Wintersemester 2018/2019, BA/MA Produkt-Design Exzellenzcluster

Tangorithm

- Interaction with learning Algorithms

 

Algorithms are increasingly becoming a part of our life. They recommend products (and friends) to us, offer suitable services to our current situation, and photograph our beloved ones in the right situation. It seems that more and more algorithmic systems are arranging our activities. They help us to filter our everyday workflow and overspill in information and therefore seem to let us experience new leisure spaces. But many of these processes in sorting and structuring data cannot be perceived. Due to the lack of transparency in decision-making processes in algorithmic systems, many people do not actively notice these interventions in their daily life.

 

Emerging debates on „Fairness, Accountability, and Transparency“ discuss these challenges, relating to intelligibility, interpretability and the ability to explain algorithmic systems. To achieve these goals, new approaches in the visualization of algorithmic processes in learning and decision-making are required. Furthermore, for the interaction with algorithmic systems, interfaces have to be rethought. The ways in which machines learn from humans is not one-sided, it is based on “exchange“ and participation. This poses the following questions: Who trains whom? To what extent does the behavior of the machine adapt to the user and vice versa?

 

Students of the Institute for Computer Science (FU-Berlin) and Product Design Department (KH-Berlin Weißensee) have formed interdisciplinary teams and developed together strategies for the day to day handling of algorithmic systems. The aim was to conceptualize new interaction patterns and interfaces on the basis of machine learning technologies and implement these in products (physical and digital). The focus was set particularly on the exchange between the user and the algorithmic system.

 

During the semester the teams have run through the following project stages: the conception stage, the stage of designing and the stage of prototypical implementation. Due to the subject of the course, it was necessary to undertake practical experiences with machine learning right from the start.

 

This course was a cooperation of the research group Human-Centered Computing (FU-Berlin) and the Product Design Department (KH-Berlin Weißensee). Taking place at the HU Clusters of Excellence - Matters of Activity, Winter Term 2018/19. 

Supervision Thomas Ness
Project categorySemester Project Project subjects BA/MA Produkt-Design