RLTasksby ulam.ai
Sign inAll tasks
Machine learning research5ML-1

5ML-1

Five machine-learning research environments for adaptive diagnosis under model misspecification. They span causal mechanisms, scaling behavior, representation geometry, continual learning, and uncertainty under distribution shift, requiring costly diagnostics, compositional mechanism inference, continuous-variation tracking, and explicit forecasts under heavy-tailed, common-mode error.

ML researchExperiment designRLVR
Version2.0.0
Environments5
RewardScalar · 0–1
DeliveryPrivate ZIP
01 Task contract

A real, versioned RL environment.

Five machine-learning research environments for adaptive diagnosis under model misspecification. They span causal mechanisms, scaling behavior, representation geometry, continual learning, and uncertainty under distribution shift, requiring costly diagnostics, compositional mechanism inference, continuous-variation tracking, and explicit forecasts under heavy-tailed, common-mode error.

Agent objective

  • Interact with the supplied stateful environment.
  • Produce verifier-checkable actions or artifacts.
  • Maximise scalar reward under the package contract.

Evaluation

  • 5 verifier-backed environments.
  • Reported reward range 0–1.
  • Package-specific public and private checks.

Delivery boundary

  • Private object stored in Cloudflare R2.
  • Authenticated entitlement required.
  • Short-lived signed URL per download.
02 What you will work on

Distinct environments, one demanding research contract.

Enough detail to understand the intellectual terrain; generated instances, hidden mechanisms, and solution paths remain inside the private package.

EnvironmentMathematical or technical frontierAdaptive research problem
CausalMechanismLabCausal discovery and intervention designCombine interventions with mean and covariance diagnostics to separate direct, mediated, and confounded effects under uncertain coefficients.
ScalingLawCartographerNeural scaling laws and compute allocationChoose training configurations that separate coupled scaling mechanisms while supporting forecasts beyond the observed regime.
RepresentationXRayRepresentation geometry, shortcuts, invarianceAudit layers and counterfactual batches through compressed probes when mechanisms share rank or separation signatures.
ContinualLearningAutopsyForgetting, replay, adapters, path dependenceDesign curricula and retention/transfer measurements revealing which mechanisms act across conflicting and revisited tasks.
UncertaintyShiftDoctorCalibration, OOD detection, selective predictionPair coverage, risk, calibration, entropy, and OOD diagnostics when pipeline components cancel in aggregate metrics.
03 Why it is interesting

What the supplied evaluation reveals.

A nominal Bayesian reference was evaluated on 100 fresh expert episodes outside the development seed range. It reached 0.7550 mean reward but only 40 passes and 51 correct identifications, with pass rates ranging from 11/20 to 3/20 across domains. This makes the suite useful for studying where one-step information gain breaks under extrapolation, path dependence, and nuisance mismatch. The package passed 66/66 tests.

We publish aggregate behavior and task structure, while withholding generated instances, hidden labels, exact successful probes, private checks, and solution trajectories.

04 Supplied evaluation

Observed evaluation result.

Shown with its provenance and limitations; it is not a performance guarantee.

i
Methodology matters

Fresh 100-episode expert calibration; 40% pass rate.

Evaluated system / policyGreedy Bayesian baseline

As identified by the supplied artifact.

Mean reward0.7550

100 reported runs.

Result artifactIncluded

greedy_bayes_expert_summary.csv

Public result record

Machine-readable provenance and the exact displayed metric are available in results.json.

Open result JSON
05 Private delivery

The package stays off the public website.

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.

Included with purchase

  • Exact package version 2.0.0
  • Environment and task contracts
  • Verifier or scoring interface
  • Supplied reference/evaluation artifacts
  • Purchase record and licence v1.1
delivery flow
Authenticated buyer + entitlement check
+ private R2 object + 2–5 minute signed URL
= direct, auditable download

Package SHA-256
e439aad667706019a88761a145f2e952b2f40ce011e43077b7992581458357a9
06 Licence v1.1

Commercial use, without exclusivity.

One 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.

Read the full licenceYotta Content LTD · business customers only
5ML-1

Ready to add this environment?

Back to marketplace
Stripe checkout

Business purchase confirmation

Sign in or create an account, then complete secure Stripe Checkout. Access is granted only by the verified payment webhook.