Avatar

Eric M. Fischer

MS Artificial Intelligence

University of California Los Angeles

About Me

I am currently editing and adding significant content to a book covering statistical modeling and learning in vision and cognition, soon to be released by my advisor, Dr. Song-Chun Zhu, and Dr. Ying Nian Wu at UCLA. I am also writing a Masters thesis for an MS in Computer Science, under Dr. Song-Chun Zhu at the Center for Vision, Cognition, Learning, and Autonomy (VCLA) at UCLA.

Before my Masters, I worked as a Full Stack Software Engineer in the San Francisco Bay Area for 3 years, most recently at NatureBox.

My research interests are in generative modeling for computer vision and natural language processing. Specifically, I am interested in applying generative learning techniques to language problems. I have contributed to several research projects that apply MCMC sampling and generative, inference, and/or energy based models to various vision and language tasks. Many projects are on my Github.

PhD Statement of Purpose (Updated 1/28/2020)

Publications

Short-Run MCMC Residual Network toward Energy Based Model

Non-convergent non-persistent short-run MCMC residual network beats VAE and GAN for image synthesis, interpolation, and reconstruction

Generative Recurrent Language Models with MCMC Inference

Utilizes image parsing to text description framework that generates text descriptions based on understanding of image and video content

Masters Thesis in Generative Modeling in Vision and Language

Thesis for Computer Science Masters program currently in progress under my advisor Dr. Song-Chun Zhu at UCLA

Research

Topology Adaptive Snakes Improve Mask Generation for Image Inpainting

T-snake deformable models, which can segment complex-shaped structures, improve generated masks passed to a GAN for image inpainting
Github Code

Exact Sampling with Coupled Markov Chains and Swendsen-Wang Cluster Sampling of the Ising Model

Convergence analysis of exact sampling with Gibbs sampler and coupled Markov chains vs. cluster sampling with Swendsen-Wang algorithm
Full Paper Github Code

Advanced Lane Finding for Self-Driving Cars

Built advanced lane-finding algorithm using distortion correction, image rectification, color transforms, and gradient thresholding
Github Code

Vehicle Detection for Self-Driving Cars

Created vehicle detection and tracking pipeline with OpenCV, histogram of oriented gradients, and SVM, and also with a deep network
Github Code

Implementation and Convergence Analysis of First-Order Optimization Methods for a CNN

Convergence analysis and Python implementations of SGD, SGD with momentum, SGD with Nesterov momentum, RMSprop, and Adam optimizers
Full Paper Github Code

Formulation of Variational Lower Bound and Application of VAE to MNIST Dataset

Formulation of the evidence lower bound (ELBO) for the variational autoencoder and an application to synthesizing binary images
Full Paper Github Code

Projects

Trigger Word Detection

Speech recognition algorithm for trigger word detection, the technology behind Alexa, Google Home, and Siri

Neural Style Transfer using Convolutional Neural Networks

Neural style transfer generates images that reflect the content of one image but the artistic “style” of another

Importance Sampling and Effective Number of Samples

Estimating the number of self-avoiding walks with importance sampling and effective sample size

Network Visualization with Saliency Maps, Fooling Images, and Class Visualization

Explores three techniques that use image gradients to generate new images

Facial Trait and Political Election Analysis by SVM

Trained SVM classifiers to infer 14 facial traits from low-level image features and use that information to make election predictions

Face Detection using AdaBoosting and RealBoosting

Used AdaBoosting, RealBoosting, Haar Filters, Non-Maximum Suppression, and hard negative mining for the task of face detection

PCA, Autoencoder, and FLD for Facial Analysis

Compares Principle Component Analysis, an Autoencoder, and Fisher Linear Discriminants for the task of analyzing human faces

Car Detection for Autonomous Driving

Achieving high-accuracy in real-time, YOLO algorithm is applied to car detection for autonomous vehicles

Experience

 
 
 
 
 

Research Student

Center for Vision, Cognition, Learning, and Autonomy

Mar 2019 – Present University of California, Los Angeles
Contribute to several research projects that apply MCMC sampling and generative, inference, and energy based models to various image and language tasks; NLP text generation project currently in progress

  • Generative Modeling
  • Computer Vision
  • Natural Language Processing
 
 
 
 
 

Full Stack Software Engineer

NatureBox

Mar 2016 – Jan 2018 Redwood City, CA
Main contributor for new React web app after Naturebox added direct-to-consumer business

  • Used Flux/React architecture with Flow and ImmutableJS additions; constructed new backend API
  • Led projects such as Litle to Stripe payment processor migration, Login and Pay with Amazon, Referrals, APIv2
  • Worked on frontend, backend, and with DB, performed most devops, security tasks, led engineering meetings
 
 
 
 
 

Software Engineer

Cinemagram

Sep 2015 – Feb 2016 San Francisco, CA
Worked with JavaScript, Ruby, SQL, and Redis to construct data management interfaces

  • Wrote and ran Snapchat client in PHP for growth campaigns, acquiring 150K users in 6 months
 
 
 
 
 

Frontend Software Engineer

Flinja

Oct 2012 – Mar 2014 Los Angeles, CA
As a main frontend contributor on a small team for startup Flinja, website won DEMO’s (VentureBeat/IDG) Fall 2012 DemoGod Award for best social platform

Keywords

AdaBoosting Autoencoder Autonomous Driving Boosting CNN Computer Vision Generative Modeling Histogram of Oriented Gradients Image Gradients Markov Chain Monte Carlo Monte Carlo Optimization Natural Language Processing Non-Maximum Suppression Object Detection Pattern Recognition RNN Sampling Sequence Model SVM Variational Autoencoder

Contact