1 code implementation • NeurIPS 2023 • Adrián Javaloy, Pablo Sánchez-Martín, Isabel Valera
In this work, we deepen on the use of normalizing flows for causal reasoning.
1 code implementation • 21 Nov 2022 • Adrián Javaloy, Pablo Sanchez-Martin, Amit Levi, Isabel Valera
Existing Graph Neural Networks (GNNs) compute the message exchange between nodes by either aggregating uniformly (convolving) the features of all the neighboring nodes, or by applying a non-uniform score (attending) to the features.
1 code implementation • 9 Jun 2022 • Adrián Javaloy, Maryam Meghdadi, Isabel Valera
We refer to this limitation as modality collapse.
no code implementations • NeurIPS 2021 • Adrián Javaloy, Isabel Valera
Multi-task learning is being increasingly adopted in applications domains like computer vision and reinforcement learning.
2 code implementations • ICLR 2022 • Adrián Javaloy, Isabel Valera
Multitask learning is being increasingly adopted in applications domains like computer vision and reinforcement learning.
no code implementations • 1 Jan 2021 • Adrián Javaloy, Isabel Valera
GradNorm eases the fitting of all individual tasks by dynamically equalizing the contribution of each task to the overall gradient magnitude.
1 code implementation • NeurIPS 2020 • Luigi Gresele, Giancarlo Fissore, Adrián Javaloy, Bernhard Schölkopf, Aapo Hyvärinen
Learning expressive probabilistic models correctly describing the data is a ubiquitous problem in machine learning.
2 code implementations • 26 Feb 2020 • Adrián Javaloy, Isabel Valera
While MTL solutions do not directly apply in the probabilistic setting (as they cannot handle the likelihood constraints) we show that similar ideas may be leveraged during data preprocessing.