Sayna Ebrahimi
PhD Candidate

EECS Department
UC Berkeley

Email : sayna at eecs dot berkeley dot edu

Download CV
Generic placeholder image Generic placeholder image Generic placeholder image


About Me

I am a double majored PhD student in Computer Science and Mechanical Engineering at UC Berkeley where I work on deep learning and computer vision problems. I am a part of Berkeley AI Research Lab (BAIR) and Berkeley DeepDrive (BDD) advised by Trevor Darrell in EECS and David Steigmann in ME. I also collaborate with Marcus Rohrbach and Mohamed Elhoseiny at FAIR.



Research Interests:

Deep Learning, Computer Vision, Lifelong learning, Meta Learning, Predictive Uncertainty, Active Learning



Recent Publications

Gradient-Free Supervised and Unsupervised Learning with Rewards
Sayna Ebrahimi, Anna Rohrbach, Trevor Darrell
WiCV workshop at CVPR , 2018 (Oral)
[PDF]



Surprisingly Simple Deep Confidence Prediction
Sayna Ebrahimi, Daniel Geng, Trevor Darrell
Under Review ECCV , 2018
[Code will be available soon]



Gradient-free Policy Architecture Search and Adaptation
Sayna Ebrahimi, Anna Rohrbach, Trevor Darrell
Conference on Robot Learning (CoRL), 2017 (Spotlight)
[PDF] [Project page] [Bibtex]








Older Work
The following represents some of my older work in ME and is not representative of my current research interests.

Journal papers:

Peridynamics analysis of the nanoscale friction and wear properties of amorphous carbon thin films
Sayna Ebrahimi, David Steigmann, Kyriakos Komvopoulos
In Journal of Mechanics of Materials and Structures 2015.
[
PDF] [Bibtex]


Dynamic analysis of cracks using the SGBEM for elastodynamics in the Laplace-space frequency domain
Sayna Ebrahimi, Anh-vu Phan
Engineering Analysis with Boundary Elements 2013.
[PDF] [Bibtex]


Dynamic crack growth modeling technique based upon the SGBEM in the Laplace domain
Sayna Ebrahimi, Anh-vu Phan
Acta Mechanica 2015.
[PDF] [Bibtex]




Conference Papers/Talks:

Parallel Algorithms for CPU and GPU Peridynamic Computation
Sayna Ebrahimi, Michael Taylor, David Steigmann
International Mechanical Engineering Congress & Exposition (ASME-IMECE), 2016, Phoenix, Arizona (Oral)


A Mechanical Sub-Element Formulation of Plasticity in Ordinary State-Based Peridynamics
Sayna Ebrahimi, Michael Taylor, David Steigmann
International Mechanical Engineering Congress & Exposition (ASME-IMECE), 2016, Phoenix, Arizona (Oral)


Peridynamic Model for Thin Shells via Descent from Three-Dimensional State-Based and Bond-Based Peridynamics
Michael Taylor Sayna Ebrahimi, David Steigmann
International Mechanical Engineering Congress & Exposition (ASME-IMECE), 2016, Phoenix, Arizona (Oral)


Peridynamics Analysis of Elastic-Plastic Contacts
Sayna Ebrahimi, David Steigmann, Kyriakos Komvopoulos,
US National Congress on Computational Mechanics (UNCCM), 2015, San Diego, California (Spotlight)


Wear process analysis of thin-films using Peridynamics
Sayna Ebrahimi, David Steigmann, Kyriakos Komvopoulos,
US National Congress on Theoretical and Applied Mechanics (UNCTAM), 2014, East Lansing, Michigan (Spotlight)


Laplace SGBEM Modeling of Dynamic Crack Propagation through a Cluster of Inclusions
Sayna Ebrahimi, Anh-Vu Phan
US National Congress on Theoretical and Applied Mechanics (UNCTAM), 2014, East Lansing, Michigan (Spotlight)


Dynamic crack growth analysis in the Laplace-space frequency domain by the symmetric-Galerkin boundary element method
Sayna Ebrahimi, Anh-Vu Phan
International Mechanical Engineering Congress and Exposition (ASME-IMECE), 2013, Raleigh, North Carolina (Oral)


Peridynamic Analysis of Crack-Inclusion Interaction in Unidirectional Fiber-Reinforced Composites
Sayna Ebrahimi, Anh-Vu Phan
Early Career Technical Conference (ASME-ECTC), 2012, Atlanta, Georigia (Spotlight)