Method and source policy
The atlas separates source discovery from publication. Automated scanning only creates review candidates; no claim is published without human approval.
Evidence status
Demonstrated
A paper or technical report with a method and observable result; this does not imply independent reproduction.
Reported
A capability or result stated by an organization without enough published method.
Inferred
A bounded editorial inference from cited evidence; rationale is required.
Editorial synthesis
A field interpretation combining multiple sources; methodology is required.
Verification state
Current, stale, needs review, and withdrawn describe when and how a source was checked; they are separate from evidentiary strength.
Unknown values
When published evidence is absent, a value is not treated as zero or failure; it remains unknown. Unlike benchmarks are never collapsed into one maturity score.
Freshness
Latest source check: Oct 7, 2026. Oldest source check: Sep 21, 2026. A source access failure preserves the last approved content; changed sources enter review instead of publishing silently.
Catalog coverage
13 concepts, 15 models, 22 primary sources, and 9 approved signals. The latest check date does not mean every claim is current or independently reproduced. Read each record’s evidence and verification labels.
Unverified publication and event dates are shown as unknown; undated records appear at the end of the timeline. Availability reflects the cited source; it is not an API test or a guarantee of access.
Download approved research data (JSON)WFM within aserdargun.com
WFM is a research atlas in the physical AI layer of the AI Learning System. Read the sources here, explore prediction and action in WML, then follow asset records, evidence, and human oversight in ITL.
These links describe a learning path. The applications run independently; they do not imply shared telemetry or automatic control. ITL examples are synthetic teaching records.