NeurIPS Debriefing

The National NeurIPS debriefing is a yearly seminar (inspired by a similar seminar in Paris), where multiple PhD students and/or senior researchers from anywhere in the Netherlands give short presentations on the paper they found the most interesting at the previous NeurIPS conference. If you are interested in NeurIPS, please consider being one of the presenters.

The format is to have two hours of short talks (15 or 20 minutes, in English). It is not required to have attended NeurIPS, and presenters would usually not present their own papers. Talks can be informal, in a friendly atmosphere, so this is an ideal opportunity for PhD students to gain experience in giving presentations.

If you are interested in presenting a paper or if you have any questions, please contact me (Tim van Erven).

Upcoming seminars will be announced on the machine learning Netherlands mailing list.

NeurIPS 2018 Debriefing

Thursday February 14, 2019
15h00-17h00
Leiden University, Snellius Building, Niels Bohrweg 1, Leiden
Room 176

Preliminary list of speakers:

Rémy Degenne, CWI Coordinate Descent with Bandit Sampling by Salehi, Thiran, Celis
Changyong Oh, UvA Geometrically Coupled Monte Carlo Sampling by Rowland et al.
Zakaria Mhammedi, Australian National University Direct Runge-Kutta Discretization Achieves Acceleration by Zhang et al.
Wouter Koolen, CWI Acceleration through Optimistic No-Regret Dynamics by Wang and Abernethy

NIPS 2017 Debriefing

Wednesday February 28, 2018
14h00-16h00
Leiden University, Snellius Building, Niels Bohrweg 1, Leiden
Room 174

Thomas Moerland Thinking Fast and Slow with Deep Learning and Tree Search by Anthony, Tian, Barber slides
Dirk van der Hoeven Parameter free online learning via model selection by Foster, Kale, Mohri, Sridharan
Wouter Koolen Safe and Nested Subgame Solving for Imperfect-Information Games by Brown, Sandholm
William Weimin Yoo Dynamic Routing Between Capsules by Sabour, Frosst, Hinton slides
Tim van Erven A high-level overview of recent thinking on why neural networks generalize, based on:

slides

NIPS 2016 Debriefing

Friday January 20, 2017
15h00-16h30
Leiden University, Snellius Building, Niels Bohrweg 1, Leiden
Room 402

Nishant Mehta: On the Recursive Teaching Dimension of VC Classes by Chen^2, Cheng, Tang
Kevin Duisters: Matrix Completion has No Spurious Local Minimum by Ge, Lee, Ma
Dirk van der Hoeven: Blazing the trails before beating the path: Sample-efficient Monte-Carlo planning by Grill, Valko, Munos
Tim van Erven: Deep Learning without Poor Local Minima by Kawaguchi