This applied AI project's goal is to model stock market trends and create a decision-making bot that leverages that info for automated trading. The project features opportunities to work on and learn more about data-mining, NLP, reinforcement learning, deep learning, and multivariate time-series forecasting using non-stationary variables. Anyone with some basic coding experience is welcome to join the team! If you're more experienced, I am actively seeking subgroup leaders for NLP, reinforcement learning and deep learning components of the project.
Group Leader: Chris Endemann
Contact: @Chris Endemann (Slack), [email protected]
Slack channel: Channel is private--message me on Slack to join.
Engineering team competing on Kaggle, an open platform for data science challenges. Workshops are held every weekend. Anyone who would like to challenge and improve their skills in real-world data science problems are welcomed to join us!
Group Leader: Zhichun Huang
Contact: @Eric Huang (Slack), [email protected]
Slack channel: engr-team-kaggle
Engineering team analyzing skin images, especially for melanoma detection, segmentation and classification. We focus on segmentation using deep learning techniques this semester. Please join us if you would like to try deep learning for computer vision applications!
Group Leader: Dandi Chen
Contact: @Dandi Chen (Slack), [email protected]
Slack channel: engr-miccai2018
Engineering team applying deep learning methods to source code related tasks. We will be focusing on predicting next revision by viewing code as a 2-D image. We are open to other tasks or ideas. Practical knowledge about neural networks, code analysis and programming language is preferred. Python will be mostly used for development. Please join us if you are interested in deep learning for code!
Group Leader: Jinman Zhao
Contact: @Jinman (Slack), [email protected]
Slack channel: engr-code-learning
This engineering team is dedicated to data mining science journals (including images and figures) to distill vast data to a more simplified human understandable data-set/outline of accepted scientific facts based on a search query.
Group Leader: Andy Leicht
GitHub: scientific-journal-mining
Contact: @AL (Slack), [email protected]
Slack channel: engr-journal-mining
The makeup industry is rapidly expanding, and the market keeps getting bigger. It can be hard sometimes to find a foundation that matches one's skin color exactly, taking into factor their undertone, and skin type. Sometimes people like myself with more olive/tan skin tone need to buy two foundations to match my exact skin tone. I want to create something using computer vision technology to create better makeup that can be matched to any individuals skin tone! If you would be interested, please contact me at [email protected].
Group Leader: Renee Kar-Johnson
Slack channel: N/A
Contact: [email protected]
Study group focusing around Andrew Ng's deep learning specialization on Coursera. The course material is completed on an independent basis, but the group meetings provide a space to ask questions and discuss the course insofar as the group has progressed in it. Meetings take place every weekend; the Slack channel will contain details about the time and location.
Group Leader: Soham Dasgupta
Slack channel: study-ml
Contact: @Soham Dasgupta (Slack), [email protected]
The most exciting advancements in A.I. have been consistently inspired by biology. Therefore, it’s not too crazy to suggest that studying the brain might be the best hope we have of constraining new algorithms that come closer to approaching artificial general intelligence. To follow this premise, this study group's intent is to discuss the intersections between the fields of cognitive science, computational neuroscience, and artificial intelligence, as well as discuss what it means to achieve artificial general intelligence. We will primarily focus our learning materials on research papers, and are always looking for additional members willing to present papers throughout the semester (we are willing to suggest some options).
A few concrete subtopics we’d like to touch on throughout the semester(s)
- Origin and evolution of deep learning and reinforcement learning algorithms
- Theories and evidence regarding how the brain updates its feature maps over time
- How plastic is the neocortex at each level of the cortical hierarchy?
- Does backprop occur in the brain? What biologically plausible alternatives to backprop exist?
- How can theoretical neuroscience help reveal mechanisms/algorithms of AGI?
- What key components are lacking from AI these days that prevent us from achieving AGI (e.g. how is the brain different from a DQN algorithm)?
Experience with deep learning, reinforcement learning, and/or neuroscience is recommended. Alternatively, for a brief introduction to the field, students can consider working through this computational neuroscience coursera course.
Group Leader: Chris Endemann
Slack channel: study-neuro-agi
Contact: @Chris Endemann (Slack), [email protected]