Edwin Simpson, DPhil
work +49 6151 16-21677
fax +49 6151 16-25295
Office: S2|02 B106
I am interested in the task of learning from weak information, how this applies to interactive learning environments, and adapting Bayesian techniques to natural language processing. Weak information for learning a model is often available when large amounts of reliable training data is not. It consists of noisy signals from different sources -- such as crowdsourced annotations, implicit feedback from users of an application, or untrusted reports of an event on social media. By combining different sources of weak information and modelling their reliability, we can learn reliable models when gold-labelled training data is not sufficient. NLP tasks such as argument mining, in which we extract and collate arguments from text corpora, are a good target case. There is a large variation in the types of arguments we may wish to look for in different domains and applications, and we need to gather data efficiently to adapt to these new situations. I'm interested in how we can address such problems by transferring knowledge between domains and reducing the cost of annotating training data through intelligent crowdsourcing and interactive learning with humans in the loop.
One tool I have developed for combining weak information sources is github.com/edwinrobots/pyIBCC, which is designed for aggregating crowdsourced annotations and was shown to outperform rival methods. I'm currently working on making the code more user-friendly and incorporating new models -- please get in touch if you would like some help using this method.
Some keywords for topics I am interested in include: natural language processing, Bayesian inference, approximate inference, scalable inference, interactive learning, crowdsourcing, computational argumentation.
I joined the UKP lab in April 2016 as a postdoctoral researcher. I was previously a member of the Machine Learning Research Group at the University of Oxford from 2010 to 2016, where I completed my doctorate and first postdoc on decision making with crowds of unreliable classifiers. Before that, I worked as a research engineer at Hewlett Packard Labs in Bristol, UK, and have a Masters in Computer Science from the University of Bristol.
Ramchurn, S. D., Huynh, T. D., Wu, F., Ikuno, Y., Flann, J., Moreau, L., … & Reece, S. (2016). A disaster response system based on human-agent collectives. Journal of Artificial Intelligence Research, 57, 661-708.
Simpson, E. D., Venanzi, M., Reece, S., Kohli, P., Guiver, J., Roberts, S. J., & Jennings, N. R. (2015, May). Language understanding in the wild: Combining crowdsourcing and machine learning. In Proceedings of the 24th international conference on world wide web (pp. 992-1002). International World Wide Web Conferences Steering Committee.
2014, D.Phil Thesis: Combined Decision Making with Multiple Agents. University of Oxford.
S. Ramchurn, T.D. Huynh, Y. Ikuno, J. Flann, F. Wu, L. Moreau, N. Jennings, J. Fischer, W. Jiang, T. Rodden, E. Simpson, S. Reece and S. Roberts (2015). HAC-ER: A Disaster Response System based on Human-Agent Collectives, Proceedings of the 2015 International Conference on Autonomous Agents and Multiagent Systems (AAMAS). Best paper of the innovative applications track.
E. Simpson, S. Roberts (2015). Bayesian Methods for Intelligent Task Assignment in Crowdsourcing Systems, Scalable Decision Making: Uncertainty, Imperfection, Deliberation and Scalability, Studies in Computational Intelligence, Springer, p. 1-32.
A. Levenberg, S. Pulman, K. Moilanen, E. Simpson, S. Roberts (2014) -- Best Paper Award. Predicting Economic Indicators from Web Text Using Sentiment Composition, International Journal of Computer and Communication Engineering, IACSIT Press
A. Levenberg, E. Simpson, S. Roberts, G. Gottlob (2013). Economic Prediction using Heterogeneous Data Streams from the World Wide Web, Scalable Decision Making: Uncertainty, Imperfection, Deliberation (SCALE), Proceedings of ECML/PKDD 2013 Workshop, Springer
E. Simpson, S. Roberts, I. Psorakis and A. Smith (2013). Dynamic Bayesian Combination of Multiple Imperfect Classifiers, Decision Making and Imperfection, Intelligent Systems Reference Library series, Springer
E. Simpson, S. Reece, A. Penta, G. Ramchurn (2013). Using a Bayesian Model to Combine LDA Features with Crowdsourced Responses, Proceedings of the 21st Text Retrieval Conference, NIST
E. Simpson, S. Roberts, A. Smith (2012) -- Best contribution award. Dynamic Bayesian Combination of Multiple Imperfect Classifiers, Presented at the NIPS 2012 workshop on Human Computation for Science and Computational Sustainability
E. Simpson, S. Reece, S. Roberts, G. Ramchurn (2012). An Information Theoretic Approach to Managing Multiple Decision Makers, Presented at the NIPS 2012 workshop on Human Computation for Science and Computational Sustainability
E. Simpson, S. Roberts, I. Psorakis, A. Smith and C. Lintott (2011). Bayesian Combination of Multiple, Imperfect Classifiers, Proceedings of NIPS 2011 workshop on Decision Making with Imperfect Classifiers
Anzahl der Einträge: 4.
Eger, Steffen ; Şahin, Gözde Gül ; Rücklé, Andreas ; Lee, Ji-Ung ; Schulz, Claudia ; Mesgar, Mohsen ; Swarnkar, Krishnkant ; Simpson, Edwin ; Gurevych, Iryna
Gurevych, Iryna ; Meyer, Christian M. ; Binnig, Carsten ; Fürnkranz, Johannes ; Kersting, Kristian ; Roth, Stefan ; Simpson, Edwin
Simpson, Edwin ; Gurevych, Iryna
Simpson, Edwin ; Reece, Steven ; Roberts, Stephen J.