# Samuel Talkington

> Incoming Assistant Professor of Electrical and Computer Engineering at the University of Michigan (Fall 2027), where he will lead the Eigenergy Group. Completed his Ph.D. at Georgia Tech, advised by Daniel K. Molzahn, and is spending the intervening year as a Postdoctoral Fellow at Harvard University, mentored by Le Xie and Na Li. Research develops efficient algorithms for societal-scale problems in electric power systems, treating randomness as a computational resource. Supported by the NSF Graduate Research Fellowship.

## Research Areas

- Randomized algorithms for power network optimization
- Optimal power flow and network reconfiguration
- Spectral graph theory applied to electric grids
- Randomized numerical linear algebra
- Differentiable optimization for power systems
- Energy affordability and energy justice
- Network tomography and topology learning

## Selected Papers

- [The Limits of Quantum Computers for Power Flow](https://arxiv.org/abs/2607.19263): Realistic grid structure precludes end-to-end quantum advantage for power flow; proofs formally verified in Lean 4.
- [Interactive Optimal Power Flow Compiled to the Browser](https://samueltalkington.com/papers/2026/talkington_tellegen/): The tellegen framework: exact power flow and OPF solvers compiled to WebAssembly.
- [Making Every Bit Count for A-Optimal State Estimation](https://arxiv.org/abs/2604.01211): Bit allocation across quantized sensors for LMMSE state estimation. IEEE CDC 2026.
- [Error Bounds for Radial Network Topology Learning from Quantized Measurements](https://samueltalkington.com/papers/2025/talkington_quantization/): Bounds on radial grid reconstruction error from quantized sensor data. IEEE Transactions on Power Systems.
- [Efficient Network Reconfiguration by Randomized Switching](https://arxiv.org/abs/2510.24458): Randomized switching policies that bypass combinatorial complexity in network reconfiguration.
- [Admittance Matrix Concentration Inequalities for Understanding Uncertain Power Networks](https://arxiv.org/abs/2510.17798): Concentration inequalities quantifying uncertainty impacts on linearized power flow models.
- [Differentiating Through Power Flow Solutions for Admittance and Topology Control](https://arxiv.org/abs/2510.17071): Implicit differentiation of power flow solutions for admittance and topology control.
- [Locational Marginal Burden: Quantifying the Equity of Optimal Power Flow Solutions](https://arxiv.org/abs/2405.12219): Introduced locational marginal burden linking OPF pricing with energy equity. ACM e-Energy 2024.
- [Strategic Electric Distribution Network Sensing via Spectral Bandits](https://arxiv.org/abs/2410.21270): Spectral bandit algorithms for smart meter polling under communication limits. IEEE CDC 2024.

## Software

- [Software page](https://samueltalkington.com/software/): Libraries and interactive demos from the Eigenergy Group.
- [tellegen](https://tellegen.dev): Interactive optimal power flow in the browser; Rust solvers compiled to WebAssembly with live KKT sensitivities.
- [PowerIO](https://eigenergy.github.io/powerio/): Fast Rust parser and converter for power system case files, with Python, Julia, and C bindings.
- [Marguerite.jl](https://samueltalkington.com/research/marguerite/): Minimal, performant, and differentiable Frank-Wolfe solver for constrained convex optimization in Julia.
- [PowerDiff.jl](https://samueltalkington.com/research/powerdiff/): Differentiable power system analysis; sensitivities of power flow and OPF solutions.

## Blog

- [Linearizing the Power Flow Equations](https://samueltalkington.com/blog/2025/linear_acpf/): Overview of linearizing the AC power flow equations.
- [Differentiating Solutions of DC Optimal Power Flow](https://samueltalkington.com/blog/2025/diff_dcopf/): How to differentiate DC OPF solutions with respect to problem parameters.
- [Writing DC Optimal Power Flow in Matrix Form](https://samueltalkington.com/blog/2024/matrix_dcopf/): DC optimal power flow in matrix notation.
- [Generating Random Graphs and Laplacians in Julia](https://samueltalkington.com/blog/2025/julia_laplacians/): Building random graphs and Laplacians in Julia with sparse matrix tricks.
- [PowerIO: A Fast Parser and Converter for Power System Case Files](https://samueltalkington.com/blog/2026/powerio/): Announcing PowerIO and its hub-and-spoke design.
- [Phase Retrieval in Power Systems](https://samueltalkington.com/blog/2023/power_phase_retrieval/): Phase retrieval and conditions for estimating complex injections from voltage magnitudes.

## Teaching

- [ECE 4320: Power System Analysis](https://samueltalkington.com/teaching/ece4320/) — Georgia Tech, Spring 2025
- [ECE 2020: Digital System Design](https://samueltalkington.com/teaching/ece2020/) — Georgia Tech, Fall 2024

## Optional

- [Eigenergy Group](https://samueltalkington.com/group/): Research group at Michigan ECE; recruiting PhD students for Fall 2027.
- [All Papers](https://samueltalkington.com/papers/): Complete list of publications.
- [Talks](https://samueltalkington.com/talks/): Slides and posters from conference presentations.
- [Interactive Demo: Optimal Power Flow in the Browser](https://tellegen.dev): Edit demands and line ratings on a live grid map.
- [Interactive Demo: Randomized Switching](https://samueltalkington.com/research/rand_switch/): Draw a power network and optimize switching probabilities.
- [Interactive Demo: Locational Marginal Burden](https://samueltalkington.com/research/lmb/): Explore how transmission congestion creates energy affordability disparities.
- [Full site content for LLMs](https://samueltalkington.com/llms-full.txt)
- [CV (PDF)](https://samueltalkington.com/talkington_cv.pdf)
- ORCID: [0000-0001-5768-8115](https://orcid.org/0000-0001-5768-8115)
- Google Scholar: [Samuel Talkington](https://scholar.google.com/citations?user=uOl2_HYAAAAJ&hl=en)