CLEMENTE ANTUNA
|
Clemente Antuna
- Majors: Mathematics/Statistics and Economics
- Research Mentor: Bodhisattva Sen, Ph.D.
- Project Title: Optimal Transport on Networks
“During the internship, Prof. Sen and I discussed the theory of Optimal Transport and some of its applications in different fields, including Statistics, Mathematics, and Economics. The final project, called “Optimal Transport on Network”, applies optimal transport theory to a matching problem on a network, using international trade as a motivating example. The problem presented in the project is that of transporting goods from suppliers to consumers at the minimum cost, which is a bipartite problem. It is an assignment problem because suppliers have to be matched to consumers, and it is also an optimization problem because the transportation cost has to be minimized. The project presents a possible method of solving the problem using Dijkstra’s algorithm and Kantorovich’s duality.”
|
ERICA INYOUNG CHOI
|
Erica Inyoung Choi
- Major: Computer Science & Minor: Applied Math
- Research Mentor: Tian Zheng Ph.D. & Chengliang Tang (Grad Student)
- Project Title: Breast Cancer Risk Prediction with Deep Learning Mammography-based Methods
“Sequential mammographic data provide important information to detect breast cancer. In this project, the goal is to develop a weakly supervised deep learning method to improve the accuracy and interpretability of breast cancer risk prediction.”
|
NICHOLAS REILLY GREENSPAN
|
Nicholas Reilly Greenspan
- Major: Computer Science
- Research Mentors: Liam Paninski, Ph.D., Dan Biderman & Taiga Abe (Grad Students)
- Project Title: Techniques and Tools for Animal Pose Estimation
“During the past number of weeks I have learned a lot about various techniques for animal pose estimation using convolutional deep learning models, and I have helped develop tools to train these models myself. I have helped develop a pytorch lightning implementation of a resnet model and data loading and training scripts for animal pose estimation. I have used the AWS boto3 API to build a pipeline that will extract frames from a video and create a labeling tool where users can easily label those frames using a browser GUI so that they can later be sent to a machine learning model and used as training data. I have also learned about semi-supervised learning, and utilized a reconstruction loss based on epipolar geometry to train a model using unlabeled frames efficiently extracted from a video in addition to labeled ones. Overall it has been a great experience, and I have enjoyed working closely with my mentors as they have been extremely accessible and insightful.”
|
S. CARLYLE MORGAN
|
S. Carlyle Morgan
- Major: Statistics & Minor: Economics
- Research Mentor: Tian Zheng, Ph.D.
- Project Title: Spectral Clustering for Directed Networks
“I’ve thoroughly enjoyed being a research intern this Summer. I’ve enjoyed learning more about the topic of spectral clustering. The problem of identifying community membership in a network transcends disciplines, appearing across the natural and social sciences. Thus the work that I’ve been doing this Summer feels relevant to addressing a wide variety of problems. The support I’ve received from my mentor has been invaluable in helping me understand the mechanics behind this topic, and their encouragement of my applying some of the techniques mentioned in the literature to simulated and real-world data has really helped develop my intuition. Overall, I have found this a very rewarding experience.”
|
DAIKI TAGAMI
|
Daiki Tagami
- Major: Mathematics & Statistics
- Research Mentor: Sumit Mukherjee, Ph.D.
- Project Title: Modeling Ranking Data Using Mallows Models
“This is my first time working on a theoretical research project, but my mentor is really nice, and I’m really enjoying my work. I think working on this project has definitely improved my knowledge in statistics, and I would like to thank Professor Mukherjee for mentoring me and the Department of Statistics for their support. I’m extremely excited to be part of this internship program.”
|
JOHN ZHOU
|
John Zhou
- Major: Computer Science
- Research Mentor: Liam Paninski, Ph.D.
- Project Title: Decoding Algorithms for Motor Cortical Activity
“Improvements in implantable intracortical probes allow us to eavesdrop on the activity of our neurons with greater bandwidth and clarity than ever before. With this data, we can then leverage machine learning and statistical algorithms to translate a series of recorded neural spikes into “thoughts.” I am specifically interested in decoding movement intentions from the motor cortex, which can be turned into a series of kinematic instructions for prosthetic limbs in order to restore lost motor functions.”
|
Columbia Home
