Scholar
Phillip Isola
Google Scholar ID: ROILf3EAAAAJ
Associate Professor, MIT
Citations & Impact
Citations
101,797
H-index
58
i10-index
86
Publications
20
Co-authors
193
list available
Contact
Publications
Resume (English only)
Academic Achievements
- - Released a free online version of the 'Foundations of Computer Vision' textbook: visionbook.mit.edu
- - Key research projects include:
- - Deep representation learning (e.g., Platonic Representation Hypothesis, Low-rank bias, Understanding contrastive learning)
- - Generative intelligence (e.g., Learning from models, Learning from NeRFs, Denoised world models)
- - World representations for agents (e.g., F3RM, Embodied representation learning, Mental imagery for robots)
- - Emergent intelligence (e.g., Learning without data, Neural MMO, PowderWorld)
Research Experience
- - Spent a year as a visiting research scientist at OpenAI
- - Worked as a postdoctoral scholar at UC Berkeley
- - Research group focused on scientifically understanding intelligence, especially human-like intelligence, involving deep nets, adaptability, general-purpose, and emergent from embodied interactions in rich ecosystems
Education
- - Bachelor's degree in Computer Science from Yale, advisor: Brian Scholl
- - Ph.D. in Brain & Cognitive Sciences from MIT, advisor: Ted Adelson, also frequently worked with Aude Oliva
- - Postdoctoral scholar in the EECS department at UC Berkeley, advisor: Alyosha Efros
Background
- Associate Professor in EECS at MIT, studying computer vision, machine learning, and AI.
Miscellany
- Personal interests not mentioned in the provided HTML content
