no code implementations • 2 Apr 2024 • Paul Best, Santiago Cuervo, Ricard Marxer
Macroscopic intelligibility models predict the expected human word-error-rate for a given speech-in-noise stimulus.
no code implementations • 31 Mar 2024 • Santiago Cuervo, Ricard Marxer
We establish a strong correlation between pre-training loss and downstream syntactic and semantic performance in SLMs and LLMs, which results in predictable scaling of linguistic performance.
no code implementations • 24 Jan 2024 • Santiago Cuervo, Ricard Marxer
Our method resulted in the winning submission in the CPC2, demonstrating its promise for speech perception applications.
1 code implementation • 5 Jun 2022 • Santiago Cuervo, Adrian Łańcucki, Ricard Marxer, Paweł Rychlikowski, Jan Chorowski
The success of deep learning comes from its ability to capture the hierarchical structure of data by learning high-level representations defined in terms of low-level ones.
1 code implementation • 29 Oct 2021 • Santiago Cuervo, Maciej Grabias, Jan Chorowski, Grzegorz Ciesielski, Adrian Łańcucki, Paweł Rychlikowski, Ricard Marxer
We investigate the performance on phoneme categorization and phoneme and word segmentation of several self-supervised learning (SSL) methods based on Contrastive Predictive Coding (CPC).
no code implementations • 19 Mar 2021 • Santiago Cuervo, Miguel Melgarejo, Angie Blanco-Cañon, Laura Reyes-Fajardo, Sergio Rojas-Galeano
We present an algorithm for multi-objective optimization of computationally expensive problems.
no code implementations • 21 Jun 2020 • Santiago Cuervo, Marco Alzate
The algorithm is based on the hypothesis that highly correlated actions are a feature of cooperative systems, and hence, we propose the insertion of an auxiliary objective of maximization of the mutual information between the actions of agents in the learning problem.