Hello! An experimental version of my new portfolio is available at beta.gerhardbliedung.com.

ntt


client

 Bachelor thesis / ZHdK

Year

 2014

Documentation

 Link

Thesis

 Link

Embracing reciprocity in ubiquitous computing.

Motivation

While technology advances rapidly and simultaneously enters personal space, our relationship with these artifacts seems to evolve at a different pace. The categorization as an obedient tool becomes debatable, when for example self driving cars have to decide over the fate of their owner, based on factors that are hidden and subject to constant change. Ntt aims to provide an outlook that encourages mutual understanding and responsibility.

Product



The voice controlled frontend consists of a three-dimensional space, containing a circle within a network of spheres. As the circle has no method to articulate itself without negotiation with its environment, it uses the spheres as interfaces for output. These could also be different devices or services, making the system device agnostic. The backend is mostly web based and can run on many devices equipped with audio and video input.

Interaction

The visitor is confronted with a constantly changing state, influenced by text mining, facial expressions as well as speech recognition. Depending on the current mood of the circle and the behavior of previous visitors towards it, the system may even develop a complete refusal until a certain amount of confidence is restored. What can be achieved, therefore depends on the amount of investment in the intersubjective relation.

Added value and future potential

The reconsideration of our relationship towards complex artifacts, especially those dealing with artificial intelligence, as well as evaluating an adequate representation beyond the problematic usage of anthropomorphism poses an interesting field of research. While this installation is a brief glimpse of this area, the foundation could be used for further interface experiments as the system can adapt new functions after a short training period.

Technologies used

  • Node.js
  • Web GL
  • Speech recognition
  • Face & expression tracking
  • Sentiment analysis
  • Supervised learning algorithms