FEBRUARY 18 / 1-5 PM

ADS

PRESENTS

TECH020

WORKSHOPS

Smart ROCE: Making the difference with AI – Smart ROCE for Fiber Rollout

WHO /

KPN

WHEN /

Round #1

13:15 - 14:00

With a customer base of > 3 mio households and > 3,5 mio mobile subscriptions in the Netherlands, our network forms part of the basic infrastructure of this country. Ever more during the pandemic, we have all experienced how essential and comfortable a good internet connection is. With our ambitious goal to bring glass fiber to an additional 500 k households every year, we will further connect and improve the digital experience of the Dutch society. While this goal will help us provide 50% of all household in NL with high-speed internet, the question is: where shall we roll out first? Given the immense digging costs for laying fiber into the ground, we aim to increase the return on investment on those network investments.

With smart ROCE, we have implemented a process that involves AI in its various forms: using capacitated k-means clustering, custom network-routing algorithms, minimal cost flow optimization, value prediction with regression and XGBoost, scenario simulation, humans in the loop, full-stack software (e.g. Vue.js, Django, Docker, AWS, Kubernetes, …), automation of decision processes, and eventually an end-to-end use case that is constantly evolving. With the feedback we receive from our business stakeholders we reevaluate the process, with new data we retrain automatically on our model management platform. The beauty of this use case lies in the joint effort of the different company divisions, the use of state-of-the-art technology and algorithms, but also its power to boost our position in the market with AI.

We will introduce the use case to you while focusing on the algorithms and technology that we employed!

Leveraging analytics for solving business problems

WHO /

3SC Analytics​

WHEN /

Round #1

13:15 - 14:00

During this presentation,  3SC Analytics, will provide an overview of the different scenarios they encounter as an analytics company through 3 use cases – 4PL services and Control Tower for a leading Swiss pharmaceutical container company, Predictive Analytics for Inventory management for a manufacturer of industrial protective gear and AI/ML driven Demand forecasting for an FMCG company.

 

The presentation will be delivered by Pinak Deshmane Director at 3SC Analytics and Gyanadev Boxi Delivery Head at 3SC Analytics.

A Path to Academia

WHO /

UvA​

WHEN /

Round #1

13:15 - 14:00

Are you thinking about doing a PhD but have no idea where to start looking or what you should be thinking about before applying? Tom van Sonsbeek is here to help you ask the right questions. Tom  is a PhD student at University of Amsterdam at the AI for Medical Imaging (AIM) lab. His project is focussed on multimodel learning from Electronic Health Records (EHR) and medical images. The AIM Lab is a collaborative initiative of the Inception Institute of Artificial Intelligence from the United Arab Emirates and the University of Amsterdam. It focuses on using artificial intelligence for medical image recognition.

Amsterdam Intelligence

WHO /

City of Amsterdam

WHEN /

Round #2

14:15 - 15:00

We live in challenging times, since the covid pandemic is currently hitting the city, and the effects of it on the city are not fully known. At the same time, we need to drastically reduce emissions in the city, and become more sustainable. Also we need Amsterdam to be a fair and just city, and prevent any obstacles from increasing societal differences.

 

At the city of Amsterdam we believe Artificial Intelligence will have a significant impact on the people living in Amsterdam. But how do we encourage positive developments, solve complex challenges and apply the necessary regulation to avoid any negative effects?

 

AI not only stands for Artificial Intelligence but also for ‘Amsterdam Intelligence’, the essence of which is the fair and just use of technology in a manner that reinforces our status as an inclusive, free and creative city.

 

The AI team of the City of Amsterdam is focused on using Artificial intelligence to solve the challenges of the city, and increase the livability for the people of Amsterdam. An example would be using machine learning to route service requests made by citizens more efficiently, computer vision scanning the streets of the city for issues (www.odk.ai) or using computer vision techniques to make pedestrians more aware of keeping sufficient distance.

 

In this presentation Maarten Sukel, AI lead at the City of Amsterdam, will talk about Data Science and AI projects at the city of Amsterdam and his experience of working as a developer for the capitol of the Netherlands.

Introduction into Startups and Venture Capital Funding for Young Founders

WHO /

ASIF Ventures​

WHEN /

Round #2

14:15 - 15:00

In the workshop Lena Hoffmann and Kyra Dresen will tell you all there is to know about startups and venture capital, which is a very specific type of investment for a very unique type of company. In short, when a startup reaches the point where it begins to commercialize its innovation, VCs can offer funding, networks, expertise, and in some cases, validation and credibility for a startup. They will give insight into the social networks, selection criteria, and secrets of VC funding. As the VC industry is dominated by (specifically white) men, special attention will be paid to female and minority founder focussed VCs! By the end of the session, you will understand the basics of VC terminology, know when and how to approach venture capitalists for funding and feel more confident and strategic when asking for financial resources. 

Modelling at ING: Different Perspectives on Artificial Intelligence

WHO /

ING Bank​

WHEN /

Round #2

14:15 - 15:00

ING has set the ambition to become a data-driven organization with analytics as a business model. This requires data and modelling processes, where both intelligence and creativity are critical. With the transition towards ‘future banking’, new types of risk models emerge, which are based on a dynamic sets of (real-time) drivers. Artificial Intelligence and Machine Learning are considered to accelerate the collection of relevant statistics and the development and validation of risk-based models (for example to improve data quality and model performance). ING is committed to advance the way development and validation units create and manage these ‘next generation’ models and the designated technologies.

 

Concrete examples and challenges will be discussed together by connecting to you and sharing our experience.

Few-Shot Object Detection ‘In the Wild’: Detecting Coaxial Cable Splitters, and AOP Wall Socket Plates

WHO /

VodafoneZiggo

WHEN /

Round #3

15:15 - 16:00

In this research project, we apply few-shot object detection ‘in the wild’, transferring a few-shot object detection for meta-learning approach that has been developed, and evaluated on benchmark datasets to a real-life VodafoneZiggo dataset with the aim to significantly reduce the number of annotated instances required to detect coaxial cable splitters, and Abonnee Overname Punten (AOPs). The preliminary results are promising. For example, after some modifications to the source code, the algorithm works properly with our custom dataset. At the end of the project, we will publish a detailed tutorial on how to successfully run the model on your own data.

Execution Management System 

WHO /

Celonis

WHEN /

Round #3

15:15 - 16:00

To achieve the kind of breakthroughs we need in this new era, we need to shift our mindset: from simply optimizing a process to maximizing its capacity to execute. By maximizing execution capacity, business potential will exponentially increase. Learn how to maximize execution capacity and visualize execution gaps by the Celonis Execution Management System (EMS), a new class of technology built to achieve execution capacity breakthroughs. The EMS helps companies manage every facet of execution management from analytics, to strategy and planning, management, actions and automations. At Celonis, we believe that every company can unlock their full execution capacity.

© 2020 by Amsterdam Data Science. QUESTIONS / Contact us at info@amsterdamdatascience.nl

  • Grey LinkedIn Icon
  • Grey Facebook Icon