Datasets:
Formats:
json
Size:
< 1K
Tags:
ai-agent
agent-knowledge-cycle
knowledge-cycle
self-improvement
cognitive-economy
signal-first
License:
Upload graph.jsonld with huggingface_hub
Browse files- graph.jsonld +13 -0
graph.jsonld
CHANGED
|
@@ -53,6 +53,7 @@
|
|
| 53 |
"https://doi.org/10.5281/zenodo.19212118",
|
| 54 |
"https://github.com/shimo4228/agent-knowledge-cycle/tree/main/examples/minimal_harness"
|
| 55 |
],
|
|
|
|
| 56 |
"definesConcept": [
|
| 57 |
"https://shimo4228.github.io/shimo4228/vocab#concept/six-phase-loop",
|
| 58 |
"https://shimo4228.github.io/shimo4228/vocab#concept/three-layer-structure",
|
|
@@ -610,6 +611,18 @@
|
|
| 610 |
"url": "https://github.com/shimo4228/shimo4228"
|
| 611 |
},
|
| 612 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 613 |
{
|
| 614 |
"@id": "https://shimo4228.github.io/shimo4228/vocab#prior-art/coala",
|
| 615 |
"@type": "ScholarlyArticle",
|
|
|
|
| 53 |
"https://doi.org/10.5281/zenodo.19212118",
|
| 54 |
"https://github.com/shimo4228/agent-knowledge-cycle/tree/main/examples/minimal_harness"
|
| 55 |
],
|
| 56 |
+
"subjectOf": "https://doi.org/10.5281/zenodo.20578272",
|
| 57 |
"definesConcept": [
|
| 58 |
"https://shimo4228.github.io/shimo4228/vocab#concept/six-phase-loop",
|
| 59 |
"https://shimo4228.github.io/shimo4228/vocab#concept/three-layer-structure",
|
|
|
|
| 611 |
"url": "https://github.com/shimo4228/shimo4228"
|
| 612 |
},
|
| 613 |
|
| 614 |
+
{
|
| 615 |
+
"@id": "https://doi.org/10.5281/zenodo.20578272",
|
| 616 |
+
"@type": "ScholarlyArticle",
|
| 617 |
+
"name": "Harness Alignment and Harness Drift: Why Intent, Unlike Correctness, Resists Automation",
|
| 618 |
+
"author": "https://orcid.org/0009-0002-6168-4162",
|
| 619 |
+
"datePublished": "2026-06-07",
|
| 620 |
+
"identifier": "10.5281/zenodo.20578272",
|
| 621 |
+
"about": "https://doi.org/10.5281/zenodo.19200726",
|
| 622 |
+
"isBasedOn": "https://github.com/shimo4228/agent-knowledge-cycle",
|
| 623 |
+
"description": "Position paper (Zenodo working paper, v1) deposited from the AKC line. Defines harness alignment — the continuous, human-gated activity of keeping an agent's harness aligned with the operator's evolving intent — and harness drift, its failure mode, against the software-evolution and alignment literatures; argues the three defining properties (continuous, human-gated, bidirectional) follow from a single root: intent, unlike correctness, cannot be automated the same way. Records a bibliographic bridge: audited 2026 drift coinages are severed from the classical software-evolution lineage, which harness drift reconnects by reference. Two-layer design: lean body for human readers, verified-verbatim footnotes as a density layer for LLM consumption. Scoped as provisional judgments from a months-old practice, offered as a position, not an empirical study."
|
| 624 |
+
},
|
| 625 |
+
|
| 626 |
{
|
| 627 |
"@id": "https://shimo4228.github.io/shimo4228/vocab#prior-art/coala",
|
| 628 |
"@type": "ScholarlyArticle",
|