Picture for Josef Teichmann

Josef Teichmann

Department of Mathematics, ETH Zürich, Switzerland

Learning Chaotic Systems and Long-Term Predictions with Neural Jump ODEs

Add code
Jul 26, 2024
Viaarxiv icon

Robust Utility Optimization via a GAN Approach

Add code
Mar 22, 2024
Viaarxiv icon

Randomized Signature Methods in Optimal Portfolio Selection

Add code
Dec 27, 2023
Viaarxiv icon

Extending Path-Dependent NJ-ODEs to Noisy Observations and a Dependent Observation Framework

Add code
Jul 24, 2023
Viaarxiv icon

Global universal approximation of functional input maps on weighted spaces

Add code
Jun 05, 2023
Viaarxiv icon

How (Implicit) Regularization of ReLU Neural Networks Characterizes the Learned Function -- Part II: the Multi-D Case of Two Layers with Random First Layer

Add code
Mar 20, 2023
Viaarxiv icon

Optimal Estimation of Generic Dynamics by Path-Dependent Neural Jump ODEs

Add code
Jun 28, 2022
Figure 1 for Optimal Estimation of Generic Dynamics by Path-Dependent Neural Jump ODEs
Figure 2 for Optimal Estimation of Generic Dynamics by Path-Dependent Neural Jump ODEs
Figure 3 for Optimal Estimation of Generic Dynamics by Path-Dependent Neural Jump ODEs
Figure 4 for Optimal Estimation of Generic Dynamics by Path-Dependent Neural Jump ODEs
Viaarxiv icon

Applications of Signature Methods to Market Anomaly Detection

Add code
Feb 08, 2022
Figure 1 for Applications of Signature Methods to Market Anomaly Detection
Figure 2 for Applications of Signature Methods to Market Anomaly Detection
Figure 3 for Applications of Signature Methods to Market Anomaly Detection
Figure 4 for Applications of Signature Methods to Market Anomaly Detection
Viaarxiv icon

Infinite width (finite depth) neural networks benefit from multi-task learning unlike shallow Gaussian Processes -- an exact quantitative macroscopic characterization

Add code
Jan 05, 2022
Figure 1 for Infinite width (finite depth) neural networks benefit from multi-task learning unlike shallow Gaussian Processes -- an exact quantitative macroscopic characterization
Viaarxiv icon

Randomized Signature Layers for Signal Extraction in Time Series Data

Add code
Jan 02, 2022
Figure 1 for Randomized Signature Layers for Signal Extraction in Time Series Data
Figure 2 for Randomized Signature Layers for Signal Extraction in Time Series Data
Figure 3 for Randomized Signature Layers for Signal Extraction in Time Series Data
Figure 4 for Randomized Signature Layers for Signal Extraction in Time Series Data
Viaarxiv icon