Papers and talks
This page links my written publications, talks, presentations and software projects.
Also a little bit of advertisement, for making the occasional quick PDF I prefer TeXmacs over LaTeX. Although these days I mostly ask Claude to generate LaTeX.
Quantum computing and quantum information
-
The Syndrome Buffer Sizing Problem in Quantum Error Correction
[Manuscript in preparation] -
Fast Bounds on Matching Decoder Soft Output for the Surface Code
[Manuscript in preparation] [code] [slides Generated by Claude. Subject to scrutiny.] -
Speculative Matching for Low Latency Surface Code Quantum Error Correction
[Manuscript in preparation] [code] -
Challenges in Barren Plateau Mitigation with Dynamic Parameterized Quantum Circuits
[Accepted as a contributed talk at QTML 2026] [arxiv] [experiments] [sympauli] -
Error Mitigation in Dynamic Circuits for Hamiltonian Simulation
[Accepted in GLS VLSI 2026] [DOI] [arxiv] [code] -
Scalable Benchmarking Framework for Dynamic Quantum Circuits
[Accepted in MICRO 2026] [arxiv] [code] [docs] -
Dynamic circuit compilation for sparse qubit connectivity
[Accepted at Quantum Week 2025] [DOI] [poster] [code] -
Scalable decoders for quantum surface codes
[letter for EE599 at USC] -
Solving the subset sum problem on a quantum computer
[report for EE520 at USC] [presentation for EE520 at USC] [talk at Microsoft Azure Quantum] [hackathon project on Github] -
A gentle introduction to quantum algorithms for combinatorial and optimization problems
[reading project for CS675 at USC] -
The hidden subgroup problem and the quantum fourier transform
[reading project for CS599 at USC] [presentation for CS599 at USC]
AI/ML
-
A flaw in the NeurIPS unlearning challenge and an algorithmic framework for entropy regularization
[Accepted at NeurIPS 2023 New in ML workshop] [poster] [paper] [talk] -
Foundation models - a 2022, 2023 review
[letter] -
Emergent structures in transformer models trained by self-supervised methods
[presentation] -
Interpretability of multi-modal neural models trained with contrastive self-supervised learning
[report for CS566 at USC]