Literature corpus / Continuously updated

The reading
behind the question.

A literature corpus assembled and maintained by a local pipeline, organised around the seven questions that make up the Dependable AI Systems agenda. Metadata is collected from OpenAlex, arXiv and Crossref; each paper is read into a structured note before it appears here.

Where the evidence is thin

Seven questions,
unequal evidence.

Each question has its own corpus. The bar shows how much of it has been read into a structured note — the gap is the honest part.

Structured notes exist

Growing

What has actually
been read.

Only papers with a completed structured note appear below. Collected-but-unread papers are counted in the totals, not listed here.

Method & limits

A pipeline, not
a literature review.

Metadata comes from OpenAlex, arXiv and Crossref. A keyword gate removes obvious off-topic results. Each remaining paper is read by a local qwen3.5 model into a fixed schema that separates what the paper states from what the model infers.

What this page does not claim. The structured notes behind this corpus are generated by a small local model and have not been verified by a human. They are not published here, and nothing on this page should be read as a finding. What is published is bibliographic metadata and counts — facts that come from the scholarly APIs, not from a model. Citation counts are OpenAlex figures and differ between sources.