Finding Similar Scroll Paintings: Using Graph Learning to Power Recommendations

A portion of a scroll painting depicting a Santhal Dance [Origin: India].

The goal is to design a graph neural network based recommendation system for scroll paintings (folk art) that can run on decentralized compute environments (such as peer-to-peer systems). Requires ability to understand and analyze machine learning problems theoretically, programming skills and performing extensive empirical analysis on a proprietary multi-modal dataset for art conservation. 

Project description

A research project is available in the Department of Management Science and Systems which involves the development of a large scale recommendation algorithm for graphs. This is an UNPAID position, and requires the incumbent to commit to at least two semesters of work (possibly more, if the project so demands). The incumbent is, eligible for independent study, research or thesis credits and is expected to gain experience in doing research. The position is available immediately.

The student is expected to be:

  • Very well-versed in programming in Python and Java with demonstrated experience in developing and debugging code beyond academic projects
  • Must have worked on / be currently engaged in research projects at the undergraduate level and be comfortable in reading research papers and implementing algorithms described in them.
  • Must have prior knowledge of graph based deep learning algorithms, including but not limited to graph neural networks and should have taken courses in them. Alternatively, they may taken courses in Neural Networks or Probabilistic Graphical Models.
  • Understand unconstrained optimization -- should have demonstrated knowledge and preferably taken prior courses in optimization.
  • Will be required to pre-process and extract features from multi-modal data (images, text, audio and video)
  • Will implement and test the graph neural network based recommendation algorithm and should be able to look into pre-existing code from other researchers if they exist.
  • Be able to interpret, visualize, analyze and plot results from empirical studies.
  • Enthusiastically participate in reading groups and brown-bag sessions and
  • Present work at conferences or journals as appropriate.

Project outcome

  • Feature Extraction, Implementation of Large Scale Recommendation System using Graph Neural Networks, Testing the Performance of the System
  • Presentation of results in a conference and journal 

Project details

Timing, eligibility and other details
Length of commitment Year-long (10-12 months)
Start time Anytime
In-person, remote, or hybrid? Hybrid Project
Level of collaboration Small group project (2-3 students)
Benefits Academic credit 
Who is eligible Seniors 

Core partners

  • Department of Computer Science
  • Data Science, School of Engineering and Applied Science 

Project mentor

Haimonti Dutta

Associate Professor

Management Science and Systems

Phone: (484) 432-1484

Email: haimonti@buffalo.edu

Start the project

  1. Email the project mentor using the contact information above to express your interest and get approval to work on the project. (Here are helpful tips on how to contact a project mentor.)
  2. After you receive approval from the mentor to start this project, click the button to start the digital badge. (Learn more about ELN's digital badge options.) 
Start the Digital Badge

Preparation activities

Once you begin the digital badge series, you will have access to all the necessary activities and instructions. Your mentor has indicated they would like you to also complete the specific preparation activities below. Please reference this when you get to Step 2 of the Preparation Phase. 

  • Demonstrated knowledge in Machine Learning, Data Science, Graph Theory, Deep Neural Networks and Probabilistic Graphical Models
  • Demonstrated ability to code in Java and Python
  • Understand unconstrained optimization -- should have taken prior courses in optimization. 

Keywords

Graph Neural Networks, Deep Neural Networks, Machine Learning, Data Science, Computer Science, Management Science and Systems