Intelligent Interactive Systems

This research is guided by the vision of closely coupling user interaction with machine learning and reasoning techniques while keeping the user in control. Several current projects aim at improving recommender systems for electronic business and knowledge work. We have developed systems and software frameworks which focus among others on rapidly integrating hybrid recommenders, producing context-aware recommendations for individuals and groups, as well as giving the user more influence on the recommendations process. Besides, adaptivity of UIs is also particularly explored in the area of automotive systems and navigation solutions.

Recommender Systems

Recommender systems have become a widely used and effective means for personalizing information access and presentation in e-commerce and other application areas. Typically, a recommender system attempts to identify a subset of items from a very large information space that meet a user’s interests and preferences best among all alternatives, and subsequently presents those items to the user in a suitable manner. We are particularly interested in user-related aspects of such systems, e.g. increasing interactivity and improving transparency.

Publications

Context-adaptivity

Context-adaptivity can be considered as a set of methods and techniques that is intended to enable users to perform their tasks with less effort or better results by taking the user’s current context into account. Depending on the respective use case, context can be virtually anything that is supposed to have an influence on the task at hand, for instance location, season, temperature, mood, or company of other people.

Publications

Smart Mobility Solutions

Navigation systems are widespread tools in automobiles while mobile applications assist smartphone users in other everyday life situations. Present solutions aim at optimizing routes without taking the user’s knowledge, experience, and preferences into account. Thus, we are investigating techniques for improving presentation of routes by incorporating personal user information and other data. Besides, we are exploring smart solutions to support intermodal mobility, i.e. using different modes of transport.

Publications

Contact

Yuan Ma

Researcher

Lovis Suchmann

Researcher

Jürgen Ziegler

Full Professor

Related publications

A comparative study of item space visualizations for recommender systems

Explaining Recommendations by Means of Aspect-Based Transparent Memories

Interactive Recommending with Tag-Enhanced Matrix Factorization (TagMF)

Impact of Item Consumption on Assessment of Recommendations in User Studies

Sequential User-based Recurrent Neural Network Recommendations

A 3D Item Space Visualization for Presenting and Manipulating User Preferences in Collaborative Filtering

Choice-based Preference Elicitation for Collaborative Filtering Recommender Systems

More related publications »

Related projects

ASSURE

Leverage argumentative statements in user reviews to improve the quality of system-generated recommendations and provide explanations

KOLEGEA++

Adaptive recommendations to support cooperative learning

SoFar

Investigating methods to automatically improve dialogue-based product recommenders

colognE-mobil II

Exploring the multifarious facets of e-mobility, with a focus on smart navigation solutions

CONTici

Context-adaptive interaction in cooperative knowledge processes

WISE

Novel approaches for Web Information and Service Engineering

colognE-mobil

Exploring the multifarious facets of electromobility

Related developments

Games with a Purpose for Understanding Latent Factors

Game that helps understanding latent factor models used in recommender systems

Movie Landscape

A 3D interface that helps expressing personal preferences in large information spaces

TagMF

Increasing transparency and offering richer interaction possibilities in today's recommender systems

Hootle

A hotel recommender system for groups that supports preference negotiation

Blended Recommending

Integrating interactive information filtering and algorithmic recommender techniques

Hybreed Framework

Programming toolkit for complex, context-aware recommenders