Past events

Jonas Teuwen | Predicting survival in lung cancer with neural networks

CWI Machine Learning Seminar by Jonas Teuwen from Netherlands Cancer Institute & Radboudumc

• iCal

Rianne de Heide | Sequential nonparametric testing with the law of the iterated logarithm - Balsubramani, Ramdas

NB Rescheduled because of the annual meeting of the VVS-OR on March 23rd. In nonparametric testing we would like to make a decision between a null hypothesis (H_0) and an alternative hypothesis (H_1) without making assumptions about underlying distribution(s) of the data being analyzed. The test should maximize the power (minimize false negatives) while maintaining the level of type I errors (false positives), and in addition we focus on a sequential testing framework that accomodates data sets that are extremely large or high-dimensional, or data that arrive sequentially. http://auai.org/uai2016/proceedings/supp/270_supp.pdf

• iCal

Sjoerd Dirksen | Dimensionality reduction with Johnson-Lindenstrauss embeddings

Machine Learning Seminar by Sjoerd Dirksen from RWTH Aachen. Dirksen will talk about dimensionality reduction using random projections.

• iCal

Peter Grunwald | "Bousquet's Version of Talagrand's Inequality": What it is, How to Prove It, and Why It Matters

"Bousquet's Version of Talagrand's Inequality": What it is, How to Prove It, and Why It Matters (from the book "Concentration Inequalities. A Nonasymptotic Theory of Independence. Stephane Boucheron, Gabor Lugosi, and Pascal Massart")

• iCal

Vivek Farias | Learning Preferences with Side Information

This is an interesting seminar hosted by Bert Zwart at CWI

• iCal

Wouter Koolen | Harder, Better, Faster, Stronger Convergence Rates for Least-Squares Regression ---- Dieuleveut, Flammarion, Bach

This will be the first Spring 2017 meeting of the CWI ML reading group. This paper is about first order algorithms (no inverse of no matrix) for stochastic least-squares regression that achieve the correct rate in terms of noise (d/n) and initial conditions (n^-2).

• iCal

National NIPS Debriefing

National NIPS debriefing in Leiden, where PhD students and/or senior researchers will briefly present the paper they found the most interesting at NIPS 2016.

• iCal

Wouter M. Koolen | MetaGrad: Multiple Learning Rates in Online Learning

This is a NIPS oral sneak preview / practice talk. It will follow the NIPS format of 15 minutes presentation plus 5 minutes questions. Your feedback will be very welcome.

• iCal

Emilie Kaufmann | Revisiting the Exploration/Exploitation trade-off in bandit models

• iCal

James Ridgway | On the properties of variational approximations of Gibbs posteriors

• iCal

Scientific Computing course in Delft

Scientific Computing Course for the ESRs of the WAKEUPCALL project Organized by VORtech and TU Delft - TU Delft, Faculty EEMCS, October 3-7, 2016.

• iCal

Gleb Polevoy | The Game of Reciprocation Habits

Gleb Polevoy from Delft Technical University will tell us about his game theoretic results on reciprocation.

• iCal

Nishant Mehta | Sample compression schemes for VC classes

Sample compression schemes for VC classes Shay Moran and Amir Yehudayoff (Journal of the ACM, 2016)

• iCal

Rianne de Heide | Bayesian regression with sparse priors

• iCal

Summer School on Quantitative Methods for Risk Management in Finance and Insurance

The Summer School on Quantitative Methods for Risk Management in Finance and Insurance (SSQM16) is one of the training activities inside WAKEUPCALL, which is a European Industrial Doctorate (EID) Marie Curie Action, funded by the Horizon2020 EU Programme.

• iCal

Wouter Koolen | MetaGrad: Online Convex Optimization in Individual-Sequence and Stochastic Settings

This talk will be about MetaGrad, a new algorithm for Online Convex Optimization.

• iCal

Steve Homer | Automatically Scalable Computation

This is a talk organised by Ronald de Wolf that may be of interest to the ML group.

• iCal

Hannu Reittu | A graph compression method inspired by Szemerédi's Regularity Lemma

• iCal

Wouter Koolen | BISTRO: An Efficient Relaxation-Based Method for Contextual Bandits ---- Rakhlin and Sridharan

This is the latest paper in the "Relaxation" n-logy.

• iCal

Haipeng Luo | Variance-Reduced and Projection-Free Stochastic Optimization

• iCal

Haipeng Luo | Efficient Second Order Online Learning via Sketching

Our visitor Haipeng Luo from Princeton will talk about his latest work.

• iCal

Cristóbal Guzmán | Efficient Noise-Tolerant Learning from Statistical Queries ---- Michael Kearns

The first paper on the statistical query model.

• iCal

Nishant Mehta | Learnability, Stability and Uniform Convergence

• iCal

Spring School in Finance -- Crash course on Credit and Counterparty risk

• iCal

Tom Sterkenburg | Prequential randomness and probability ---- Vovk and Shen

I will start with an accessible introduction to the philosophical/historical background of the problem of defining a randomness notion for infinite outcome sequences. Then I will present the main result of this paper: the equivalence of two different ways -- following the game-theoretic and the measure-theoretic paradigm, respectively -- of defining the randomness of sequences of prediction-outcome pairs. If time permits, we might discuss in some more detail the prequential approach to probability, that motivated this paper.

• iCal

Wouter Koolen | On the Complexity of Best Arm Identification in Multi-Armed Bandit Models ---- Kaufmann, Cappé and Garivier

This will be the first meeting of the CWI ML reading group. We will look at best arm identification, a particular stochastic bandit problem. We will see what these problems are, how their complexity is defined and what bounds we can get from algorithms and hard cases. --- Wouter

• iCal

Supervisory Board Meeting

• iCal

Machine Learning Reading Group goes live

Empty test event

• iCal

Soft skills education for PhD students

• iCal