Avatar

Suman Kalyan Bera

Researcher in Computer Science

I am part of the Graph AI team at Katana Graph, a start-up founded by Prof. Keshav Pingali and Prof. Chris Rossbach in 2020. Katana Graph provides a distributed graph intelligence platform with highly scalable and lightning fast graph query, analytics, mining, and AI solutions. If you are passionate about graph algorithms/graph AI and looking for opportunities, feel free to reach out.

Before joining Katana Graph, I have spent a couple of years as a postdoctoral researcher at UC Santa Cruz where I worked with Prof. C. Seshadhri. I obtained my Ph.D. in Computer Science from Dartmouth College, advised by Prof. Amit Chakrabarti. In the past, I have worked at IBM Research Lab and Adobe India.

 

Research Interests

I am broadly interested in the topic of large-scale graph analysis. My current research involves developing scalable graph AI models for real-world applications. I also work on designing efficient and practical graph algorithms with theoretical guarantees in various Big Data models such as the streaming/distributed models. Over the past few years, I have enjoyed working on a diverse range of topics including algorithmic fairness, scalable clustering, communication complexity, arithmetic circuit complexity, and approximation algorithms. My research has been published at premier theoretical computer science conferences such as SODA, ITCS, ICALP, etc. as well as at top-tier ML/data mining conferences such as NeurIPS, KDD, VLDB, etc. For more details, see my publications.

Recent News

See all news>>

  • [Mar 2022] At Katana Graph, we have developed a simple graph neural network (gnn) based AI model SimGCN that is dominating the leaderboards at TDC ADMET Benchmark Group! Check out our report and Github repo.

  • [Feb 2022] Excited to serve on the program committee of ACM FAccT Conference 2022.

  • [Jan 2022] Two papers accepted at the Web Conference! Details to be updated soon.

  • [Nov 2021] Our paper titled Counting Subgraphs in Degenerate Graphs has been accepted for publications at the Journal of the ACM! Arxiv preprint will be posted soon.

  • [Oct 2021] Serving on the program committee of The Web Conference 2022.

  • [June 2021] Excited to join the Graph AI team at Katana Graph, a start-up that is providing a distributed scalable graph intelligence platform!

Publications

Click on each title to see the abstract and other details.

Counting Subgraphs in Degenerate Graphs
To appear in the Journal of the ACM (JACM).

Approximation Algorithms for the Partition Vertex Cover Problem
TCS 2014.

PDF

Teaching

I have served as a Teaching Assistant for the following courses.

  • Machine Learning and Statistical Data Analysis (CS74/CS174)
  • Algorithms (CS31)
  • Discrete Mathematics in Computer Science (CS30)
  • Theory of Computation (CS39)
  • Computer Architecture