Zoltán Szabó: recent talks.

Invited Talk (Global):
Computer Science Colloquium, Department of Computer Science, University of Warwick, UK. The Power of Cumulants in Reproducing Kernel Hilbert Spaces. [abstract, slides; Feb. 12, 2024]
CMStatistics-2023: Economic Data Analysis and Statistical Inference to Unfold Uncertainty session, Berlin, Germany. Nyström M-HSIC. [abstract, slides; Dec. 18, 2023]
CMStatistics-2023: Statistical Machine Learning with Kernels and Non-linear Transformations session, Berlin, Germany. Beyond Mean Embedding: Cumulants in RKHSs. [slides; Dec. 17, 2023]
CMStatistics-2023: Advances in Kernel Methods and Gaussian Processes session, Berlin, Germany. Kernel Cumulants. [abstract, slides; Dec. 16, 2023]
Lifting Inference with Kernel Embeddings (LIKE23), Bern, Switzerland. Kernel Cumulant Embedding. [slides; June 26-30, 2023]
Plenary talk @ Advanced Course on Data Science and Machine Learning (ACDL-2023), Tuscany, Italy. Shape-Constrained Kernel Machines and Their Applications. [abstract, slides; June 10-14, 2023]
Plenary talk @ Advanced Course on Data Science and Machine Learning (ACDL-2023), Tuscany, Italy. Beyond Mean Embedding: The Power of Cumulants in RKHSs. [abstract, slides; June 10-14, 2023]
BIRS workshop on New Interfaces of Stochastic Analysis and Rough Paths. Kernel Machines with Shape Constraints. [abstract, slides; Sept. 8, 2022]
International Conference on Econometrics and Statistics (EcoSta): Recent Advances in Machine Learning session. When Shape Constraints Meet Kernel Machines. [abstract, slides; June 4, 2022]
DataSig Seminar, Mathematical Institute, University of Oxford. Tensor Product Kernels for Independence. [abstract, slides; May 26, 2022]
Gatsby Unit, London, UK. When Kernel Machines Meet Shape Constraints. [abstract, slides; Jan. 26, 2022]
Lifting Inference with Kernel Embeddings (LIKE22) winter school and workshop. Continuous Emotion Transfer using RKHSs. [slides; Jan. 12, 2022]
CMStatistics-2021: Advanced Statistical Methods for High Dimensional Data session. Vector-Valued Infinite Task Learning in Style Transfer. [abstract, slides; Dec. 19, 2021]
Data Science Seminar at Eurecom. Shape Constraints Meet Kernel Machines. [abstract, slides; Nov. 4, 2021]
EUROPT-2021. Kernel Regression with Hard Shape Constraints. [slides; July 8, 2021].
LIS, Aix-Marseille University. Vector-valued Prediction with RKHSs and Hard Shape Constraints. [abstract, slides; May 20, 2021].
Department of Statistics, LSE. Information Theory, Kernels and Applications. [March 4, 2021].
Texas A&M University: Department of Statistics. Shape-Constrained Kernel Machines. [abstract, slides; Feb. 12, 2021]
Meeting on Mathematical Statistics (MMS): Robustness and Computational Efficiency of Algorithms in Statistical Learning. Kernel Machines with Hard Shape Constraints. [abstract, slides; Dec. 15, 2020]
D. E. Shaw Group, New York, US. Kernel Information Theory and Finance. [Jan. 21, 2020]
International Conference on Modern Mathematical Methods and High Performance Computing in Science & Technology (M3HPCST-2020), Ghaziabad, India. Towards Large-Scale Approximation of Tasks with Derivatives -- A Kernel Perspective. [abstract, slides; Jan. 9, 2020]
International Indian Statistical Association Conference (IISA-2019), IIT Bombay, India. Orlicz Fourier Features. [abstract, slides; Dec. 26-30, 2019]
EPFL, Lausanne, Switzerland. Consistency of Orlicz Random Fourier Features. [abstract, slides; Sept. 23, 2019]
Gatsby 21st Birthday Symposium, London, UK. Orlicz Random Fourier Features. [slides; July 11-13, 2019]
Third International Conference on Mathematics of Data Science (MathoDS 3), City University of Hong Kong (CityU), Hong Kong, China. Outlier-Robust Divergence Estimation on Kernel-Endowed Domains with Median of Means. [abstract, slides, code; June 19-23, 2019]
Workshop at RIKEN AIP. Towards Outlier-Robust Statistical Inference on Kernel-Enriched Domains. [abstract, slides, code; Apr. 15, 2019]
Statistical Seminar in Rennes. From Distance Covariance to Hilbert-Schmidt Independence Criterion. [abstract, slides; Oct. 26, 2018]
EPFL: Laboratory for Information and Inference Systems (LIONS). HSIC, A Measure of Independence? [abstract, slides; Feb. 28, 2018]
ETH Zürich: Department of Biosystems Science and Engineering (D-BSSE): Machine Learning & Computational Biology Lab. HSIC, An Independence Measure? [slides; Feb. 26, 2018]
INRIA Saclay: Tao Seminar. Linear-time Divergence Measures with Applications in Hypothesis Testing. [abstract, slides; Feb. 13, 2018].
Pennsylvania State University: Department of Statistics. Characterizing Independence with Tensor Product Kernels. [slides; Dec. 13, 2017]
Google Brain, Mountain View. Tensor Product Kernels: Independence and Beyond. [abstract, slides; Dec. 1, 2017]
Cubist Systematic Strategies: Advanced Methods Group, New York. Tensor Product Kernels: Characteristic Property and Beyond. [abstract, slides; Nov. 28, 2017]
Yahoo Research, New York. Independence with Tensor Product Kernels. [abstract, slides; Nov. 28, 2017]
ETH Zürich: SfS: Research Seminar. Tensor Product Kernels: Characteristic Property and Universality. [abstract, slides; Nov. 3, 2017]
ENSAE: CREST Statistics Seminar. Characteristic Tensor Kernels. [abstract, slides; Oct. 9, 2017]
Télécom ParisTech: PASADENA Seminar. Data-Efficient Independence Testing with Analytic Kernel Embeddings. [abstract, slides, code; May 17, 2017]
Henri Poincaré Institute: Parisian Statistics Seminar. Distribution Regression: A Simple Technique with Minimax-optimal Guarantee. [abstract, slides, code; Mar. 27, 2017]
Marseilles: Signal Processing and Machine Learning Seminar. A linear-time adaptive nonparametric two-sample test. [abstract, slides, code; Mar. 24, 2017]
Orsay: Probability and Statistics Seminar. Minimax-optimal Distribution Regression. [abstract, slides, code; Mar. 16, 2017]
Télécom ParisTech: Machine Learning Seminar. T-testing: A Linear-time Adaptive Nonparametric Technique. [abstract, slides, code; Feb. 2, 2017]
Dagstuhl Seminar: New Directions for Learning with Kernels and Gaussian Processes. Distribution regression. [slides, code, Dagstuhl report; Dec. 1, 2016]
Facebook AI Research. Adaptive linear-time nonparametric t-test. [abstract, slides, code; Nov. 21, 2016]
Realeyes, Budapest, Hungary. Distinguishing Distributions with Maximum Testing Power. [slides, code; Aug. 24, 2016]
International Workshop on Pattern Recognition in Neuroimaging (PRNI), Trento, Italy. Hypothesis Testing with Kernels. [abstract, slides; June 22-24, 2016]
University of California, San Diego. Kernel-based learning on probability distributions. [slides, code; Apr. 25, 2016]
MASCOT-NUM 2016 @ Institut de Mathématiques de Toulouse, INSA Toulouse. Distribution Regression with Minimax-Optimal Guarantee. [abstract, slides, code; Mar. 23-25, 2016]
MPI, Tübingen: Special Symposium on Intelligent Systems. Performance guarantees for kernel-based learning on probability distributions. [abstract, slides, code; Mar. 15-16, 2016]
École Polytechnique. Optimal Rates for the Random Fourier Feature Technique. [abstract, slides; Mar. 14, 2016]
Imperial College London: Department of Computing. Learning from Features of Sets and Probabilities. [abstract, slides; Mar. 9, 2016]
CMStatistics 2015. Learning Theory for Vector-Valued Distribution Regression. [abstract, slides, code; Dec. 12, 2015]
Pennsylvania State University. Optimal Uniform and Lp Rates for Random Fourier Features. [slides; Dec. 4, 2015]
Carnegie Mellon University: Statistical ML Reading Group. Optimal Rates for the Random Fourier Feature Method. [abstract, slides; Dec. 1, 2015]
Carnegie Mellon University: ML Lunch Seminar. Distribution Regression: Computational and Statistical Tradeoffs. [abstract, slides, code; Nov. 30, 2015]
Princeton University. Distribution Regression: Computational and Statistical Tradeoffs. [abstract, slides, code; Nov. 26-27, 2015]
University of Alberta. Optimal Rates for Random Fourier Feature Approximations. [abstract, slides; Nov. 23-24, 2015]
UC Berkeley: AMPLab. Optimal Rates for Random Fourier Feature Kernel Approximations. [abstract, slides; Nov. 20, 2015]
University of Sheffield: ML@SITraN group. Performance Guarantees for Random Fourier Features - Limitations and Merits. [abstract, slides; arXiv: abstract, paper; June 25-26, 2015]
University of Warwick: Department of Statistics: Centre for Research in Statistical Methodology (CRiSM) Seminars. Regression on Probability Measures: A Simple and Consistent Algorithm. [event, abstract, slides, code; May 29, 2015]
University of Oxford: Department of Statistics: Computational Statistics and Machine Learning reading group. Vector-valued Distribution Regression - Keep It Simple and Consistent. [event, abstract, slides, code; May 1, 2015]
University of Birmingham: Artificial Intelligence and Natural Computation seminars. A Simple and Consistent Technique for Vector-valued Distribution Regression. [event, abstract, slides, code; Jan. 26, 2015]
Max Planck Institute for Intelligent Systems (Tübingen): Bernhard Schölkopf's lab. Consistent Vector-valued Regression on Probability Measures. [abstract, slides, code; Jan. 14-18, 2015]
University of Hertfordshire: Computer Science Research Colloquium. Consistent Distribution Regression via Mean Embedding. [abstract, slides, paper, code; Mar. 5, 2014]

Invited Talk / Poster (Local):
LSE Statistics Research Showcase, Department of Statistics, LSE. The Minimax Rate of HSIC Estimation. [abstract, slides; 20 June, 2024]
LSE Statistics Research Showcase, Department of Statistics, LSE. Kernelized cumulants: Beyond mean embeddings. [slides; 5 June, 2023]
Ph.D. Open Day, Department of Statistics, LSE. Kernel Machines with Shape Constraints. [poster; Nov. 28, 2022]
Combinatorics, Game Theory, and Optimisation (CGO) Seminar, Department of Mathematics, LSE. Support Vector Machines with Hard Shape Constraints. [abstract, slides; Sept. 29, 2022]
Data Science Research Lightning Talks event, DSI, LSE. Shape-Constrained Kernel Machines [slides; Sept. 21, 2022]
LSE Statistics Research Showcase, Department of Statistics, LSE. Continuous Emotion Transfer. [slides; 15 June, 2022]
Research Showcase at the Data Science Institute, LSE. Information Theory, Kernels & Applications. [slides, Dec. 13, 2021]
LSE Statistics Open House. Distribution Regression and Beyond. [slides; Oct. 14, 2021]
LSE Statistics Open House. Applications of Kernel-based Information Theoretical Measures. [slides; Oct. 14, 2021]
B.Sc. Day, CMAP, Ecole Polytechnique. Kernels. [slides; Sept. 4, 2018]
CMAP seminar, Ecole Polytechnique. Adaptive linear-time nonparametric two-sample testing. [abstract, slides, code; Nov. 22, 2016]
'Statistics with coffee' seminar, CMAP, Ecole Polytechnique. Random Fourier Features: Optimal Uniform Bounds. [slides, Oct. 5, 2016]
UCL eResearch Domain launch event, London, UK. Optimal Regression on Sets. [poster; June 29, 2016]
UCL: Statistical Science Seminars: Vector-valued Distribution Regression. A Simple and Consistent Approach. [abstract, slides, code; Oct. 9, 2014]
UCL: CSML Lunch Talk Series. Distribution Regression - the Set Kernel Heuristic is Consistent. [abstract, slides, paper, code; May 2, 2014]

Machine Learning Journal Club (CMAP):
Examples are not enough, learn to criticize! Criticism for Interpretability. [slides; May 4, 2017]

Research Talk (Gatsby):
Optimal Distribution Regression [slides, code; May 23, 2016]
Optimal Uniform and Lp Rates for Random Fourier Features [slides; see also arXiv: abstract, paper; Sept. 7, 2015]
Optimal Rate for Random Kitchen Sinks - Journey to Empirical Process Land [slides; May 18, 2015]
Distribution-to-Anything Regression [slides, code; Sept. 8, 2014]
Consistent, Two-Stage Sampled Distribution Regression [slides, paper, code; Mar. 10, 2014]

Tea Talk (Gatsby):
9 Initialization Strategies [slides; June 28, 2016]
Nim & Friends [slides; Jan. 12, 2016]
The Khintchine Constant and Friends [slides; Sept. 18, 2015]
...k [slides; June 19, 2015]
Supervised Descent Method and its Applications to Face Alignment [slides; Mar. 16, 2015]
Word Storms: Multiples of Word Clouds for Visual Comparison of Documents [slides; Dec. 18, 2014]
The Dvorak Element of the Symmetric Group [slides; Aug. 15, 2014]
Divide and Conquer Kernel Ridge Regression: A Distributed Algorithm with Minimax Optimal Rates [slides; June 10, 2014]
Wasserstein Propagation for Semi-Supervised Learning [slides; Mar. 21, 2014]
Rubik’s on the Torus [slides; Feb. 20, 2014]
On the Chi Square and Higher-Order Chi Distances for Approximating f-Divergences [slides; Dec. 20, 2013]
Smoothing Proximal Gradient Method for General Structured Sparse Regression [slides; Oct. 25, 2013]
Characterizing the Representer Theorem [slides; Oct. 3, 2013]

Machine Learning Journal Club (Gatsby):
Statistical Depth Function [slides; June 26, 2016]
Nonparametric Independence Testing for Small Sample Sizes [slides; Apr. 4, 2016]
Autodiff [slides; Jan. 12, 2016]
Kernels Based Tests with Non-asymptotic Bootstrap Approaches for Two-sample Problems [slides; Nov. 2, 2015]
Random Kitchen Sinks - Revisited [slides; Mar. 12, 2015]
Elementary Estimators for High-Dimensional Linear Regression [slides; Nov. 24, 2014]
Scalable Kernel Methods via Doubly Stochastic Gradients [slides; Oct. 20, 2014]
Fastfood - Approximating Kernel Expansions in Loglinear Time [slides; May 16, 2014]

Kernel Reading Group (Oxford):
Iterative Hessian sketch: Fast and accurate solution approximation for constrained least squares [Aug. 17, 2015]

Quinquennial Review Symposium (Gatsby):
Optimal Uniform and Lp Rates for Random Fourier Features [poster; Sept. 23, 2015]
Bayesian Manifold Learning: Locally Linear Latent Variable Model (LL-LVM) [poster; Sept. 23, 2015]

External Review (Gatsby):
Two-Stage Sampled Distribution Regression on Separable Topological Domains [poster; Oct. 29, 2014]
Bayesian Manifold Learning: Locally Linear Latent Variable Model (LL-LVM) [poster; Oct. 29, 2014]