Use AI to kickstart your digital journey
Artificial Intelligence is becoming integral to digitalisation – a fundamental component of the journey that allows us to reduce repetitive, time-consuming and redundant tasks, resulting in creating capacity for more lucrative added-value activities. The journey that can enable us to become more connected on a human level again, and to focus on our human values that no AI is going to be able to replace.
The first CognitiveTank of 2019 puts the spotlight on use-cases varying between industries to the Moon and back. Literally. We will bring close to ten hands-on examples on stage, demonstrating how AI can underpin digital transformation; combined with a panel discussion focusing on four different fields where AI has been making its impact already.
Date & Time
14th January 2019, @13:30
Host & Location
T-Systems Schweiz, Kloten
Presentations are held in German and English
All slides are in English
Event participation is by invitation only.
Welcome by Host – Good and Bad AI
When discussing AI initiatives, both public and experts are encountering its key dilemma very early in the process: Whether it is about sorting out job applicants or deciding on what the least damage in an unresolvable traffic situation is – we expect AI to learn, to advise or even decide autonomously in a “right” and still bias-free way. This opening note will demonstrate that a “correct” decision without bias is an impossibility, and I’ll propose a framework on how to judge if AI solutions are ethically good or bad.
Dr. Bert Klöppel
Real world examples of computer vision
Analysis of pictures and video streams are still a reasonable challenge in AI because of the proximity to practical usage and because of the easy performance benchmarking with a very elaborated and praxis proven real-world vision system: us! Some practical examples demonstrate the power and usefulness of solutions based on even mid-range implementations: so computer vision becomes common practice even in everyday situations.
Dr. Daniel Angerhausen
NASA Frontier Development Lab / Universität Bern / Explainables Science Communication
Bringing Space Exploration AI applications back to Earth
Dr. Daniel Angerhausen will present AI applications in Space Exploration developed at NASA’s frontier development lab (FDL) and put them in context with use cases down on Earth. FDL is an applied AI accelerator that maximizes new AI technologies and in academia and the private sector and apply them to challenges in the space sciences. Commercial, international and academic partners, such as Nvidia, Intel, and Google provide capital, expertise and vast compute resources necessary for rapid experimentation and iteration in data intensive areas. The FDL projects range from predicting solar storms and navigating rovers on the moon to “digitally” repairing failed satellite instruments and the search for life in space. Comparable problem sets on Earth can come from various fields such as Fintech, logistics or customer analysis.
Dr. med. Drazen Jurjevic
Landmark detection in medical images: Early-stage glaucoma diagnosis with AI
Medical image analysis often depends on the accurate detection of landmarks as anchor points for subsequent diagnoses. We will present CellmatiQ’s AI platform strategy and technology for pattern detection on the example of early-phase glaucoma identification in ophthalmology, which has the potential to surpass human experts.
Using AI for unstructured client data in the private banking industry
Our bank relationship managers are using valuable information out of different documents for their meeting preparation. Such manual activities are repetitive, costly and take a lot of time to complete. This keynote will reveal a practical solution to this problem that potentially could also be leveraged for other industries too. Milestones of the project will be presented and future possible with business value will be investigated.
Improve your Monday morning coffee experience with AI
See how you can enhance the experience of your customer’s everyday life through AI. During this session, you will get an insight into what it means to train an AI and how to avoid common pitfalls. See how you can put the result into your customer’s hands without any privacy concerns. The use case is to configure a coffee machine for a given coffee bean via Smartphone. The trained AI (DL), embedded in an iOS and Android App, can detect the coffee based on the packaging and then configure the coffee machine accordingly.
Break-Out Session Leaders
Responsible AI – Creating fair and understandable AI applications.
In this workshop, we will look at the current playing field of AI that is being built in various industries, and discuss the facets of their inherent risk. Alleviating and/or minimizing these risks calls for new global platforms to be developed which entail a new field of research called Responsible AI (RAI). We will brainstorm about how RAI can be developed and implemented in YOUR industry & organisation, and we will also put the spotlight on one area of RAI called model interpretability. If time permits, an example in Python will be also given and discussed.
Book Author (Creating value with AI) // Google Developers // Efma
Implementing AI in your product for end-user adoption.
In this interactive workshop we will brainstorm about how we can identify the “best problems” for using AI/ML and what the conditions need to be satisfied for being able to select the right problems. We will also explore how to overcome challenges concerning with data, product development, and user adoption. The workshop leader will also share lessons learned from over 10 years of his experience with building AI products in research, startups and corporations, and explain the five-step process for building AI products. The workshop is suitable for people who are interested to use AI/ML for solving problems and building products in their departments. No technical knowledge or programming skill is required.
Cross-company ML: needs and challenges illustrated with Swiss German Voicebots.
ML Training often requires data sources from different organizations and poses a challenge in terms of protecting personal data, especially language resources. The Swiss-German corpora initiative for voice bots therefore follows a decentralized training approach. In this workshop, we will work in smaller groups, and then, together, we will jointly examine whether this method is suitable and can be used for other purposes.
Dr. Bert Klöppel
Experiences and Challenges of AI.
AI-based innovative solutions is and will be a rapidly growing field. Easy available AI frameworks, the abundance of data and scalable computing powers boost
this development. However: designing useful and reliable solutions based on well-defined requirements and combining all needed IT capabilities is essential for the practical success covering technical and business aspects as well as legal and ethical constraints. Based on some real-world examples, the workshop discussion
will face those demands with practical experiences.