3 papers
cs.CL2026
Causal Latent Workspaces for Looped Transformers
Ilia Zaznov
Reusing a Transformer block across depth increases computation without replicating its parameters, but successive passes have no dedicated state through which positions can exchang…
cs.AI2026
TT-DAC-PS: Twin-Target Deterministic Actor-Critic with Policy Smoothing for Optimal Trade Execution
Ilia Zaznov, Atta Badii, Julian Kunkel +1
This study addresses the optimal execution of large stock sell programs by introducing TT-DAC-PS (Twin-Target Deterministic Actor-Critic with Policy Smoothing), a deterministic act…
cs.LG2024
AdamZ: An Enhanced Optimisation Method for Neural Network Training
Ilia Zaznov, Atta Badii, Alfonso Dufour +1
AdamZ is an advanced variant of the Adam optimiser, developed to enhance convergence efficiency in neural network training. This optimiser dynamically adjusts the learning rate by…