I am a chief research scientist of the Systems and Machine Learning (SysML) group at NEC Labs Heidelberg. From 2012-2015 I was a postdoctoral research associate at the Allen School of Computer Science, University of Washington, working primarily with Pedro Domingos. I was also a member of the Data and Web Science Research Group at the University of Mannheim.

My research interests include representation learning for graph-structured data, unsupervised and semi-supervised learning, probabilistic graphical models, and statistical relational learning. My group's methods are concerned with learning, inducing, and leveraging relational structure with applications in vision, natural language processing, and the (bio-)medical domain.

I am also co-founder of several open-source digital humanities projects such as the Indiana Philosophy Ontology Project and the Linked Humanities Project.

News

• Two papers accepted to ICML 2019: State-Regularized Recurrent Neural Networks and Learning Discrete Structures for Graph Neural Networks
• I am an invited speaker at the IPAM workshop "Geometric Deep Learning for Big Data and Applications" in LA, the Karlsruhe.AI and Heidelberg.AI speaker series, the RIKEN AI Insitute in Tokyo, and the ECML workshop "New Trends in Representation Learning with Knowledge Graphs"
• One paper accepted to ESWC 2019: MMKG: Multi-Modal Knowledge Graphs.
• One paper accepted to AKBC 2019. Answering Visual-Relational Queries in Web-Extracted Knowledge Graphs.
• We released MMKG a collection of multi-modal knowledge graphs with images and numerical features. The KGs are sub-KGs of DBpedia and Yago whose entities are those found in the commonly used KG completion benchmark FB15k.
• Our work on augmenting hourglass networks with shortcut connections between conv layers of different spatial extent was accepted at MIDL, the Medical Imaging with Deep Learning conference.
• Two papers accepted at EMNLP 2018: Learning Sequence Encoders for Temporal Knowledge Base Completion and
• Our ongoing work on defining a spectrum of graph convolutional networks was accepted to IEEE Data Science Workshop
• Our work on representation learning for knowledge bases with numerical, latent, and relational features accepted at UAI 2018.
• I am an invited speaker for the 27th International Conference on Inductive Logic Programming.
• The paper "Discriminative Gaifman Models" has been accepted at the Thirtieth Annual Conference on Neural Information Processing Systems (NIPS).
• The paper Learning Convolutional Neural Networks for Graphs has been accepted at the International Conference on Machine Learning (ICML).
• (03/27/16) I am co-organizing the IJCAI-16 workshop Statistical Relational AI. Please visit the workshop's website for more information (including some exciting invited talks) and consider submitting your work!
• The paper "Learning and Inference in Tractable Probabilistic Knowledge Bases" was accepted at the 31st Conference on Uncertainty in Artificial Intelligence. (UAI)
• Nominated by the University of Washington, I will present my work at the Information Theory and Applications Workshop (ITA) in San Diego
• The paper "Out of Many, One: Unifying Web-Extracted Knowledge Bases" was accepted at the NIPS Workshop on Automated Knowledge Base Construction (AKBC) 2014
• The paper "Lifted Probabilistic Inference for Asymmetric Graphical Models" was accepted at the 29th AAAI Conference on Artificial Intelligence
• The paper "Generalized Conditional Independence and Decomposition Cognizant Curvature: Implications for Function Optimization" was accepted at the NIPS Workshop on Discrete and Combinatorial Problems in Machine Learning
• I was selected to participate in the 2nd Heidelberg Laureate Forum, to be held September 21-26 in Heidelberg, Germany
• The paper "Tractability through Exchangeability: A New Perspective on Efficient Probabilistic Inference" is one of 5 papers (out of 1400 submissions) nominated for the AAAI outstanding paper award.
• The paper "Exchangeable Variable Models" was accepted at the International Conference on Machine Learning (ICML).
• The paper "Tractability through Exchangeability: A New Perspective on Efficient Probabilistic Inference" was accepted at this year's AAAI conference
• I am serving on the PC of the AAAI workshop StarAI and the SIGMOD/PODS workshop BUDA.
• Pedro Domingos, Daniel Lowd, and I are organizing the ICML workshop Learning Tractable Probabilistic Models. Please visit the workshop's website for more information (including some exciting invited talks) and consider submitting your work!
• I will serve on the PC of AAAI, ECAI, and UAI 2014.