EPSRC Centre for Doctoral Training in
Intelligent Games and Game Intelligence (IGGI)
Hi, I’m Christian, a PhD student in the Computational Creativity Group. I’m also a member of IGGI, the EPSRC Centre for Doctoral Training in Intelligent Games and Game Intelligence.
I am interested in how we can create artificial systems that would be deemed creative in their own right by unbiased observers. I address this challenge with formal models of intrinsic motivation. I show both theoretically and in applied studies that models of intrinsic motivation can give rise to more general, robust and adaptive creative systems. Video games make an ideal application domain for this work, as they represent arbitrarily complex, reliable abstractions of the real world, and a challenging, multi-faceted domain for designers.
I have a Magister Artium in Computer Science, History of Art and Business Studies and a BSc in Computer Science. For my theses, I worked on intrinsic motivation in multi-agent systems, and on algorithms for the serendipitous recommendation of art. During my studies, I gained industrial experience at the research department of SAP SE and worked as a research assistant at Technical University Darmstadt. I’m an open minded, passionate researcher and draw on an interdisciplinary approach. My general interests are in Computational Creativity and AI, Machine Learning and Information Theory.
My core research focusses on three different challenges. I employ formal models of intrinsic motivation to:
- Realise intentional agency in General AI and Computational Creativity.
- Create General Non-Player Characters which exhibit creative, companion-/enemy-like behaviour in an arbitrary video game.
- Predict a player’s experience of video game content without involving a human player or designer.
Together with Christoph Salge and Tobias Mahlmann, I co-presented the first tutorial on Intrinsic Motivation in General Game Playing and NPCs at the Computational Intelligence and Games conference (CIG’16, Link).
I also contributed to STATSREP-ML, an open-source tool for automating the process of evaluating machine-learning results. You can find the code and instructions here.
You can contact me on .[at].., or find me on Twitter @CreativeEndvs.
In: Semantic Web, 2015.
In: Entropy, 16 (6), pp. 3357–3378, 2014, ISSN: 1099-4300.
In: Proc. 8th Int. Conf. Computational Creativity, 2017 , 2017.
In: Proc. 35st ACM Conf. Human Factors in Computing Systems (CHI), ACM 2017.
In: Proc. IEEE Conf. Computational Intelligence in Games (CIG’16), IEEE, 2016.
In: Proc. 15th Int. Conf. Synthesis and Simulation of Living Systems (ALIFE), 2016.
In: Proc. 7th Int. Conf. Computational Creativity, 2016.
In: Proc. 7th Int. Conf. Computational Creativity, 2016.
In: Proc. Conf. Empirical Methods in Natural Language Processing (EMNLP), 2015.
Computational Poetry Workshop: Making Sense of Work in Progress (Inproceedings)
In: Proc. 6th Int. Conf. Computational Creativity, 2015.
Towards the Automatic Generation of Fictional Ideas for Games (Inproceedings)
In: Experimental AI in Games (EXAG’14), a workshop collocated with the tenth annual AAAI conference on artificial intelligence and interactive digital entertainment (AIIDE’14). AAAI Publications, 2014.
US 20160132954 A1, 2016.
Patent: Reporting and Managing Incidents (Miscellaneous)
US 8786433 B2, 2014.
Telecooperation Group, Technical University Darmstadt 2014.
I have been invited to give the following tasks. Please click on the accordion tab for details.
University of Hertfordshire, 2016
As part of the RAGS (Research in Adaptive systems Group Seminar) series
Collaboration in Co-Creative Scenarios via Coupled Empowerment Maximization: A Case-Study in Video Games
Recently, embodied and situated agents have become increasingly popular in co-creative systems (where humans and artificial agents jointly work on creative tasks). Intrinsically-motivated agents are particularly successful here, because of their capacity to act flexibly and adapt in open-ended interactions without clearly specified goals. Unfortunately, existing implementations do not manage to establish and maintain collaboration as a core mechanic in such systems without constraining the flexibility of the agents by means of explicitly specified interaction rules. This talk introduces the information-theoretic principle of coupled empowerment maximization as a means to establish a frame for both collaborative and antagonistic behaviour within which agents can interact with maximum flexibility. We study this mechanism in a dungeon-crawler video game testbed, to drive the behavior of an NPC supporting the human player. We demonstrate our progress, future challenges, and argue that the principle could eventually allow for the emergence of truly creative behavior.
Tungsten Centre for Intelligent Data Analytics, 2016
Does Empowerment Allow for Fully Enactive Artificial Agents?
The enactive AI framework wants to overcome the sense making limitations of embodied AI by drawing on the biosystemic foundations of enactive cognitive science. While embodied AI tries to ground meaning in sensorimotor interaction, enactive AI adds further requirements by grounding sensorimotor interaction in autonomous agency. At the core of this shift is the requirement for a truly intrinsic value function. We suggest that empowerment, an information-theoretic quantity based on an agent’s embodiment, represents such a function. We highlight the role of empowerment maximization in satisfying the requirements of enactive AI, i.e. establishing constitutive autonomy and adaptivity, in detail. We then argue that empowerment, grounded in a precarious existence, allows an agent to enact a world based on the relevance of environmental features in respect to its own identity.
My talk on “Addressing the “Why?” in Computational Creativity: A Non-Anthropocentric, Minimal Model of Intentional Creative Agency” at the 8th International Conference on Computational Creativity in Atlanta, Georgia, June 19 – June 23, 2017 (paper link).
This video showcases our research on intrinsically motivated, general companion NPCs (paper link).