Agent objective
- Interact with the supplied stateful environment.
- Produce verifier-checkable actions or artifacts.
- Maximise scalar reward under the package contract.
Five apex mathematics environments where 64 candidate mechanisms share exactly non-identifying seed evidence and agents get at most three costly probes. The suite spans rough paths, weighted automata, lattice geometry, replicator dynamics, and inverse conductivity, demanding noncommutative reasoning, complementary measurements, calibrated posteriors, and transfer to sealed probes.
Five apex mathematics environments where 64 candidate mechanisms share exactly non-identifying seed evidence and agents get at most three costly probes. The suite spans rough paths, weighted automata, lattice geometry, replicator dynamics, and inverse conductivity, demanding noncommutative reasoning, complementary measurements, calibrated posteriors, and transfer to sealed probes.
Enough detail to understand the intellectual terrain; generated instances, hidden mechanisms, and solution paths remain inside the private package.
| Environment | Mathematical or technical frontier | Adaptive research problem |
|---|---|---|
| RoughPathOracle | Rough paths and noncommuting flows | Combine endpoint and operator-level probes to recover higher Lie interactions hidden by identical first-level calibration. |
| AutomatonHankel | Weighted automata and rational series | Choose complementary word, Hankel, reversal, and continuation probes to expose hidden transition couplings. |
| VoronoiCipher | Lattices, Voronoi cells, nearest-vector geometry | Sample complementary hard and soft cell-boundary behavior when covolume and origin quantization agree. |
| ReplicatorResonance | Game theory and replicator dynamics | Use signed global trajectories across horizons to recover hidden strategic orientation beyond shared equilibrium spectra. |
| CalderónSketch | Inverse conductivity and graph Laplacians | Combine signed and quadratic boundary responses across sectors to infer conductivity fields after harmonic elimination. |
A public exact-likelihood complementarity planner was tested on 60 fresh expert episodes. It averaged 0.7781 and passed 21/60, versus 11/60 for the bundled greedy control. VoronoiCipher passed 12/12 while RoughPathOracle and ReplicatorResonance passed none, demonstrating distinct noncommutative and geometric bottlenecks rather than reskinned tasks. Package integrity included 47/47 tests and 107 verified checksums.
We publish aggregate behavior and task structure, while withholding generated instances, hidden labels, exact successful probes, private checks, and solution trajectories.
Shown with its provenance and limitations; it is not a performance guarantee.
Reported result from the evaluation artifact supplied with this package.
As identified by the supplied artifact.
60 reported runs.
ulam_self_score_bundle.zip:SELF_SCORE_REPORT.md
Machine-readable provenance and the exact displayed metric are available in results.json.
The paid ZIP will live in a private R2 bucket. Vercel authorizes the buyer and issues a 2–5 minute object URL; R2 serves the bytes directly.
Authenticated buyer + entitlement check
+ private R2 object + 2–5 minute signed URL
= direct, auditable download
Package SHA-256
c3b54c253e4813f538077ddf5f5a05311c2f65c82d9a33accfe586db14ca8910One purchase licenses this identified item to one legal organisation for worldwide, perpetual commercial model training, evaluation, research and development. Redistribution and resale of the package are not permitted.