Adriana Romero

Papers

Total Citations

H-Index

About

Adriana Romero is a leading researcher at the intersection of machine learning, computer vision, and robotics, with a particular focus on multimodal perception and representation learning. Her work explores how intelligent systems can integrate diverse sensory inputs—such as vision and touch—to build richer, more robust understandings of the physical world. In her highly cited 2020 paper, "3D Shape Reconstruction from Vision and Touch," Romero draws inspiration from human developmental cognition, demonstrating how combining high-fidelity local tactile data with global visual information enables more accurate 3D object reconstruction. This work, which has garnered 25 citations, exemplifies her broader contributions to self-supervised learning and embodied AI. Romero is also widely recognized for her pioneering research in network compression and efficient deep learning architectures, including the influential "FitNets" framework. Her work has been published in top venues such as NeurIPS, ICML, and CVPR, and she has received multiple awards for her contributions. Through her innovative fusion of sensory modalities and her commitment to building computationally efficient models, Romero continues to shape how machines perceive, learn, and interact with their environments.