Analytical Study of Momentum-Based Acceleration Methods in Paradigmatic High-Dimensional Non-Convex Problems
–Neural Information Processing Systems
In this work, we use dynamical mean field theory techniques to describe analytically the average dynamics of these methods in a prototypical non-convex model: the (spiked) matrix-tensor model. We derive a closed set of equations that describe the behaviour of heavy-ball momentum and Nesterov acceleration in the infinite dimensional limit.
Neural Information Processing Systems
Oct-1-2025, 21:07:05 GMT
- Country:
- Africa > Middle East
- Tunisia > Ben Arous Governorate > Ben Arous (0.04)
- Asia > Russia (0.04)
- Europe
- France (0.04)
- Russia (0.04)
- United Kingdom > England
- Oxfordshire > Oxford (0.14)
- North America > United States
- Louisiana > Orleans Parish > New Orleans (0.04)
- Africa > Middle East
- Technology: