See also my CV for more information on myself.
My work concerns
Intelligence is the ability to act under uncertainty. It exists on a broad range, not just of physical, but also of computational scales: From simplistic ideas like gradient descent, which may be a microbe's strategy to get closer to a source of nutrients, to an adult human's reasoning about career goals. Much of modern research in machine learning and artificial intelligence aims for the top of this hierarchy: algorithms capable of building highly structured models, and taking complicated decisions, at high computational cost. I believe that there is still plenty of room for improvement left at the bottom, too.
Algorithms for the bottom end of the intelligence hierarchy are those constructed by numerical mathematics. They are methods that take as input a function and return elementary properties of that function that are not tractable from the analytic form alone: Optimizers return the location of (local or global) extrema. Quadrature methods return the values of integrals. Sampling methods interpret the function as an unnormalised probability distribution to draw random numbers from. Differential equation solvers And control algorithms treat the function as describing a dynamical system to simulate. It is not a new, but still a little-known idea that all these methods can be seen as performing inference: Making statements about an uncertain quantities given certain observations of related quantities.
These algorithms are the building blocks for the more complex, expensive, fancy top level intelligence. So they have to be modular, to be re-usable. They have to be robust, because their failure may cause big problems upstream. And of course they have to be cheap. In my work, I try to address theses requirements. Here is a selection of some of it. See "publications" for pdfs and detailed citations, and my CV for more information.:
Stochastic gradient descent is still the dominant algorithm for the training of many online learning algorithms, like neural networks. All just because more elaborate ideas, like quasi-Newton methods, cannot deal with noise? See what can be done about that: Hennig. "Fast Probabilistic Optimization from Noisy Gradients". ICML 2013
Did you know that BFGS is a least-squares regressor? See what happens when you make it nonparametric: Hennig & Kiefel. "Quasi-Newton methods, a new direction". ICML 2012
When optimizing experimental parameters in search of a global optimum, algorithms shouldn't try evaluating close to the optimum. They should try to evaluate where they expect to learn most about the optimum. Hennig & Schuler, "Entropy Search for Information Efficient Global Optimization". JMLR 13 (2012).
Probability theory offers a uniquely coherent view on the infamous exploration/exploitation tradeoff: From the Bayesian view, reinforcement learning is about modelling the effect of possible future observations on the optimality of decisions taken in the present. In general, this decision process is intractable. But under Gaussian process assumptions (which, depending how on look on it, is either a quite general, or a quite limited set of assumptions), the right answer moves within reach of numerical analysis. Hennig, "Optimal Reinforcement Learning for Gaussian Systems", NIPS 2011
Topic modelling is a very popular area of machine learning at the moment. Documents come with metadata, and topics change over time, and from document to document depending on the author, the subject, and many other features. The probabilistic extension of topic models that allows modelling such effects requires an algorithmic link between discrete distributions and continuous domains, often realised as a set of "dependent Dirichlets". We pointed out how to do this, in a numerically extremely efficient way. Hennig, Stern, Herbrich and Graepel, "Kernel Topic Models". AISTATS 2011.
Tree search, finding the optimal leaf of a tree, is exponentially hard in the depth of the tree, because trees are exponentially big in their depth. But what happens during that exponentially long search? If you have a probabilistic belief over the value and location of the optimal leaf, and get one more observation of one individual leaf's values? Shouldn't updating the belief cost only linear time? It does. Hennig, Stern and Grapel. "Coherent Inference on Optimal Play in Game Trees". AISTATS 2010
and (2013) Quasi-Newton Methods: A New Direction
Journal of Machine Learning Research 14 807-829.|
, and (2013) Analytical probabilistic modeling for radiation therapy treatment planning
Physics in Medicine and Biology . in press|
and (2012) Entropy Search for Information-Efficient Global Optimization
Journal of Machine Learning Research 13 1809-1837.|
and (2007) Point-spread functions for backscattered imaging in the scanning electron microscope
Journal of Applied Physics 102(12) 1-8.|
Conference papers (8):
and (2012) Quasi-Newton Methods: A New Direction
29th International Conference on Machine Learning (ICML 2012), 1-8.
, , and (2012) Learning Tracking Control with Forward Models
IEEE International Conference on Robotics and Automation (ICRA 2012), 259 -264.
, , and (2012) Kernel Topic Models
Fifteenth International Conference on Artificial Intelligence and Statistics (AI & Statistics 2012), 1-9.
, and (2012) Approximate Gaussian Integration using Expectation Propagation
(2012) Fast Probabilistic Optimization from Noisy Gradients
30th International Conference on Machine Learning (ICML).
(2011) Optimal Reinforcement Learning for Gaussian Systems
In: Advances in Neural Information Processing Systems 24, (Ed) , , , and , Twenty-Fifth Annual Conference on Neural Information Processing Systems (NIPS 2011), 325-333.
, and (2010) Using an Infinite Von Mises-Fisher Mixture Model to Cluster Treatment Beam Directions in External Radiation Therapy
(Ed) Draghici, S. , T.M. Khoshgoftaar, V. Palade, W. Pedrycz, M.A. Wani, X. Zhu, Ninth International Conference on Machine Learning and Applications (ICMLA 2010), IEEE, Piscataway, NJ, USA, 746-751.
, and (2010) Coherent Inference on Optimal Play in Game Trees
In: JMLR Workshop and Conference Proceedings Volume 9: AISTATS 2010, (Ed) Teh, Y.W. , M. Titterington, Thirteenth International Conference on Artificial Intelligence and Statistics, JMLR, Cambridge, MA, USA, 326-333.
Technical reports (1):
: Expectation Propagation on the Maximum of Correlated Normal Variables, Cavendish Laboratory: University of Cambridge, (2009).
, and (2009): Bayesian Quadratic Reinforcement Learning, NIPS 2009 Workshop on Probabilistic Approaches for Robotics and Control, Whistler, BC, Canada.|
: Approximate Inference in Graphical Models, University of Cambridge, (2010).