Darby Losey Email & Phone Number

Pittsburgh, Pennsylvania, United States

• I developed interpretable deep learning methods to enhance transparency of large artificial neural networks.
• Created a novel method for interpretable deep learning using influence functions and matrix approximation methods.
• Formulated and derived robust convergence bounds for the proposed method and its variants, providing theoretical accuracy guarantees underscored by detailed proofs.
• Trained large convolutional and feedforward neural networks for method validation.
• Streamlined model debugging and data cleansing workflows by implementing automation processes using explainable A.I., resulting in a substantial increase in time efficiency.
• Led a series of seminars to train fellow research engineers on the use and application of explainable A.I.

Pittsburgh, Pennsylvania, United States

• Leveraged advanced machine learning methodologies to analyze complex patterns within larger neuronal populations, uncovering rules that govern learning. Adapted these algorithms for the training of artificial neural.
• Implemented continual deep learning models using PyTorch and JAX, with a focus on biological plausibility.
• Modeled the evolution of noisy, high dimensional time series data using reinforcement learning algorithms.
• Led and participated in deep reinforcement learning journal club and machine learning journal club.
• Guest lecturer for graduate level classes, focusing on Kalman filters and dimensionality reduction methods. TA for two computational courses.
• Awarded a five-year National Science Foundation Graduate Research Fellowship.

Seattle, Washington, United States

• Implemented deep learning models for the classification of noisy time series data.
• Designed, trained, and evaluated various deep learning models such as autoencoders and convolutional neural networks for the prediction of cognitive states.
• Attained over 90% accuracy in the real-time classification of a user’s mental state by monitoring their brain’s electrical signals.
• Presented prototypes to venture capitalists, resulting in angel investment.

Seattle, Washington, United States

• Implemented software for scientific experiments and conducted data analysis.
• Developed and deployed real-time classification algorithms for high-dimensional, noisy timeseries data.
• Designed and implemented robust, high-performance software for brain-computer interface experiments.
• Executed comprehensive data analysis on large image-based datasets, utilizing statistical methods and machine learning algorithms to identify key patterns and trends.
• Awards: Honors in Computer Science, Computer Science Outstanding Undergraduate Thesis Award, Mary Gates Foundation Research Scholarship, Washington Research Foundation Innovation Fellowship, North American Computing.
• Published three academic papers and filed one patent. More details available on personal website.

Doctor Of Philosophy - Phd, Machine Learning And Neural Computation

Master Of Science - Ms, Machine Learning Research

Bachelor Of Science - Bs, Computer Science

Honors in Computer Science Computer Science Outstanding Undergraduate Thesis Award Mary Gates Foundation Research Scholarship Washington.