Forrest Iandola

B.S. in Computer Science, University of Illinois at Urbana-Champaign, May 2012.
Illinois Math and Science Academy, May 2008.

I am a PhD student in computer science at University of California, Berkeley. I'm particularly interested in making computer vision and deep learning faster and more efficient. I co-founded DeepScale, where we provide accelerated deep neural network training as a cloud service.

I'm involved in several collaborations including the ParLab, ASPIRE, BVLC, Kurt Keutzer's PALLAS group, and Trevor Darrell's computer vision group. I have also worked at a variety of government and industry R&D institutions including LLNL, Fermilab, SLAC, and NVIDIA. Over the years, my work has spanned computer architecture, supercomputing, medical physics, and computer vision. I did my undergraduate thesis at UIUC with Tarek Abdelzaher.

In my spare time, I like to follow tech news and to see the latest developments in the world of software startups. I also like to stay current with automotive industry news. Jalopnik, Autoblog, and Top Gear are three of my favorite automotive news sources.

Forrest Iandola


FireCaffe: near-linear acceleration of deep neural network training on compute clusters
Forrest N. Iandola, Khalid Ashraf, Mattthew W. Moskewicz, and Kurt Keutzer
arXiv technical report, November 2015. (Paper on arXiv)

DeepLogo: Hitting Logo Recognition with the Deep Neural Network Hammer
Forrest N. Iandola, Anting Shen, Peter Gao, and Kurt Keutzer
arXiv technical report, October 2015. (Paper on arXiv) (Reference in MIT Technology Review) (News article on The Stack)

From Captions to Visual Concepts and Back
Hao Fang, Saurabh Gupta, Forrest Iandola, Rupesh Srivastava, Li Deng, Piotr Dollar, Jianfeng Gao, Xiaodong He, Margaret Mitchell, John C. Platt, C. Lawrence Zitnick, Geoffrey Zweig
Computer Vision and Pattern Recognition (CVPR), 2015. (Paper on arXiv)

Deformable Part Models are Convolutional Neural Networks
Ross Girshick, Forrest Iandola, Trevor Darrell, and Jitendra Malik
Computer Vision and Pattern Recognition (CVPR), 2015. (Paper on arXiv) (Code on GitHub)

libHOG: Energy-Efficient Histogram of Oriented Gradient Computation
Forrest Iandola, Matthew Moskewicz, Kurt Keutzer.
International Conference on Intelligent Transportation Systems (ITSC), 2015.

Audio-Based Multimedia Event Detection with DNNs and Sparse Sampling
Khalid Ashraf, Benjamin Elizalde, Forrest Iandola, Matthew Moskewicz, Julia Bernd, Gerald Friedland, Kurt Keutzer.
International Conference on Multimedia Retrieval (ICMR), 2015. (Paper) (Poster) (Pre-release version of code)

DenseNet: Implementing Efficient ConvNet Descriptor Pyramids
Forrest Iandola, Matt Moskewicz, Sergey Karayev, Ross Girshick, Trevor Darrell, Kurt Keutzer
arXiv technical report, April 2014. (Paper on arXiv) (Code in Caffe featpyra_int_rc branch on GitHub)

Quantifying the Energy Efficiency of Object Recognition and Optical Flow
Michael Anderson, Forrest Iandola, and Kurt Keutzer
UC Berkeley technical report, March 2014. (Paper)

Communication-Minimizing 2D Convolution in GPU Registers
Forrest N. Iandola, David Sheffield, Michael Anderson, Phitchaya Mangpo Phothilimthana, and Kurt Keutzer
IEEE International Conference on Image Processing (ICIP), 2013. (Paper) (Slides) (Code on GitHub)

Deformable Part Descriptors for Fine-grained Recognition and Attribute Prediction
Ning Zhang, Ryan Farrell, Forrest Iandola, and Trevor Darrell
International Conference on Computer Vision (ICCV), 2013. (Paper) (Project Page)

Optimal Load Management System for Aircraft Electric Power Distribution
Mehdi Maasoumy, Pierluigi Nuzzo, Forrest Iandola, Maryam Kamgarpour, Alberto Sangiovanni-Vincentelli and Claire Tomlin
IEEE Conference on Decision and Control (CDC), 2013. (Paper) (Code on GitHub)

Techniques for Assigning Priorities to Streams of Work
Vivek Kini, Forrest Iandola, and Timothy Murray
US Patent #US20140344822.

Representing Range Compensators in the TOPAS Monte Carlo System
Forrest N. Iandola, Jan Schuemann, Jungwook Shin, Bruce Faddegon, Harald Paganetti, and Joseph Perl
European Workshop on Monte Carlo Treatment Planning, Seville, Spain, 2012. (Paper) (Slides)

On Schedulability and Time Composability of Multisensor Data Aggregation Networks
Fatemeh Saremi, Praveen Jayachandran, Forrest Iandola, Yusuf Sarwar, and Tarek Abdelzaher
International Conference on Information Fusion, Singapore, 2012. (Paper) (Slides)

MethMorph: Simulating Facial Deformation due to Methamphetamine Usage
Mahsa Kamali, Forrest N. Iandola, Hui Fang, and John C. Hart
International Symposium on Visual Computing (ISVC), Las Vegas, NV, 2011. (Paper)

Linear Clutter Removal from Urban Panoramas
Mahsa Kamali, Ido Omer, Forrest Iandola, Eyal Ofek, and John C. Hart
International Symposium on Visual Computing (ISVC), Las Vegas, NV, 2011. (Paper)

Real-Time Capacity of Networked Data Fusion
Forrest Iandola, Fatemeh Saremi, Tarek Abdelzaher, Praveen Jayachandran, and Aylin Yener
International Conference on Information Fusion, Chicago, IL, 2011. (Paper) (Slides)

PyMercury: Interactive Python for the Mercury Monte Carlo Particle Transport Code
Forrest N. Iandola, Matthew O'Brien, and Richard Procassini
International Conference on Mathematics and Computational Methods Applied to Nuclear Science and Engineering (M&C), Rio de Janeiro, Brazil, 2011. (Paper) (Poster)

Electron Beam Focusing for the International Linear Collider
Forrest Iandola and Michael Syphers
American Association for the Advancement of Science Annual Meeting, Boston, MA, 2008. (Paper)

Research and Industry Experience

DeepScale, Berkeley, CA
Co-founder and CEO, September 2015 to present
  • We provide cloud-based rapid deep neural network training for advanced and expert users of deep neural networks.
  • I meet with customers, I work on algorithms and infrastructure, and I do whatever needs to be done.

University of California, Berkeley, CA
Graduate Student Researcher, August 2012 to present
Visiblend, Berkeley, CA
Co-founder, May 2014 to May 2015
  • Using Deep Learning to produce uplift on video ad campaigns.

Microsoft Research, Redmond, WA
Research Intern, June 2014 to September 2014
  • Worked on deep and multimodal (images+language) learning at scale.

NVIDIA Corporation, Santa Clara, CA
CUDA Intern, June 2012 to August 2012
  • Added kernel priority scheduling to the CUDA programming language. My co-authors and I added the cudaStreamCreateWithPriority() function to CUDA.

SLAC National Accelerator Laboratory, Stanford University, Menlo Park, CA
Research Intern, June 2011 to December 2011
  • Worked with Joseph Perl on the Geant4 Monte Carlo particle transport simulation toolkit.
  • Designed computational geometry methods for modeling and simulating radiation cancer therapy systems.

University of Illinois, Urbana, IL
Research Assistant, December 2009 to May 2012
Lawrence Livermore National Laboratory, Livermore, CA
Research Intern, May 2010 to August 2010
  • Worked with Matthew O'Brien, Scott McKinley, and Richard Procassini to develop the Mercury parallel Monte Carlo particle transport simulation.
  • Designed and implemented a Python API for Mercury.

State Farm Research Center, Champaign, IL
Systems Analyst Intern, September 2009 to December 2009
  • Was one of the people who dreamed up the "take a picture of a car, and get an insurance quote" app that's in the app store now.
  • Five years later, I briefly worked as a computer vision consultant to make some improvements.

Northwestern University Articulab, Evanston, IL
Research Intern, May 2008 to August 2008
Fermi National Accelerator Laboratory, Batavia, IL
Research Intern, Summer 2006 and Summer 2007

Talks and Posters

Application-Driven Research in the ASPIRE Lab
Michael Anderson, Khalid Ashraf, Gerald Friedland, Forrest Iandola, Peter Jin, Matt Moskewicz, Zach Rowisnki, Kurt Keutzer
ASPIRE Retreat, May 2014 (Slides)

DenseNet: Efficient Computation of Deep Neural Networks for Object Detection
Forrest Iandola, Matt Moskewicz, Sergey Karayev, Ross Girshick, Kurt Keutzer, and Trevor Darrell
Presented versions of this at BVLC and ASPIRE retreats, March-May 2014. (Poster)

Boda: Towards a Framework for Evaluating Accuracy/Efficiency Tradeoffs in Object Detection
Matt Moskewicz, Forrest Iandola, and Kurt Keutzer
ASPIRE Retreat, Jan 2014 (Poster)

Scaling Up Deformable Parts Models for Object Detection
Forrest Iandola, Ning Zhang, Ross Girshick, Trevor Darrell, and Kurt Keutzer
ASPIRE Retreat, May 2013 (Poster)

Minimizing Memory Communication for 2D Image Convolution in GPU Registers
Forrest Iandola, David Sheffield, Michael Anderson, Mangpo Phothilimthana, and Kurt Keutzer
GPU Technology Conference, March 2013 (Poster)


650-200-0082 (C)

My Blog

Forrest Iandola on LinkedIn Forrest Iandola on Twitter Forrest Iandola on Facebook