Feature

The structure your sources never made explicit.

You're starting with three textbooks on the same topic, a stack of papers that reference each other in circles, and a codebase README that assumes you already know the architecture — and none of them tells you how it all fits together. SILKLEARN reads across all of them, maps what depends on what, and shows you the structure your source material never made explicit — dependencies, contradictions, and gaps.

Why this matters

Mapping dependencies turns a pile of sources into an explicit knowledge structure.

This capability exists to make hidden knowledge structure inspectable and reusable instead of leaving it embedded in long source material.

What it does

Dependency mapping surfaces the structure your multiple sources never made explicit.

Capability 01

Researchers: find which papers assume which priors — and stop assembling your model backwards.

Capability 02

Developers: see which architecture decisions your code depends on — before you ship something that breaks one of them.

Capability 03

Students: see where your textbooks disagree before that gap distorts your reasoning.

In depth

Most sources don't fail because the information is wrong — they fail because their structure is invisible. When you're dropped into a new research area, a large codebase, or a pile of conflicting textbooks, there's no signal about what connects to what, where sources disagree, or what is simply unknown. You build a mental model by accident, filling gaps in whatever order you happen to encounter them. SILKLEARN's dependency mapping makes that structure explicit. It traces what each concept assumes you already know, finds where two sources give you different answers to the same question, names the gaps neither source covers, and builds the graph your brain was trying to construct anyway — surfaced, inspectable, and ready to build on.

How it works

Extract the prerequisite structure from your material.

  1. 01

    Ingest your sources

    Upload whatever you're trying to learn from — research papers, technical docs, textbooks, codebases, PDFs. SILKLEARN begins identifying the conceptual relationships between them.

  2. 02

    Identify dependencies

    Using structural analysis, it finds where one concept assumes another has been understood — and makes those implicit assumptions into explicit prerequisite edges.

  3. 03

    Build the graph

    The result is a directed acyclic graph (DAG) showing which concepts must come before others. The structure is extracted from your sources, not invented.

  4. 04

    You review it

    Before you build on it, you inspect the graph. You see every edge, every source reference, every place where your sources conflict, every gap left unnamed. You decide what to trust.

  5. 05

    Read in the order it implies

    Reading order is one downstream use. The approved graph can drive the exact sequence your material implies — with contradictions flagged, dependencies visible, and every claim source-anchored.

Common questions

Does this work with research papers?
Yes. Research papers are one of the best inputs for SILKLEARN — they're dense, heavily cross-referenced, and assume a lot of prior knowledge. SILKLEARN surfaces those assumptions as explicit edges so you can see what you need to read before a given paper will make sense.
What if my sources contradict each other?
That's exactly what SILKLEARN is built to surface. When two sources give conflicting accounts of the same concept, the graph flags the contradiction so you can see it — and decide which source to trust — instead of unknowingly building a mental model on inconsistent foundations.
How is this different from just asking an AI to summarize my sources?
Summarization answers questions about what's in your sources. SILKLEARN builds a knowledge graph from your sources — surfacing dependencies, contradictions, and relationships automatically. It doesn't summarize. It structures.
Does it work with unstructured sources?
Yes. SILKLEARN is designed for dense, unstructured source material — PDF papers, technical wikis, raw documentation, textbook chapters. You don't need clean formatting to start.

Related capabilities

See where knowledge structure changes how teams approach onboarding and research.

Next step

Surface the hidden structure across all your sources.