Adrián Bazaga

Adrián is an A.I. scientist in Microsoft’s Applied Sciences Group (ASG), where he works on empowering agentic A.I. experiences for Copilot+ devices, leveraging large language models (LLM’s).

He holds a Ph.D. in machine learning from the University of Cambridge, where he conducted research on LLM’s, multimodality, and Generative A.I.  His research has been published in leading machine-learning conferences such as ICLR, ICML, and EMNLP, as well as nature journals.

Before joining ASG, he gained research experience at Microsoft Research and Amazon AGI, where he explored novel training schemes to enhance few-step generation in diffusion models, and test-time scaling for temporal reasoning with LLM’s.  Before that, he spent about 5 years in various startups working at the intersection of A.I. and biology.

To learn more about Adrián, visit his personal webpage and his page on Google Scholar.

• Unsupervised Pretraining for Fact Verification by Language Model Distillation

  International Conference on Learning Representations (ICLR) 2024

  By Adrián Bazaga, Pietro Liò, Gos Micklem
• HyperBERT: Mixing Hypergraph-Aware Layers with Language Models for Node Classification on Text-Attributed Hypergraphs

  Empirical Methods in Natural Language Processing (EMNLP) 2024

  By Adrián Bazaga, Pietro Liò, Gos Micklem
• TabMDA: Tabular Manifold Data Augmentation for Any Classifier using Transformers with In-context Subsetting

  International Conference on Machine Learning (ICML) 2024

  By Andrei Margeloiu, Adrián Bazaga, Nikola Simidjievski, Pietro Liò, Mateja Jamnik