no code implementations • 11 Oct 2016 • João P. Oliveira, Ana Bragança, José Bioucas-Dias, Mário Figueiredo, Luís Alcácer, Jorge Morgado, Quirina Ferreira
In this article, we present a denoising algorithm to improve the interpretation and quality of scanning tunneling microscopy (STM) images.
no code implementations • 14 Nov 2023 • Arlindo L. Oliveira, Tiago Domingos, Mário Figueiredo, Pedro U. Lima
We argue that the internal architecture of LLMs and their finite and volatile state cannot support any of these properties.
1 code implementation • 14 Jul 2020 • João Pedro Araújo, Mário Figueiredo, Miguel Ayala Botto
This paper introduces single-partition adaptive Q-learning (SPAQL), an algorithm for model-free episodic reinforcement learning (RL), which adaptively partitions the state-action space of a Markov decision process (MDP), while simultaneously learning a time-invariant policy (i. e., the mapping from states to actions does not depend explicitly on the episode time step) for maximizing the cumulative reward.