Overparametrisation and the Bias-variance Dilemma

EI Seminar
Person writing on a whiteboard

For several machine learning methods such as neural networks, good generalisation performance has been reported in the overparametrised regime. In view of the classical bias-variance trade-off, this behavior is highly counterintuitive.

Speaker
Johannes Schmidt-Hieber
Date
Thursday 17 Apr 2025, 12:00 - 13:00
Type
Seminar
Room
ET-14
Location
Campus Woudestein
Add to calendar

Joint work with Alexis Derumigny (Delft).

We will present a general framework to establish universal lower bounds for the bias-variance trade-off.

See also

Optimal Maximin GMM Tests for Sphericity in Latent Factor Analysis of Short Panels

Patrick Gagliardini (University of Lugano)
Campus with students

FinEML Conference 2025

Financial Econometrics Meets Machine Learning
Google Deep mind

Assessing the Value of Vessel Information Sharing Through Simulation

Pim van Leeuwen (Erasmus School of Economics)
Image - Container Ship

Shapley Instruments

Giovanni Mellace (University of Southern Denmark, SDU)
People with Post IT notes

4th International Econometrics PhD Conference 2025

Bringing together PhD candidates from universities around the world specialising in Econometrics to present and discuss their research.
Campus Woudestein met het oog op het fontein

Early Birds Get the Vol: Morning Volatility Uncertainty and Variance Risk Premium

Rodrigo Hizmeri (University of Liverpool Management School)
A person jumping off from gray concrete building
More information

Do you want to know more about the event? Contact the secretariat Econometrics at eb-secr@ese.eur.nl.

Compare @count study programme

  • @title

    • Duration: @duration
Compare study programmes