Peter Conant's Portfolio
SJSU Master Student | Artificial Intelligence and Software Engineering
Embedded Files
I'm a Masters student Major Artificial Intelligence. My portfolio includes data analysis/prediction models, Computer Vision Tasks, Natural Language Projects projects . I'm seeking AI-driven internship and full time opportunities.
An tonal contort extractor designed for dolphin whistles created by Mary Roche. I had a pleasure of contributing to Dr. Roche research by implementing a Convolutional Neural Network into her open source software Silbido, improving the accuracy of contort extraction by 33%.
This software and associated research was presented at DCLDE 2022 where it was well received by the aquatic mammal research community. Software like this goes on to help population estimation, localization, and density research. Further research in this field aims to apply the same techniques to mammals with similar communication methods (whales, birds etc.) or classify shapes of contort extraction and map them to dolphin behaviors, creating a rudimentary dolphin translator.
Audio Signal Processing • Deep Learning • Research • MATLAB
This project improves wildfire risk assessment by leveraging regression modeling and data analysis to improve upon the Fire Weather Index's (FWI) fire perdition accuracy.
R • Regression • Hypothesis Testing • Data Visualization
An CNN Model for identifying fruits and vegatables "too unappealing for sale" super markets, but fit for donation to food backs.
Computer Vision • Deep Learning • Fine Tuning • TensorFlow
An exploration into improving diagnosis with smaller and efficient neural network models.
PyTroch • Deep Learning • Fine Tuning
A research projected aimed at improving predictions of competitive NHL games by finding hidden relations in individual contributor data with Machine Learning and Deep Learning.
Data Acquisition and Preprocessing • Regression • Time Series Predictions
A FasterRCNN script for identifying common objects in home videos.
Computer Vision • Deep Learning • PyTorch
Page updated
Google Sites
Report abuse