|COOKIES: By using this website you agree that we can place Google Analytics Cookies on your device for performance monitoring.|
Large-scale convex optimization for machine learning
If you have a question about this talk, please contact Richard Samworth.
Many machine learning and signal processing problems are traditionally cast as convex optimization problems. A common difficulty in solving these problems is the size of the data, where there are many observations (“large n”) and each of these is large (“large p”). In this setting, online algorithms which pass over the data only once, are usually preferred over batch algorithms, which require multiple passes over the data. In this talk, I will present several recent results, showing that in the ideal infinite-data setting, online learning algorithms based on stochastic approximation should be preferred (both in terms of running speed and generalization performance), but that in the practical finite-data setting, an appropriate combination of batch and online algorithms leads to unexpected behaviors, such as a linear convergence rate with an iteration cost similar to stochastic gradient descent.
This is joint work with Nicolas Le Roux, Eric Moulines and Mark Schmidt.
This talk is part of the Statistics series.
This talk is included in these lists:
Note that ex-directory lists are not shown.
Other listsRede Lectures Ver Heyden de Lancey Lectures Madingley Lunchtime Seminars
Other talksBreaking Card: Reverse-Engineering the Smart-Card Application Protocol Data Unit The Liverpool Epidemic Strain of Pseudomonas aeruginosa: Transmission and Transatlantic Travels Personal Principles and the Political Game Land and credit in the asset allocations of the Jews in late 14th-century Zaragoza Sensor CDT - Sensors Day Dale’s principle and the neural processing of latent factors