments (Pedro Baiz)

TCFD (Task Force on Climate related Financial Disclosures) recommendations continue to gather support across the financial sector (e.g. Central Banks, Blackrock, etc) demonstrating how the entire financial system is changing. The talk will provide a guiding light on the current ESG maze, starting from a comprehensive overview of what is Sustainable Finance (starting from “materiality” and focusing on “metrics” and “analytics”), using as an example the Real Estate sector (VaR estimations).

3. Image Construction and Recognition for Finance (Rafaëlle Botter)

Convolutional Neural Networks (CNNs) are great for images recognition because they can identify patterns between images despite changes of color, location or orientation. Nonetheless, “images” of the financial market are not the same as an image of an animal or an object or a letter. This presentation is about how to build “images” for finance and how to set the CNNs in order to optimize image recognition of the financial market.

Choose the topic that is relevant to you. In this session we offer focused topics to dig in deeper into specific content segments. So this is the perfect opportunity to share your story, your specific question and get first hand advice from your peers and experts.  All attendees can participate in the roundtable of their choice and it’s a camera on event.

Other Roundtable Topics will be announced soon

Analytics & AI Governance (Frank Poerschmann)

Discussion between experts on the rising demand for actively managing and governing AI & Analytics activities, as i.e. managing model portfolios, continuous evaluation of the value of analytics solutions and their adaption, how to scale limited resources and manage the interface between business, data science, data management and IT. Participants shall exchange personal learnings and questions around frameworks, best practices as well as personal experiences on challenges and solutions.

Machine learning and transparency: Why and How? (Tom Alby)

With more and more advanced algorithms entering the data labs of the finance industry, some stakeholders feel uncomfortable with the lack of transparency that some algorithms offer. Regulatory authorities may even demand that the internals of a model are disclosed, e.g. to understand whether risks such as discrimination exists. In this table discussion, we want to discuss experiences and approaches to this challenge.

Driving a Data-Driven Mindset Across the Enterprise (Nora Reich)

Advanced analytics solutions imply significant value for a company but also high complexity. Crucial success factors are corporate values and norms that include a comprehensive understanding for the contribution of data-driven work for sustainable company development. However, we often see intuitive instead of data-driven decision marking, lack of understanding for data potentials, spotlight on near-term solutions, etc. Let’s talk about how we can enhance data literacy and understanding.

1. Case for the AI Regulator (Shashin Mishra and Sray Agarwal)

In this talk we make a case that an independent regulator is needed to create the standards and the guidelines for the adoption of the technology across industries. Expecting regulators for specific industries will lead to inconsistent standards and may also leave most of the industries without properly defined standards at best or at worst with no regulatory oversight on how the technology is being used.

2. Artificial Intelligence in Regulatory Processes: Chance or Risk? (Clemens Kraus)

Regulatory authorities find it difficult to assess artificial intelligence. The use of machine learning components is solely viewed critically and a detailed risk assessment is often neglected despite benefits of the solution. We would like to shed light on why the supervisory authorities still have such a hard time with artificial intelligence and highlight the chances of using AI systems in audit for repetitive tasks without sacrificing quality and accuracy.

Join the moderator of the day for a wrap up of content, discussions, findings and an outlook.

Coffee Roundtables – grab your real coffee and share experiences virtually with your peers to explore the new and old challenges. Just like pre-show breakfast in a regular conference you’ll join a “round table” with fellow attendees and see where the conversation takes you.

Kick-starter: Share the impact the pandemic has had on your working environment, interaction with colleagues, management of projects and processes. Do you see digital transformation in your organization being accelerated as a result and what lasting effects do you think it will have on your career and working environment once the pandemic is over?

1. Gaining Efficiencies through Scalable Data Science Assets (Tobias Lampert)

Putting Data Science projects into production is one of the main challenges organisations struggle with nowadays. Projects often get stuck in the PoC phase and need to be fully reimplemented to meet production quality requirements. Development frequently starts from scratch and rarely benefits from previous work. Tobias shows how these issues can be overcome by making components reusable, which criteria scalable components need to meet and how organisational challenges can be solved.

2. Beyond Data Sanity Checks: Machine Learning Data Quality Assurance (Lisa Maag and Lana Caldarevic)

GfK is hosting the world’s largest retail panel to track products and deliver insights based on actual sales data. One essential aspect of the reliability of our insights is the continuous quality assurance of our ML lifecycle, particularly the quality of the models, which differs from ordinary quality checks. In this talk you will learn how we built an extensive data validation pipeline that goes far beyond usual panel data health checking routines to ensure the precision of the models.

3. Code Quality – Bridging the Gap between Data Science & Engineering (Jean Metz and Pamela Hathway)

One major challenge of deploying machine learning models to production is technical debt created during the research phases. In this talk, we show how GfK speeds up the transition from proof-of-concept to production by ensuring easy access to data, high-quality coding standards, and automation. We present our learnings, discuss challenges, and how to find the right balance of code quality improvements whilst respecting core responsibilities and different skills- and mindsets.

Predictive analytics reaches out into more and more areas of business, industrial, and research applications. The sheer number of different algorithms and technologies is staggering. In this presentation we give a top-level review of the most popular tree-based algorithms now easily available to anyone through Minitab. From individual trees to powerful modern ensembles, we highlight their strengths, weaknesses, uses, and limitations, especially when compared to the conventional modeling techniques like multiple linear and logistic regression. We illustrate the flexibility of the modern tree-based predictive analytics by building a series of models to predict power generation of a solar energy power plant. In this case, gradient boosting ensemble achieves 20% more accuracy compared to the conventional regression. In addition, the winning model provides great insights into the nature of multivariate dependencies.

1. Lessons Learned: Building a Keyword Extraction Service for Bibliographic Metadata (Daniel Wrigley)

   The right keywords drive revenue: They ensure better visibility, optimised findability and higher rankings of publications. But curating metadata is a tedious process and outsourcing is expensive. This was the motivation to build a service to automatically extract keywords from metadata. Finding the right toolset in order to find the right keywords was a huge challenge. I want to share what we learned in the areas of NLP (e.g. lemmatization, ML on unstructured data) and ETL.
   
   

2. Data-driven Personas – Understand Your Customers’ Behavior (Alwin Haensel & Paul Kleinschmidt)

   Marketing is all about understanding customer behavior and anticipation of their responses and engagement. Classical marketing relies on general knowledge and a lot of gut feeling. But data can tell us who our customers are and how they react to campaigns and interact with offers over time. We propose an approach for customer segmentation based on temporal behavior clusters. This provides a customer interaction understanding, which enables marketeers to optimal target & time marketing campaigns.

1. Personalization for Regional News Media  (Jonah Kresse)

Regional publishers produce online content every day, but only a small fraction of this content gets a chance to be seen by readers on the publishers website. Personalization techniques used in other areas of business (e.g. e-commerce) can help publishers to show the right content to the right users, thereby increasing media time and gaining subscriptions. In our presentation, we present findings from personalization experiments conducted with several regional publishers from Germany.

2. Beyond automation: how to create innovations with AI – Where is the Tesla of the media industry (Peter Gentsch & Sebastian Döring)

Most AI applications in the media area focus on optimizing and automating processes rather than creating innovations. The talk shows various use cases and practical examples for AI-augmented processes and functions. Furthermore, it will be discussed how AI can be used to expand business areas and may be create completely new business models. Where is the Tesla of the media industry?

3. Demonstrate the Benefits of AI Applications Through Prototypes (Steffen Kühne)

AI does not live in a box and the usefulness of AI depends on the infrastructure around it. However, creating the necessary infrastructure for AI and machine learning is fairly difficult and tedious. Rapid prototyping helps us to showcase the benefits of AI and initiate a debate on what data we really need to thrive as a media organisation. Our prototypes show, that structured content and metadata is key for building new products and repurposing valuable content that already exists.

Meet the speakers, fellow attendees, sponsors, moderators – randomly for a quick chat, just like in real life. If you are a match you can exchange contact details with one click. If not, you simply move on to the next contact.

1. Expert Round: Extracting Insights with Cluster Inspection Toolkit  (Annemarie Paul)

   Extracting meaningful insights from clustering of a given dataset is hard and laboursome. We propose a toolkit containing three techniques to derive insights from arbitrary clustering solutions and / or low-cardinality classifications faster:
   1. Cluster plotting: Feature inputs are usually high dimensional. To represent the cluster solution in a low dimensional (~plottable) space, Uniform Manifold Approximation and Projection (UMAP) is employed.
   2. Cluster assignment: Which features and what interactions drive the clustering? The cluster assignment can be reverse-engineered in an interpretable manner using gradient boosted trees and shaply values.
   3. Feature inspection: The contribution of a specific feature on the global clustering space can be approximated by combining a low dimensional UMAP projection with a kernel ridge regression.
   The python-based toolkit will be shared with the audience.
   
   

2. How Deep Learning Helped to Classify Large-Scale B2B Marketing, Sales, and Web Traffic (Rohit Kewalramani & Justin Chien)

   The account engagement platform at 6sense involves analysing traffic flow through customers’ website, marketing automation, and customer relationship tools. Prospect accounts visit certain webpages and interact with marketing / sales teams – all influencing their buying decision. Our algorithm standardises data from these sources and uses NLPs context and word sequence to predict categories that are part of buying intent models. We leverage ONNX that boosts inference by 10x (i.e. 8.5M URLS/day).

3. Expert Round: Predicting Customer Attrition during COVID-19 – (Evelina Stoikou & Daniel Hellwig)

   As the world adjusts to the post-pandemic environment, the nature of customer attrition is changing. The pandemic has shortened the attrition cycle, altered the underlying causes, and changed the profile of customers who call in to cancel their service. At the same time, historic data if often not applicable anymore. Predictive analytics can be used to better understand customer attrition by tracking geo-specific indicators, segmenting customer groups in new ways, and creating new features.

Choose the topic that is relevant to you. In this session we offer focused topics to dig in deeper into specific content segments. So this is the perfect opportunity to share your story, your specific question and get first hand advice from your peers and experts.  All attendees can participate in the roundtable of their choice and it’s a camera on event.

AI at Scale – the Good, the Bad and the Ugly (Erika McBride)

Businesses both large and small are attempting to keep up with, or get ahead of, the competition by developing, deploying and adopting AI at scale. This journey is riddled with pitfalls, from technology to talent, data quality to culture. Join fellow attendees as we discuss hurdles and triumphs, along with lessons learned at Dow, Inc., on the journey towards Dow’s Analytics Everywhere vision.

Data Science and Design: Driving User Adoption and Business Value through UX (Brian O´Neill)

The dirty secret of analytics and data science, especially in the enterprise, is that a whole lot of the work never gets used. Low or no adoption is rampant, and in fact, one leading expert from the International Institute for Analytics thinks the problem is getting worse. However, more tech and data isn’t the solution. If you work with AI/ML/Analytics or Design/UX, join Brian T. O’Neill for a lively Q/A on how design and UX can help drive adoption and business value in your organization.

Selling Data Science: How to Explain Deep Learning Performance to Business Leaders ( Benjamin Ziomek)

“99% Accurate:” Often the phrase that gets the most attention from business leaders looking at a new data science-based product. But as practitioners know, Accuracy is a meaningless term. Unfortunately, throwing out “F1,” “Recall,” and such don’t usually move the needle with business executives. This presentation talks through how to frame data science products in a way that business leaders can understand, and how to get to the heart of what they care about.

The Data Science group at The New York Times develops and deploys machine learning solutions to newsroom and business problems. Re-framing real-world questions as machine learning tasks requires not only adapting and extending models and algorithms to new or special cases but also sufficient breadth to know the right method for the right challenge. This session first outlines how unsupervised, supervised, and reinforcement learning methods are increasingly used in human applications for description, prediction, and prescription, respectively. It then focus on the ‘prescriptive’ cases, showing how methods from the reinforcement learning and causal inference literatures can be of direct impact in engineering, business, and decision-making more generally.

Join the moderators of the day for a wrap up of content, discussions, findings and an outlook.

Coffee Roundtables – grab your real coffee and share experiences virtually with your peers to explore the new and old challenges. Just like pre-show breakfast in a regular conference you’ll join a “round table” with fellow attendees and see where the conversation takes you.

Kick-starter: Share the impact the pandemic has had on your working environment, interaction with colleagues, management of projects and processes. Do you see digital transformation in your organization being accelerated as a result and what lasting effects do you think it will have on your career and working environment once the pandemic is over?

With the severity of the COVID-19 outbreak, we characterize the nature of the growth trajectories of counties in the United States using a novel combination of spectral clustering and the correlation matrix. We capture the nature of counties’ growth in cases using growth communities, demographic factors, and social distancing performance to help overnment agencies utilize known information to make appropriate decisions regarding which potential counties to target resources and funding to.

1. Reducing Distressing Ambulance Transports of Older Adults with Predictive Modeling (Jorn Op den Buijs)

Germany – a super-aged “society” with a growing number of chronically ill elderly – is among the top spenders on healthcare. We use machine learning based on home health data to timely intervene with high risk patients and avoid distressing, costly emergency department visits. Initial results show significant reduction in ambulance dispatch rate. This talk will cover how to deploy prediction models with targeted prevention to facilitate independent living by older adults.

2. Leveraging Graphical Models to Assist Healthcare System (Vidhi Chugh)

The use of machine learning is needed to aid the healthcare system today more than ever. Pandemic has put a lot of pressure on the medical system and doctors are working round the clock to diagnose the symptoms and save the mankind. We are leveraging the power of the graphical networks to learn the joint probability distribution of the multivariate system and assist the end to end analytical solution including predictive and prescriptive analytics.

3. Implementing a Bayesian Adaptive Design for a Clinical Trial Sample Size. A Case Study From LivaNova (Giovanni Ranuzzi)

It is presented a use case of Bayesian adaptive design for a clinical trial to select the ‘optimum’ sample size based on accumulated data. During the study, frequent sample size selection analyses are produced using ‘R statistics’, and predictive probabilities are calculated with a simulation approach and used to drive decisions to stop/continue the trial. The methodology is known as Goldilocks, as it is constantly asking the question, “Is the sample size too big, too small, or just right?”

1. TriAl: Text and Data Mining for Drug Project Prioritization (Markus Bundschus)

The TriAI methodology is based on the observation that publication patterns can be considered as early success indicators for clinical developments. The ultimate goal of this project was to complement portfolio
decisions by providing rankings to decision makers that are based on machine learning on retrospective data. TriAI won the German Roche Innovation Award for Digitalization. In this short presentation we will present lessons learned how to best deliver machine learning results to the end-user in order to enhance trust for the algorithms.

2. Discovering Key Topics from Real-World Medical Inquiries via Natural Language Processing at Bayer – (Angelo Ziletti)

Millions of unsolicited medical inquiries are received by pharmaceutical companies every year. It has been hypothesized that these inquiries can give insight into medicinal products and associated treatments. Here, we use natural language processing and unsupervised learning to discover key topics in real-world medical inquiries. The discovered topics are meaningful and medically relevant, thus demonstrating that unsolicited medical inquiries are a source of valuable customer insights.

3. Healthcare Data Platforms and Machine Learning at Cochrane (Julian Everett)

Cochrane’s mission is to promote evidence-informed health decision-making by producing high quality systematic reviews and other synthesized research evidence. Their work is internationally recognized as the benchmark for high-quality information about the effectiveness of health care, and informs over 90% of the guidelines produced by the WHO. They have recently undertaken a number of initiatives which are collectively transforming their core data and content production pipelines via a mix of machine learning, crowd and knowledge graph technologies. This is reducing the lead time and cycle time for new evidence to be published, greatly enhancing its discoverability, and enabling a near-realtime capability to produce baseline datasets that summarise the current state of knowledge about specific clinical research questions. This is in turn enhancing access to the latest research findings for point-of-care clinical decision making. In this talk, you will learn how Wardley Mapping and Domain Driven Design techniques can be used to guide strategic decisions about where and how machine learning, crowd-sourcing and knowledge graph techniques can best be deployed to solve real-world business problems. It will also explore how to ensure competitive advantage can be maximised within budget limitations via a relentless focus on core domain market differentiators, and techniques for successfully managing the change impacts of machine learning on business workflows.

1. Explainable AI in Deep Brain Medicine (Afsaneh Asaei & Ozgur Polat)

This talk summarizes the key elements of explainable AI in development of medicine for brain pathology. The project is a joint collaboration between Digital Product School of UnternehmerTUM and the Schoen Clinic of Neurology at Munich for objective evaluation of patients with Parkinson’s disease. We will compare two paradigms of deep learning: One fully explainable and rich for defining the biomarkers of neurodegeneration contrasted with a typical architecture for end-to-end learning.

2. Computer Vision Methods for Skin Cancer Recognition (Karol Przystalski)

Pattern recognition of images is one of the most popular approaches that is used in machine learning solutions supporting medical doctors. We show to use image processing methods to do simple image analysis to find skin cancer cases. In the next step, we use neural networks and simple white-box methods to recognize skin cancer patterns on multilevel images. As it is a medical solution we show the process of commercialization of an AI solution for medtech and explain what kind of explainable AI methods can be used to be compliant with the FDA requirements.

3. How AI Can Help Us Unlock Nature’s Immense Therapeutic Potential (Sona Chandra)

Despite millenia of knowledge on the vast therapeutic potential of plants, we have explored less than 1% of the molecules found within this kingdom of life. Traditional methods for understanding the activity of natural products require laborious, resource-intensive experimental assays. We will discuss how machine learning can be used to accelerate the discovery of bioactive compounds, with a concrete example of how we illuminated the MOA of polyphenols with regards to their anti-inflammatory properties.

Meet the speakers, fellow attendees, sponsors, moderators – randomly for a quick chat, just like in real life. If you are a match you can exchange contact details with one click. If not, you simply move on to the next contact.

1. How to Bring AI Medical Devices to the Market. A Case Study From Roche (Gloria Macia)

Healthcare is emerging as a prominent area for AI applications, but innovators aiming to seize this chance face one major issue: achieving regulatory compliance. Using a real industry case study, Gloria Macia walks you through the current American and European regulatory frameworks for medical devices and provides a step-by-step guide to market for AI applications, highlighting the main challenges and pitfalls to avoid as well as the key issues a company needs to consider to succeed in this endeavor.

2. The Role of TensorFlow, PyTorch, and Friends in a Medical Device – the Regulatory Perspective (Oliver Haase)

Machine learning libraries such as TensorFlow, PyTorch, or XGBoost play a key role in the development of virtually any ML-based software as a medical device. From the regulatory perspective, they are third-party software that must be validated to be run in a medical device. This presentation will give an overview on why ML libraries are much more critical than 3rd party software in traditional software development, and how validation reuse can be leveraged to minimize the validation effort.

Choose the topic that is relevant to you. In this session we offer focused topics to dig in deeper into specific content segments. So this is the perfect opportunity to share your story, your specific question and get first hand advice from your peers and experts.  All attendees can participate in the roundtable of their choice and it’s a camera on event.

Digital and Date Literacy in Pharma. Nice-to-have or Must-have? (Boris Adryan)

Digital and data skills are in demand across all industries. Likewise, there’s a growing gap between those “who know” and those “who don’t”. How much data literacy is needed in the workforce along the pharma value chain? For example, how much conceptual knowledge does a bench scientist need to fully understand and adhere to the FAIR principles? In pharma production, are traditional Six Sigma tools enough, or should practitioners have a notion of machine learning? Does everyone need upskilling?

What can Healthcare learn from the Financial Sector? (Dennis Diederix)

Data science brings enormous potential for our society. As the techniques are fit to solve a multitude of problems, we see the use in different sectors. So how far is each sector with data science an AI? And what can healthcare learn from other sectors, such as the financial sector? What makes the healthcare sector different in use of data science? And which problems are common? As a consulting firm working in both the healthcare, financial and public sector, we understand the differences.

Challenges & Trends in the Health Sector During and After Corona (Martin Szugat)

Corona changed everything, especially the health sector. It has put a lot of pressure on the system, but also fostered new innovations. The pharmaceutical companies and hospitals have shown that they could react quickly and work effectively. But many weaknesses have also been revealed. In this round we discuss what has changed with Corona and especially what a post-Corona world might and should look like.

1. Corona Community Pitch: Maps Matter – Realistic Hot Spot Detection Across Regional Boundaries (Sebastian Cattes)

One standard visualisation of Covid-19 is a colored map showing the total infections per region. This does not inform about variations within and across regional boundaries and therefore diminishes the insights that could be derived from clustered data. This talk presents a novel kernel-heaping technique that generates a realistic picture of the spatial-temporal evolution of Covid-19. This approach allows identifying emerging infection hotspots at 30% higher accuracy than district-level data.

2. Corona Community Pitch: The Interplay of Demographic Variables and Social Distancing Scores in U.S. Covid-19 Cases (Francesca Tang)

With the severity of the COVID-19 outbreak, we characterize the nature of the growth trajectories of counties in the United States using a novel combination of spectral clustering and the correlation matrix. We capture the nature of counties’ growth in cases using growth communities, demographic factors, and social distancing performance to help government agencies utilize known information to make appropriate decisions regarding which potential counties to target resources and funding to.

3. Corona Community Pitch: The Process Perspective in Healthcare: Leveraging Process Mining for COVID-19 Event Data Analytics (Marco Pegoraro)

The recent increase in the availability of medical data, possible through automation and digitization of medical equipment, has enabled more accurate and complete analysis on patients’ medical data through many branches of data and process science. This talk illustrates some preliminary findings obtained with established process mining techniques in regard of the medical data of patients of the Uniklinik Aachen hospital affected by the recent epidemic of COVID-19.

Join the moderator of the day for a wrap up of content, discussions, findings and an outlook.