cd /news/machine-learning/a-transformer-derived-iterative-prec… · home › topics › machine-learning › article
[ARTICLE · art-171617] src=arxiv.org ↗ pub= topic=machine-learning verified=true sentiment=↑ positive

A Transformer-Derived Iterative Preconditioner

A paper submitted to arXiv on 17 November 2026 shows that a softmax-attention transformer trained to solve linear systems Xw=y implicitly improves matrix conditioning, and tracing XX^T revealed an iterative preconditioning primitive the authors developed into RAPID, a standalone algorithm with no learned components. RAPID is an anytime preconditioner for symmetric positive definite matrices that uses randomized mixing and sparse local corrections; proof-of-concept experiments on synthetic and real-world problems consistently improved conditioning and reduced the conjugate gradient iterations needed to reach a target accuracy. For a Haar-mixing idealization, the authors prove convergence to any fixed target condition number regardless of spectral distribution, with only logarithmic dependence on initial conditioning.

key takeaway

A transformer trained on in-context linear-system prediction yielded RAPID, an anytime SPD preconditioner whose Haar-mixing idealization provably reaches any fixed target condition number with only logarithmic dependence on initial conditioning and cut conjugate gradient iterations in proof-of-conce

read2 min views7 publishedNov 18, 2026
A Transformer-Derived Iterative Preconditioner
Image: source
  [Submitted on 17 Nov 2026]


[View PDF](https://arxiv.org/pdf/2611.22941)

[HTML (experimental)](https://arxiv.org/html/2611.22941v1)

Abstract:Transformers trained on in-context prediction problems can learn known numerical algorithms, e.g., gradient descent. We go further, showing that their internal computations can reveal new algorithms with provable guarantees. A softmax-attention transformer trained to solve linear systems $Xw=y$ implicitly improves matrix conditioning. Tracing $XX^\top$ reveals an iterative preconditioning primitive, which we develop into RAPID, a standalone algorithm requiring no learned components.

RAPID is an anytime preconditioner for symmetric positive definite (SPD) matrices, using randomized mixing and sparse local corrections to improve conditioning. Proof-of-concept experiments on synthetic and real-world problems consistently improve conditioning and reduce conjugate gradient iterations needed to reach a target accuracy. For a Haar-mixing idealization, we prove convergence to any fixed target condition number regardless of spectral distribution, with only logarithmic dependence on initial conditioning. Together, these results show how a transformer can discover a broadly applicable numerical algorithm.

Additional Features

References & Citations

...

Bibliographic Explorer

(What is the Explorer?) Connected Papers

(What is Connected Papers?) Litmaps

(What is Litmaps?) scite Smart Citations

(What are Smart Citations?) alphaXiv

(What is alphaXiv?) CatalyzeX Code Finder for Papers

(What is CatalyzeX?) DagsHub

(What is DagsHub?) Gotit.pub

(What is GotitPub?) Hugging Face

(What is Huggingface?) ScienceCast

(What is ScienceCast?) Influence Flower

(What are Influence Flowers?) CORE Recommender

(What is CORE?) IArxiv Recommender

(What is IArxiv?) arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

── more in #machine-learning 4 stories · sorted by recency
── more on @arxiv 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/a-transformer-derive…] indexed:0 read:2min 2026-11-18 · —