Reproduce “Agentic Autoresearch for CT Reconstruction” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Andreas Maier, Lucas Kachelriess, Siming Bayer, Yixing Huang, Yan Xia, Amber Simpson, Moritz Zaiss
The authors present an LLM agent that benchmarks CT-reconstruction solvers by editing established methods, running short jobs with a shared differentiable projector, and applying one frozen score. Under mild test noise, a learned primal-dual method rose from hr 0.7233 to 0.9349, while the clean champion fell from 0.8948 to 0.0000; retraining partly restored rankings in this breast task.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.