Oscar Guarnizo

Hi, I'm a computer scientist from Ecuador with experience in AI/ML. I recently completed a master's degree in AI/ML at the University of Birmingham, fully funded by the Google DeepMind International Scholarship.

My current research interests include how to create more robust AI systems capable of adapting, focusing, or reusing knowledge. I'm intrigued by the question of why current AI struggles to preserve and reuse knowledge. Currently, I'm keen on deepening my understanding exploring topics such as:

  • Reinforcement learning
  • Representation learning
  • Continual learning
  • Meta learning

After my undergrad, I worked as a part-time Junior Researcher at Scientific Computing Group, supervised by Israel Pineda and at DeepARC Research supervised by Eugenio Morocho. Then, I held a research internship at KAUST, working on StyleGANs supervised by Peter Wonka. Additionally, I had some experience in the industry, working as a Machine Learning engineer at Mobia Latam (start-up) and DIGEVO.

During the past years, I was actively participating in Oxford ML 2024, EEML 2022, EEML 2021 (organized by DeepMind), IssoDL2021, and Workshop MathRL 2021, strengthening my basic knowledge and interpersonal skills.

I firmly believe in the power of knowledge and that anyone can learn/practice whatever they want without needing to be an expert.

 /  /  / / CV / Google Scholar / Medium

profile photo
Skills

Spanish (Native)
English
(IELTS C1 - 7/9)
(TOEFL iBT C1 - 104/120)

CONFIDENT :
Python | Pandas | Latex | PyTorch | TensorFlow | Nvidia Deepstream | Nvidia TAO Toolkit

FAMILIAR :
OpenCV | Docker | Kafka | Unix | C

Coursework

Artificial Intelligence & Neural Networks
Fundamentals of Programming *
Databases Management
Operating Systems
Functional Programming **
Computer Graphics & Image Processing
Probabilities and Statistics **

*Teaching Assistant
**Research Assistant

MOOC Training

Deep Learning Specialization
Video AI Applications on Jetson Nano
Machine Learning
Data Structures + Algorithms
Learning How to Learn
Neuromatch Deep Learning Academy

Research

I have experience in computer vision and deep learning for performing tasks such as classification, detection, segmentation, and image generation (using StyleGAN).

Bisimulation Prioritized Experience Replay: Enhancing Online Reinforcement Learning through Behavioral-Based Priorities
Oscar Guarnizo, Mirco Giacobbe, Leonardo Stella
Master Thesis, 2023-2024
project page thesis / poster article / demo slides / code

Bisimulation Prioritized Experience Replay (BPER) introduces a novel approach to enhance value-based reinforcement learning by incorporating a bisimulation metric to prioritize behaviorally relevant transitions. By balancing traditional TD-error with state behavior similarities, BPER improves data diversity and learning efficiency, as demonstrated in a 31-state Grid World and pixel-based environments.

CovarianceNet: CNN Feature Extraction Using Covariance Tensor Decomposition
Ricardo Fonseca, Oscar Guarnizo, Diego Suntaxi, Alfonso Cadiz, Werner Creixell
IEEE Access, 2021
EEML, 2022 (Oral Presentation, Best Poster Award)
project page / paper / poster / video / blog / code

During a single step forward, we compute the covariance tensor and factorize it by Tucker decomposition to generate kernels for a CNN. The kernels were then plugged into the CNN (CovarianceNet) for classification.

StyleGAN Embedding Algorithms: Image2StyleGAN (I2S) & ImprovedImage2StyleGAN (II2S)
Oscar Guarnizo, Peter Wonka
EEML, 2021
project page / poster article / blog / code

We reproduced StyleGAN embedding algorithms, adding further experimentation such as inpainting, super-resolution, colorization, morphing, style transfer, and expression transfer.

Three Dimensional Adaptive Path Planning Simulation Based On Ant Colony Optimization
Oscar Guarnizo, Israel Pineda
LA-CCI, 2019   (Oral Presentation)
project page / paper / thesis / video / code / extended code

Path planning based on Ant Colony Optimization was studied to deal with stuck conditions and speed up convergence.

Successive Adaptive Linear Neural Modeling for Equidistant Real Roots Finding
Joseph Gonzalez, Fernando Zhapa, Oscar Guarnizo, Francisco Ortega-Zamorano
ETCM, 2018   (Oral Presentation)
project page / paper / video / code

Successive Adaptive Linear Neural Modeling is a two-step approach based on a Self Organized Map (SOM) and an Adaptative Linear Neuron (Adaline).

Software

Additionally, I developed some software projects related to deep learning, computer vision, and video games.

Hangman Model: Finetuning a BERT Model for Token Classification using Character-Level Masked Language Modeling (MLM)
Oscar Guarnizo,
Self-Learning , May 2024
code

This project aims to fine-tune a BERT-based model, called HangmanNet, for character-level masked language modeling to play the Hangman game. The model predicts hidden characters by leveraging bidirectional context from both sides of a word. Additional techniques, such as incorporating prior character frequency and custom data collators, were explored to improve accuracy, particularly for shorter words.

Minimal RL: Fundamental Reinforcement Learning Algorithms
Oscar Guarnizo,
Self-Learning, 2023 - Present
code

This project compares and implements key model-free Reinforcement Learning (RL) algorithms, focusing on clarity over complexity. Using online resources for guidance, it aims to provide accessible, easy-to-understand implementations, serving as an educational tool for those exploring RL's fundamental concepts.

Automatic License Plate Recognition (Nvidia TAO + Deepstream Application + Python Bindings)
Oscar Guarnizo, Leduin Cuenca
Mobia Latam, 2022
project page / code

We developed a multistream Automatic License Plate Recognition (ALPR) system by assembling an Nvidia deepstream pipeline, training 3+ models with the Nvidia TAO toolkit, and periodically sending data streams to a Kafka Apache for further use cases.

Segmentating and Counting Grapes Bunches using MaskRCNN + Tracker DeepSort
Oscar Guarnizo, Diego Suntaxi, Fabricio Crespo
Digevo, 2021
project page / video / code

We developed a system to segment and count grapes bunches using Mask RCNN and DeepSort Tracker. After that, we extrapolate the counting information to satellite images to generate heat maps that show the number of grapes per parcel in a yield.

Retro Video Game Programming Challenge - GMTK Game Jam 2020
Esteban Lasso, Oscar Guarnizo, Joseph Gonzalez, Jose Seraquive
GMTK Game Jam , 2020
project page / video game

We developed a retro video game in 48 hours using Unity3D to measure our expertise and continue improving. Game Jam Topic: Out of Control

Learning
Learning Notes: 2st Semester - MSc in AI/Ml, University Birmingham, United Kingdom
Jan 2024, April 2024 (In Progress)
Oscar Guarnizo
notes

A compilation of my notes during the second semester of the master program in Artificial Intelligence and Machine Learning. I covered the following topics: Natural Language Processing, Computational Vision and Imaging, and Current Topics of AI/ML.

Learning Notes: 1st Semester - MSc in AI/Ml, University Birmingham, United Kingdom
Sept 2023, November 2023 (In Progress)
Oscar Guarnizo
notes

A compilation of my notes during the first semester of the master program in Artificial Intelligence and Machine Learning. I covered the following topics: Machine Learning, Deep Learning, and Mathematics for AI/ML.

Introduction to Pandas
Release Date: 2023 (In Progress)
Oscar Guarnizo

Introduction to Pandas library. This short introduction includes 8 jupyter notebooks to understand the fundamentals and minimum requirements for an job interview.

Miscellaneous
Activities Deep Learning Reading Group, University of Birmingham
Eastern European Machine Learning Summer School
International Summer School on Deep Learning
Resources Artificial Intelligence and Games by Georgios Yannakakis & Julian Togelius
How to Take Smart Notes by Sönke Ahrens (Zettelkasten Method)

Powered by Jon Barron.