FIRE Capital One Machine Learning is a part of the FIRE: First-year Innovation & Research Experience Discovery-Based Learning Pilot Program that provides a CURE: Course-based Undergraduate Research Experience that immerses first-year (new freshman, freshman connection, transfer) undergraduate students in state-of-the-art research and faculty-led mentorship experience in machine learning, deep learning, and artificial intelligence. We focus on projects using recently developed techniques, perspectives, and applications for market-relevant areas such as computer vision, natural language processing, and data analytics.
Our recent 2019 machine learning projects include:
- Instance-level Object Localization and Segmentation of Camera, LIDAR, and RADAR Images from Self-Driving Cars
- Instance-level Object Recognition and Tracking for Video Surveillance and Self-Driving Cars
- Deep Image Caption Generation for Visually Impaired Users
- Extreme Learned Data Compression for Images and Videos
The FIRE Capital One Machine Learning Innovation & Research Stream Experience provides a 2-semester course sequence that spans over a full year, with the first day of the course starting at the Spring semester each year and the last day ending with the Fall semester. Students from all degree majors enroll in our FIRE171/271 course during the Spring and Fall semesters after completing their FIRE Semester 1 (FIRE120) course in the Fall semester.
Other than 1 hour of scheduled class meetings per week, the FIRE Capital One Machine Learning Stream Experience requires each student to commit 6 additional hours of independent and team work each week, most of which will be in the FIRE Capital One Machine Learning lab.
Each student throughout the semester will work both individually and as a team member of a faculty-led machine learning project. Research sessions in the lab setting will focus on faculty-led research, including collaboration with peers, communication of ideas, troubleshooting unexpected outcomes, as well as giving students relevant experiences that seek to build resiliency and critical analysis skills.
Scheduled class meetings will focus on training in current discipline-specific methods and practices, discussion of primary literature, troubleshooting research issues, and continual review of individual and group research progress.
Upon completion of this course, students will be able to:
- Demonstrate knowledge of fundamental concepts and ideas in neural networks and deep learning.
- Demonstrate the capacity to drive machine learning projects.
- Demonstrate proficiency with machine learning tools, programming, and implementation.
- Understand how to preprocess the data for machine learning.
- Demonstrate capacities to work with and manage large datasets.
- Understand how to build neural networks for applications such as computer vision, natural language processing, or data analytics.
- Understand how to fine tune and optimize the model for real-world applications.
- Critically evaluate findings on the implementation results.
- Communicate ideas and research findings through academic writing, data visualizations, and presentations.
During the mid-point of the first-semester (FIRE semester 2) of the FIRE Capital One Machine Learning Innovation & Research Stream Experience, students can apply for the competitive opportunity to become a FIRE summer fellow. As a summer fellow, each students will receive a moderate stipend to take part in an 8-week immersive research experience and commit 12-18 hours per week in the FIRE Capital One Machine Learning lab during the summer period. Each summer fellow will truly be a member of a working research group led by the Research Educator. The summer fellowship program offers an excellent opportunity to rapidly advance the student's research, professional, and leadership skills.
|Week||Date||Online Modules||Assignments||Class Topics|
|1||1/30/2019||Introduction to FIRE Capital One Machine Learning|
|2||2/6/2019||OMC1 Introduction to Machine Learning|
OMC2 Introduction to deep learning
|ASN1 Introduction to Python||Supervised Learning & Linear Regression|
|3||2/13/2019||OMC3 Linear Regression with Multiple Variables|
OMC4 Logistic Regression
|ASN2 Supervised Learning with scikit-learn||Neural Network Basics|
|4||2/20/2019||OMC5 Neural Networks Basics|
OMC6 Shallow neural networks
|ASN3 Deep Learning in Python||Deep Learning Training Techniques|
|5||2/27/2019||OMC7 DeepNeural Networks|
OMC8 Foundations of Convolutional Neural Networks
|ASN4 Data Classification Challenge||Convolutional Neural Network Basics|
|6||3/6/2019||OMC9 Practical aspects of Deep Learning|
OMC10 Deep convolutional models: case studies
|ASN5 Convolutional Neural Networks for Image Processing||Deep Convolutional Neural Networks|
|7||3/13/2019||OMC11 Object detection|
OMC12 Special applications: Face recognition & Neural style transfer
|ASN6 Image Segmentation Challenge|
ASN7 In-Lab Hours I
ASN8 Team Meetings I
|Object Segmentation & Similarity Learning|
|8||3/20/2019||NO CLASS (SPRING BREAK)|
|9||3/27/2019||OMC13 ML Strategy (1)|
OMC14 ML Strategy (2)
|Instance Segmentation & Similarity Learning|
|10||4/3/2019||OMC15 Optimization algorithms|
OMC16 Hyperparameter tuning, Batch Normalization and Programming Frameworks
|ASN9 Team Project Proposal||Similarity Learning & Re-Identification|
|11||4/10/2019||OMC17 Recurrent Neural Networks|
OMC18 Natural Language Processing & Word Embeddings
|ASN10 Team Project Setups||Recurrent Neural Network Basics|
|12||4/17/2019||ASN11 Image Segmentation Challenge V2||Long Short Term Memory & Word Vector Representation Basics|
|13||4/24/2019||Word Embedding Techniques|
|14||5/1/2019||ASN12 Team Project Milestone Paper|
ASN13 Team Project Code Presentation
|Team Project Code Presentations|
|15||5/8/2019||ASN14 Peer Review|
ASN15 In-Lab Hours II
ASN16 Team Meetings II
ASN17 FIRE Survey
|In-Class Competition: Dog vs Cat Image Classification|
|16||5/15/2019||NO CLASS (END OF COURSE)|
We sincerely thank the following sponsors for providing us with the funds, services, and tools to help FIRE Capital One Machine Learning provide a better innovation & research experience for our students.