Anticipation Geometry: Domain-General Trajectory Characterization with Knowledge Graph-Grounded Rewards
Anticipation Geometry defines seven scalar signals over trajectories in arbitrary state spaces. It tests whether motion, conversation, and graph traversal share a measurable geometry of convergence from uncertainty into commitment.
Paper workspace
Live draft structure
Artifacts
Markdown paper source
Anticipation Geometry is mapped as source-first until the final paper render is selected.
source-only
Editable source
Paper source exists in Comp-Core and a parallel standalone paper directory. Public PDF render pending.
Source anchors
Comp-Core/papers/anticipation-geometry/paper.md
anticipation-geometry/paper/paper.md
Comp-Core cc-anticipation Rust implementation
Method tags
Ingest intersections
Status
Paper drafted; Rust implementation exists in Comp-Core; downstream task lift remains the hard proof gate.
Key claims
01
Transition pressure is a reusable signal across domains.
02
Trajectory geometry can become a dense reward layer before final task success is known.
03
The hard remaining proof is downstream task improvement, not only scalar significance.
Public reading note
Working paper public summary safe; attach full text after final review.
Standard skeleton
What this paper must keep proving
problem
Prediction and confidence are too sparse to monitor reasoning or motion while the trajectory is still unfolding.
method
Define seven scalar signals over vector trajectories and test whether they transfer across motion, conversation, and graph paths.
implementation
Comp-Core cc-anticipation kernel, conversation embeddings, KG path reward experiments, and Rust/Python analysis.
data
Simulated/physical motion, conversation embeddings, and graph-path snapshots with hard negatives.
evaluation
Transition-pressure prediction, KG path discrimination, scalar stability, and downstream reward usefulness.
references
Anticipatory computing, geometric deep learning, CoALA, DSS graph rewards, GraphMERT.
openQuestions
The hard public proof is downstream lift when the geometry is used to improve agent or model decisions.
Checkpoints and references
Proof chain
Claim checkpoint
central-claim slot
Every central claim must point to a proof anchor or remain labeled as speculative.
Implementation checkpoint
implementation-map slot
Every method should identify the code path, harness, schema, or protocol that embodies it.
Evidence checkpoint
evidence-manifest slot
Every reported result should point to run IDs, packet IDs, data snapshots, commits, or review artifacts.
Reference checkpoint
references slot
Every external claim should resolve to a cited paper, benchmark, standard, or documented prior system.
Release checkpoint
release-gate slot
Every PDF needs a named condition before it can move from draft to citation-ready.