5 citations · 8 across the 8 of their papers we have counts for
5 papers · 1 filter
Improving Neural Network Training by Decoupling the Magnitude and Direction of Weight Vectors
Alexander Hägele, Alejandro Hernández-Cano, Atli Kosson +1
Modern neural network training relies on optimizers such as Adam and Muon which act on each weight matrix as a single object. Yet every weight matrix carries two distinct quantitie…
Training Dynamics of the Cooldown Stage in Warmup-Stable-Decay Learning Rate Scheduler
Aleksandr Dremov, Alexander Hägele, Atli Kosson +1
Learning rate scheduling is essential in transformer training, where the final annealing plays a crucial role in getting the best performance. However, the mechanisms behind this c…
The Surprising Agreement Between Convex Optimization Theory and Learning-Rate Scheduling for Large Model Training
Fabian Schaipp, Alexander Hägele, Adrien Taylor +2
We show that learning-rate schedules for large model training behave surprisingly similar to a performance bound from non-smooth convex optimization theory. We provide a bound for…
Scaling Laws and Compute-Optimal Training Beyond Fixed Training Durations
Alexander Hägele, Elie Bakouch, Atli Kosson +3
Scale has become a main ingredient in obtaining strong machine learning models. As a result, understanding a model's scaling properties is key to effectively designing both the rig…
BaCaDI: Bayesian Causal Discovery with Unknown Interventions
Alexander Hägele, Jonas Rothfuss, Lars Lorch +3
Inferring causal structures from experimentation is a central task in many domains. For example, in biology, recent advances allow us to obtain single-cell expression data under mu…