Projects
Brain Tumor Detection
Multi-Cognitive Domain Training
Offer-Up Redesign
Summary:
Developed a logistic regression model and a multi-layer perceptron classifier network to classify MRI images with the presence of a brain tumor and those without.
Summary:
Synthesized and facilitated a class discussion on the outcomes from the paper Structural Brain Changes after Traditional and Robot-Assisted Multi-Domain Cognitive Training in Community-Dwelling Healthy Elderly (Kim Et. Al., 2018).
Summary:
Collaborated with classmates to review and redesign the Offer-Up mobile app.
Skills:
- Pre-processed and recoded images of MRI scans using singular value decomposition
- Utilized the grid search method to find the best combination of hyperparameters
- Evaluated model quality using k-fold cross-validation
- Julia, Python
- Peer collaboration
Skills:
- Analyzed and critiqued a cognitive science research paper
- Facilitated a class discussion comparing the paper and theories of memory covered in the course
Skills:
- Worked with classmates to create a research plan
- Developed a user persona and Heuristics Evaluation in according with common UX Principles
- Protoyped an alternate design using Figma
Instagram Redesign
Human Action Recognition
Endogenous Attention
Summary:
Developed a diagnosis of Instagram based on the theories and method discussed in one of All Tech is Human's Design & Techology Ethics mentorship pod.
Summary:
Developed an unsupervised learning model from the ground up (without any machine learning models) to classify whether the model is familiar or not with the action depicted in an image.
Summary:
Designed an asynchronous presenation to explain the concept of exogenous attention to peers.
Skills:
- Analyzed critical works by technology ethicists such as W. Brian Arthur, Albert Borgmann, Paul Nightingale, Ruth Scwartz Cowan, Ursula M. Franklin, Langdon Winner, Sharon Bardzell, Laura Forlano, Helen Nissenbaum
- Evaluated the question "How to responsibly engage technologies as designers" applied to the context of Instagram
Skills:
- Preprocessed data and recoded images for model processing
- Initialized model's weights using a separately trained supervised learning model
- Tuned the model by varying model's parameter's such as the regularization term, train-test-split and other factors
Skills:
- Reviewed academic journal articles on attention from a multidisciplinary lens
- Distilled complex scientific information in an engaging and comprehensible manner to a general audience
Brain Tumor Detection
Summary:
Developed a logistic regression model and a multi-layer perceptron classifier network to classify MRI images with the presence of a brain tumor and those without.
Skills:
- Pre-processed and recoding images of MRI scans using singular value decomposition
- Utilized the grid search method to find the best combination of hyperparameters
- Evaluated model quality using k-fold cross-validation
- Julia, Python
- Peer collaboration
Multi-Cognitive Domain Training
Summary:
Synthesized and facilitated a class discussion on the outcomes from the paper Structural Brain Changes after Traditional and Robot-Assisted Multi-Domain Cognitive Training in Community-Dwelling Healthy Elderly (Kim Et. Al., 2018).
Skills:
- Analyzing and critiquing a cognitive science research paper
- Facilitating a class discussion comparing the paper and theories of memory covered in the course
Offer-Up Redesign
Summary:
Collaborated with classmates to review and redesign the Offer-Up mobile app.
Skills:
- Worked with classmates to create a research plan
- Developed a user persona and Heuristics Evaluation in according with common UX Principles
- Protoyped an alternate design using Figma
Instagram Redesign
Summary:
Developed a diagnosis of Instagram based on the theories and method discussed in one of All Tech is Human's Design & Techology Ethics mentorship pod.
Skills:
- Analyzed critical works by technology ethicists such as W. Brian Arthur, Albert Borgmann, Paul Nightingale, Ruth Scwartz Cowan, Ursula M. Franklin, Langdon Winner, Sharon Bardzell, Laura Forlano, Helen Nissenbaum
- Evaluated the question "How to responsibly engage technologies as designers" applied to the context of Instagram
Human Action Recognition
Summary:
Developed an unsupervised learning model from the ground up (without any machine learning models) to classify whether the model is familiar or not with the action depicted in an image.
Skills:
- Preprocessed data and recoding images for model processing
- Initialized model's weights using a separately trained supervised learning model
- Tuned the model by varying model's parameter's such as the regularization term, train-test-split and other factors
Endogenous Attention
Summary:
Designed an asynchronous presenation to explain the concept of exogenous attention to peers.
Skills:
- Reviewed academic journal articles on attention from a multidisciplinary lens
- Distilled complex scientific information in an engaging and comprehensible manner to a general audience