Vinodchandran Variyam
University of Nebraska-Lincoln, NE·Computer Science
Most algorithms assume the whole dataset sits in memory. Modern systems often face the opposite situation: data arrives as a never-ending, high-volume stream, and the machine cannot store it all.
- Fields
Computer Science
Mathematics & Statistics
- Topics
data streams
streaming algorithms
sublinear algorithms
- Grants
- 2 known grants, through Sep 2027
What they work on
- Initiate the study of a "right to forget" data stream model, in which items can be required to be deleted, motivated by privacy concerns and the growth of data generation.
- Explore pseudodeterministic computation, where a randomized algorithm must output the same canonical answer on almost every run, in the setting of streaming algorithms.
- Investigate the Delphic set streaming model, where each stream item is a succinctly represented set, exploiting the recently found link between data streaming algorithms and model counting.
- Define and formalize new notions of algorithmic replicability across different computational contexts, and develop general methodologies for designing replicable randomized algorithms.
- Establish the costs and inherent limitations of replicability by proving lower bounds, with application to machine learning and large data analysis.
More about the research
A second modern worry is that algorithms which use randomness can give different answers on different runs, which makes results hard to reproduce. Both projects in this lab tackle the theoretical foundations of computing under such constraints, designing algorithms with provable guarantees and proving limits on what is possible.
Why it matters
Streaming algorithms are used in network monitoring, security and data mining, where memory is far smaller than the data. A "right to forget" streaming model could support privacy rules that require deleting individual records. Work on algorithmic replicability addresses a recognized problem across the sciences: randomized, data-driven methods that give different answers on repeated runs, which affects machine learning and large-scale data analysis.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Andrew McGregor
University of Massachusetts, MA·Computer Science
Develop adversarially robust data stream algorithms that stay accurate when inputs are chosen adaptively in response to the algorithm's earlier outputs, as happens in interactive data analysis.
Aleksandr Golovnev
Georgetown University, DC·Computer Science·Early-career
Find explicit constructions of rigid matrices, which must have many entries changed to lower their rank, to better understand the power and limits of logarithmic-depth circuits.
Michael Bender
Stony Brook University, NY·Computer Science
Design efficient methods to build graph-based vector search indexes such as HNSW and DiskANN, reducing their currently very high construction costs.
The school
University of Nebraska-Lincoln: 17 labs on PI Pages, $657M a year on research, 261 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.