Wilker Aziz, University of Amsterdam
Date & time: From 27-29 May, 9:00h – 12:30h (each session)
Location: Room 11.26, 11th floor, North Tower, IST
Neural networks are taking NLP by storm. Yet they are mostly applied in contexts were we have complete supervision. Many real-world NLP problems require unsupervised or semi-supervised models, however, because annotated data is hard to obtain. This is where generative models shine. Through the use of latent variables they can be applied in missing data settings. Furthermore they can complete missing entries in partially annotated data sets.
This tutorial is about how to use neural networks inside generative models, thus giving us Deep Generative Models (DGMs). The training method of choice for these models is variational inference (VI). We start out by introducing VI on a basic level. From there we turn to DGMs that employ discrete and/or continuous latent variables. We justify them theoretically and give concrete advise on how to implement them. For continuous latent variables, we review the variational autoencoder and use Gaussian reparametrisation to show how to sample latent values from it. For discrete latent variables, for which no reparametrisation exists, we explain how to use the score-function or REINFORCE gradient estimator. More Information..
Sir Peter Knight, Quantum Metrology Institute, UK National Physical Laboratory
& Blackett Laboratory, Imperial College
Date & time: Tuesday, May 21st, 16:00h
Location: VA6, Civil Engineering Building, IST, Lisbon
Quantum Physics has focused on light, matter and their interaction, from the early days of quantum mechanics right down to the present day. Much of this work has concentrated on the nature of quantum correlations and coherence beyond what is allowed classically. Quantum Information Science in part has grown out of this progress: the quantum world allows information to be encoded, manipulated and transmitted in ways quite different from classical physics and quantum approaches to measurement to transform metrology. Now we are seeing the transition to technology, applying these insights in spectacular new ways.
Parallelism and entanglement, the characteristic features of the quantum world, enable us to perform precise measurements and to undertake information processing tasks which are peculiar to the quantum world: secure encryption, teleportation of quantum states and the speed up of certain classes of algorithms. The 21st Century has seen the emergence of a networked world, connected by global fibre-optic communications and mobile phones, with geo-location provided through GPS, and all this has changed our lives more dramatically than at any time since the industrial revolution. Quantum-enabled technology is ate the heart of this change. I will describe these developments and how the UK has invested substantially in the basic science and its exploitation. Some of these include communication systems immune to GPS jamming (of real importance for global security), as well as quantum sensors for medical applications (including cardiography, neurophysiology, etc.),
sensitive magnetometry, gyros, and geophysical surveying.
Sir Peter Knight is one of the pioneers of quantum information science. A former president of the UK Institute of Physics and of the Optical Society of America, he has been the driving force behind the UK National Quantum Technologies Programme, an initiative that led to the creation of similar programmes in other countries around the world, as well as at EU level.
For more information about Sir Peter Knight, see: https://www.imperial.ac.uk/people/p.knight