Meta
Apple
University of Toronto
Albert GuAssistant Professor, Machine Learning Department · Carnegie Mellon UniversityResearcher↗
Researcher · Academic · Founder · Executive
Ruslan Salakhutdinov
Sooth LabsCo-founder and CSO
Carnegie Mellon UniversityProfessor
🇨🇦 CanadianBorn 1980
Salakhutdinov helped lay the foundations of modern deep learning. With Geoffrey Hinton, he showed how deep neural networks could learn useful patterns in images and documents without human labels.[1] He also co-authored Dropout, a method that helps neural networks generalize beyond their training examples by randomly switching off units during training.[2]
Career & education
Current rolesSooth Labs2026-presentCo-founder and CSOCarnegie Mellon University2016-presentProfessor
Sooth Labs2026-presentCarnegie Mellon University2016-presentEducation
Work and education history is primarily focused on AI-relevant roles and may not be comprehensive.
Citation Trend
Total citations to scholarly work, including research papers, technical reports, books and theses, as recorded by Google Scholar. Each point is a dated snapshot.
citations282,183As of Nov 1, '26
citations+4,286Since last snapshot, Oct 1, '26
Selected contributions
Dropout: A Simple Way to Prevent Neural Networks from Overfitting
Dropout reduces memorization by switching off parts of a neural network during training.
XLNet: Generalized Autoregressive Pretraining for Language Understanding
XLNet learns language context by predicting words in varied orders.
Reducing the Dimensionality of Data with Neural Networks
With Geoffrey HintonShowed how deep neural networks can compress complex data into compact representations while preserving useful structure.
References
- Reducing the Dimensionality of Data with Neural Networks — Science (2006)cs.cmu.edu
With Geoffrey Hinton, he showed how deep neural networks could learn useful patterns in images and documents without human labels.
- Dropout: A Simple Way to Prevent Neural Networks from Overfittingjmlr.org
He also co-authored Dropout, a method that helps neural networks generalize beyond their training examples by randomly switching off units during training.
Based on public sources. If accuracy is important, verify details using the links or references.
For corrections, updates, or requests to remove information, contact us.
Reviewed August 7, 2026 · Checked for accuracy and new information.
