To kick off the new Data Science & Artificial Intelligence course of study, the University of Applied Sciences Wedel is devoting a series of lectures to the topic of the application of artificial intelligence. For the lecture series "How companies benefit from AI", well-known companies are invited to give insight into their practical application examples. The aim of the series of events is to provide an overview of the multifaceted areas of application of the processes using practical examples from a wide range of industries. In addition to the students of FH Wedel, interested parties from outside the university are also welcome to be inspired by the application examples for possible own plans or projects.
From 28 October, the lecture series will take place digitally via zoom every Wednesday at 5 pm. The 45-minute lectures in English are followed by an open discussion with questions and answers. It is not necessary to register for the event, but it is recommended to register with Eventbrite under the event link to be informed of any updates. Any organizational questions should be directed to the Public Relations Department at email@example.com or by phone at 04103-8048-50. Questions regarding the content of the event will be answered by the head of the Data Science & Artificial Intelligence program, Dr. Hendrik Annuth, by E-Mail.
From the idea to the release - how does the AI get into the program
Basic research and product development are two extreme points in the spectrum of modern software development. So how do results from research find their way into the finished product? How does collaboration between researchers and developers work, and how can the balancing act between the researchers' desire for publications and the obligation of industrial secrecy be successfully resolved? Dr. Nico Becherer, who has been working with Adobe Research scientists for many years and develops features in products such as Photoshop, Lightroom, Audition and Premiere, will talk about interpreting between two worlds in his presentation.
Dr. Nico Becherer is a Senior Engineering Manager for Applied Research in the Adobe Video & Audio Team. Together with his team he works on evaluating state of the art research and designing code and algorithms for Audition, Premiere Pro, After Effects and other Adobe products. His current focus lies in workflow simplification - making video and audio processing more accessible to non-expert users through the use of advanced technologies and machine learning. Nico gained a PhD in Computer Science in 2008 and has worked with several major audio software companies in the past 10 years. He holds several patents in audio and video processing. When he’s not working on sound and music he’s building robotic spiders or is teaching university students on how to play improv theater.
Automating Machine Learning for Recommender Systems
At XING, recommender systems are build as ensembles that consist of different recommender strategies, filters, diversification components and ranking models. Various machine learning models are applied to better estimate interests of the users, learn thresholds for filtering or train models that rank the top k items that are presented to the users. Since the XING platform (including the recommender services themselves), user behavior and therefore the data is constantly changing, it is key to continuously adapt the involved machine learning models. In this talk, the speakers will give an overview about recommender systems at XING, give insights in selected algorithms and discuss the approaches towards automating involved machine learning processes.
Fabian builds search and recommender systems at XING. Before joining XING in 2012, he was working on topics related to user modeling and personalization in academia at TU Delft, the Netherlands. Nils worked for a long time as a Software Developer before he switched to machine learning and worked as a research assistant in the field of autonomous robotics and image recognition at the University of Hamburg. In 2018, he started at XING as a Data Scientist building recommender.
The impact of radical process automation and (nearly) autonomous organizations
Artificial intelligence has been around for more than 60 years. In various AI summers and winters, the topic was seen to come and go. But this time, it is likely that it has come to stay. Artificial intelligence and intelligent automation in the business world is only starting to scratch underneath the surface. This talk is taking a deeper look into the technological implications of any AI application as well as on the underlying technical, societal, and economic implications. The automation of knowledge work is likely to have a similar effect on human work as industrial automation, electrification, and the internet had. This implies a change in our need for know-how, talent, mindset, and ultimately our traditional career paths.
Tim Behrens is a Senior Consultant with Connected Innovations, a Digital and AI Consultancy. Tim works closely with leading thinkers and entrepreneurs in the field of artificial intelligence, he supports innovation projects at leading companies in all aspects of how AI technology can be used to develop new products, services, and business models. Before Connected Innovations, Tim has been a consultant and later Chief of Staff with Arago, a European AI platform company specializing in machine reasoning and AI-based process automation. Tim holds a degree in Economics, he has studied in Karlsruhe, at ESB Business School in Reutlingen, and in Querétaro Mexico.
Creating business insights through text classification
At Qualitize, thousands of comments are collected from users every week. These comments can be aggregated in smart ways to quickly gain relevant business insights, e.g. by classifying them by sentiment (positive or negative) or topic. At scale, this requires the support from machine learning to automatically classify comments based on previously labeled training data. To support changing requirements for business insights resulting in new classification tasks, the classification software has to be flexible and adaptive. This talk covers the basics of active learning and text classification, as well as how Qualitize developed an open source labelling tool that leverages machine learning and active learning to train new and update existing machine learning models. Finally, the speakers present real-world examples of business insights that can be gained from classifying and aggregating textual feedback for our customers.
Lukas Skibowski, Chief Information Officer (CIO) of Qualitize, holds a Bachelor (B.Sc.) degree in Computer Science Engineering from the Technical University of Hamburg and subsequently completed a Master (M.Sc.) in IT Management & Consulting at the University of Hamburg. At Qualitize, he is responsible for the strategic orientation of the IT and is the interface between the product requirements and the corresponding IT implementation.
Max Wiechmann, Software Engineer at Qualitize, holds a B.Sc. in Media Systems from HAW Hamburg and an M.Sc. in Computer Science from the University of Hamburg. In addition to the areas of software engineering and architecture, his focus is on the application of machine learning in the context of natural language processing (NLP). At Qualitize, he is responsible for the design and implementation of backend systems and the implementation of machine learning solutions.
Data integration using semantic graphs
Data integration is a key challenge in many enterprises since existing IT and thus data architectures are often organized in silos mirroring the organizational structure. Every cross-silo activity such as process automation or business analytics therefore needs data integration. Traditional approaches often fail when functional data definitions are ambiguous and data can only be understood in its context. But there is a solution: ontologies and semantic graphs – originally coming from linguistics and sociology in the 1960s – are the only known structures that allow context-bound data pools. It is thus not surprising, that semantic graphs have been (re-)established in digital research since the 1990s to link data and their relationships with each other. Today, tech platforms like Google use sematic graphs to organize huge amounts of data ensuring high performance and flexibility. This talk invites to explore what a semantic graph is and how it can be deployed in various business contexts.
Dr. Edeltraud Leibrock is Co-Founder, Partner, and Managing Director with Connected Innovations, a Digital and AI Consultancy. She also engages as Supervisory and Advisory Board Member of several Fintech companies and AI Initiatives such as ARIC (Artificial Intelligence Center Hamburg). Previously, she served as Executive Board Member responsible for IT and processes for KfW Bankengruppe, the world's largest national development bank. Edeltraud studied Physics and Biology in Regensburg, received her PhD from Technical University Hamburg, and worked as Postdoc at NOAA, Boulder, Colorado. She strongly believes in the innovative power of interdisciplinary exchange and cooperation.
AI in the Enterprise Software Environment
With its unique business model, Tchibo combines a supply chain, two completely different product divisions and three distribution channels – an exciting environment for data-based optimisation. In his presentation, Omar Hairani will explain the technical and organisational approaches Tchibo is taking to this challenge and how artificial intelligence in the enterprise software environment has increasingly become an essential part of Tchibo's IT strategy.
Omar Hairani is Head of BI & Big Data Application at Tchibo. He and his team are responsible for the development of analytics & AI applications in the SAP BI environment for all core systems as well as for the open source area on the Google Cloud platform.