Research
My research focus lies in statistical learning and information theory. Specifically, I am interested in applying tools from information theory to (high-dimensional) statistical inference problems, such as compressed sensing, low-rank matrix estimation, changepoints detection and communications in large user networks. I design efficient algorithms for these problems, and provide probabilistic guarantees on the tradeoff between sample complexity and estimation error.
I’ve been particularly interested in studying message passing algorithms for the inference problems mentioned above.
I review papers for International Symposium on Information Theory (ISIT) and International Symposium on Topics in Coding (ISTC).
Key words:
- Message passing algorithms: Belief Propagation and Approximate Message Passing (AMP)
- Many-user communications
- Bayesian changepoint detection
- Compressed sensing, matrix sketching
- Random graph processes (e.g. Wormald’s differential equation method)
Publications:
X. Liu and R. Venkataramanan, “Sketching Sparse Low-Rank Matrices With Near-Optimal Sample- and Time-Complexity Using Message Passing,” in IEEE Transactions on Information Theory, vol. 69, no. 9, pp. 6071-6097, Sept. 2023, doi: 10.1109/TIT.2023.3273181.
X. Liu and R. Venkataramanan, “Sketching sparse low-rank matrices with near-optimal sample- and time-complexity,” 2022 IEEE International Symposium on Information Theory (ISIT), Espoo, Finland, 2022, pp. 3138-3143, doi: 10.1109/ISIT50566.2022.9834693. (presentation slides)
In review:
X. Liu, P. Cobo and R. Venkataramanan (2024) “Many-user multiple access with random user activity”, in review, submitted version available on request. (Poster at IEEE European School of Information Theory 2023 (ESIT))
X. Liu, K. Hsieh and R. Venkataramanan (2024) “Coded many-user multiple access via Approximate Messsage Passing”, in review.
G. Arpino, X. Liu and R. Venkataramanan (2024) “Inferring Change Points in High-Dimensional Linear Regression via Approximate Message Passing”, in review. (Poster at Yale Workshop Honoring Andrew Barron: Forty Years at the Interplay of Information Theory, Probability and Statistical Learning)