Víctor Valls


Víctor Valls


Staff Research Scientist


IBM Research Europe - Ireland Dublin, Ireland


I am a research scientist specializing in mathematical optimization. My expertise lies in first-order methods for convex optimization, which are widely used in various fields such as machine learning (ML), network control, and quantum computing. In the past, I have designed online algorithms that can effectively operate under noisy conditions that arise in practical systems. For example, in cloud computing, a job scheduler must allocate resources in a data center without prior knowledge of the job duration. Another example is when an ML algorithm needs to update the model parameters with only a partial view of the dataset. Currently, my research focuses on accelerating the training of Graph Neural Networks (GNNs) for digital healthcare and drug discovery applications.



Active feature acquisition.png

Active Feature Acquisition

Learning complex patterns in knowledge graphs.

Top collaborators

Alessandra Pascale

Alessandra Pascale

Senior Research Scientist and manager, Accelerated Discovery, NLP and Inference on Graphs