Charlie Lehman

Charlie Lehman

PhD Student

Georgia Institute of Technology

Biography

Charlie Lehman is a Machine Learning Ph.D. student at the Georgia Tech Omni Lab fro Visual Engineering and Science (OLIVES). His research interests include robustness and explainability of deep vision models. He is also an Engineering Duty Officer with the U.S. Navy Reserves, where he specializes in the maintenance and repair of surface vessels.

Interests

  • Machine Learning
  • Robustness
  • Digital Image Processing
  • Visualization
  • Dry Docking

Education

  • PhD in Machine Learning, On going

    Georgia Institute of Technology

  • MSc in Electrical and Computer Engineering, 2018

    Georgia Institute of Technology

  • BSc in Electrical and Computer Engineering, 2009

    The Citadel

Skills

Python

90%

Data Visualization

100%

Dry Docking Ships

100%

Experience

 
 
 
 
 

Computer Vision and Machine Learning Intern

Stealth Startup

May 2020 – Present
 
 
 
 
 

President

ConvexMind

Mar 2018 – Present Georgia
 
 
 
 
 

Graduate Research Assistant

Georgia Institute of Technology - OLIVES

Sep 2016 – Present Atlanta, Georgia

Advisor - Ghassan AlRegib, Ph.D.

Researching methods and the underlying properties of deep vision models that lead to improvements in robustness.

 
 
 
 
 

Graduate Research Assistant

Georgia Institute of Technology - AMBER Lab

Aug 2016 – Dec 2016 Atlanta, Georgia

Advisor - Aaron Ames, Ph.D.

Designed and installed a computer vision system for the DURUS bipedal robot for use in estimating translations and heading.

 
 
 
 
 

Graduate Research Assistant

George Mason University - Lofaro Lab

Feb 2016 – Apr 2016 Fairfax, Virginia

Advisor - Daniel Lofaro, Ph.D.

Robotics Simulation

 
 
 
 
 

Engineering Duty Officer

United States Navy Reserves

Oct 2015 – Present Georgia
Currently assigned to the NAVSEA Heavy Lift Unit and Dry-docking Reserve Unit.
 
 
 
 
 

Engineering Duty Officer

United States Navy

May 2009 – Oct 2015 Georgia

Officer in the United States Navy responsible for the design, acquisition, construction, repair, maintenance, conversion and overhaul of ships, submarines, aircraft, carriers and their respective systems.

My focus areas are in the maintenance and repair of surface ships. In particular, I am an expert in dry docking.

 
 
 
 
 

Undergraduate Student

The Citadel

May 2007 – Mar 2009 Charleston, South Carolina

Advisor - Robert Barsanti, Ph.D.

Wavelet Receiver Design

Recent Posts

On the Structures of Representation for the Robustness of Semantic Segmentation to Input Corruption

SS-08:Dynamic Background Reconstruction/Subtraction for Challenging Environments
Oct 22, 2020 25 min read

Visualizing Softmax

Introduction I have been faced with several situations where visualizing the output of a trained classifier has helped explain some interesting behaviors that are present within learned representations. I want to share a simple technique to visualize outputs for SoftMax-based classifiers.
Dec 9, 2019 5 min read
Visualizing Softmax

Learning to segment CIFAR10

Import Important Things import torch from torch import nn import torch.nn.functional as F from torch.utils.data import DataLoader from torchvision import transforms from torchvision.models import resnet18 from torchvision.datasets import CIFAR10 from tqdm import tqdm_notebook as tqdm from torchvision.
Oct 17, 2019 5 min read
Learning to segment CIFAR10

Speechworks Interview

I was interviewed by Julie Lindsay from Speechworks for a podcast regarding my experiences starting ConvexMind during the CreateX Startup Launch last Summer. There is also an article that summarizes the main points we discussed.
Aug 13, 2019 1 min read
Speechworks Interview

Visualizing Adversarial Attacks

Pixelwise classifiers with Large-margin Gaussian Mixture representation These are several visualizations I dug up from really really messy code, which I won't share just yet. This was some quick and dirty work to test out the feasibility of an idea: I wanted to visualize the effects on the representation space at each pixel during an FGSM attack.
Dec 13, 2018 2 min read
See all posts

Projects

Blackjack Simulator

Project for ECE8823 Convex Optimization course at Georgia Tech
Code
Blackjack Simulator

Recent & Upcoming Talks

Robustness and Overfitting Behavior of Implicit Background Models

SS-13:Explainable Machine Learning for Image Processing
Oct 28, 2020 8:00 AM — 10:25 AM SS-13:Explainable Machine Learning for Image Processing
PDF Code Video
Robustness and Overfitting Behavior of Implicit Background Models

On the Structures of Representation for the Robustness of Semantic Segmentation to Input Corruption

SS-08:Dynamic Background Reconstruction/Subtraction for Challenging Environments
Oct 27, 2020 9:30 AM — 12:55 PM SS-08:Dynamic Background Reconstruction/Subtraction for Challenging Environments
PDF Code Video
On the Structures of Representation for the Robustness of Semantic Segmentation to Input Corruption

S6:Semi-Supervised Self-Supervised Semantic Segmentation

ARS-14: Machine Learning for Image and Video Classification IV
Oct 27, 2020 9:00 AM — 9:25 AM ARS-14: Machine Learning for Image and Video Classification IV
PDF Video
S6:Semi-Supervised Self-Supervised Semantic Segmentation

Publications

On the Structures of Representation for the Robustness of Semantic Segmentation to Input Corruption

Semantic segmentation is a scene understanding task at the heart of safety-critical applications where robustness to corrupted inputs …
PDF Code Poster
On the Structures of Representation for the Robustness of Semantic Segmentation to Input Corruption

Robustness and Overfitting Behavior of Implicit Background Models

In this paper, we examine the overfitting behavior of image classification models modified with Implicit Background Estimation …
PDF Code Poster Video
Robustness and Overfitting Behavior of Implicit Background Models

S6:Semi-Supervised Self-Supervised Semantic Segmentation

Semi-supervised learning provides a means to leverage unlabeled data when labels are expensive to obtain. In this work, we propose a …
PDF Poster Video
S6:Semi-Supervised Self-Supervised Semantic Segmentation

Implicit Background Estimation for Semantic Segmentation

Scene understanding and semantic segmentation are at the core of many computer vision tasks, many of which, involve interacting with …
PDF Code Poster
Implicit Background Estimation for Semantic Segmentation

Application of a Wavelet-Based Receiver for the Coherent Detection of FSK Signals

This paper investigates the application of wavelet analysis to the problem of coherent detection of digital binary frequency shift …
PDF

Popular Topics

Class Project Digital Signal Processing ICIP Implicit Background Estimation Machine Learning Semantic Segmentation Special Session

Contact