aurora-mix-v1 — cascade challenger generator

A twelve-family synthetic time-series prior built to out-train the genesis king (tempopfn-base-mix-v1). numpy + scipy only — no torch, no vendored package, one self-contained generator.py.

Study the live king first (cascade fetch)

Every committed generator is public and content-addressed. Pull the reigning king and diff your prior against the actual code that is winning right now (the current testnet king is a GP-based smoothgp, so the kernel_gp family below competes it head-on):

cascade fetch king --network test --chain-toml chain.testnet.toml   # → ./fetched-king-uidN
python local_validator/validate.py --repo my_generator --king ./fetched-king-uidN

--king accepts any generator dir, so the compare stage scores your feature-space coverage against the live king, not just the shipped base_generator. You win by improving on the visible best — a byte-identical copy is dropped before it trains, so you must genuinely beat it.

Strategy vs the king

The king mixes ten TempoPFN families. This generator:

1. Keeps the highest-signal families (per the TempoPFN ablation), with its own implementations: compositional GP/kernel priors (KernelSynth-style, coarse-grid Cholesky + cubic upsampling) and rich trend × multi-seasonality × structured-noise composition.

2. Adds seven regime classes the king does not emit:

   family	real-world shape it teaches	literature
   chaotic (Lorenz/Rössler/Duffing/Mackey-Glass/logistic/Hénon)	nonlinear dynamics, traffic/weather transfer	DynaMix (arXiv 2505.13192)
   fgn (fractional Gaussian noise / fBm)	long-range dependence	classic Hurst
   regime_garch (Markov-switching AR + GARCH)	volatility clustering, regime shifts	econ/finance
   intermittent (zero-inflated counts)	retail demand, sparse events	Croston
   calendar (daily profile × weekly factors × holidays)	energy/traffic load curves	—
   growth (logistic/Gompertz/lifecycle)	adoption curves, saturation	Bass
   bursts_anomaly (subcritical Hawkes + anomaly-injected AR)	bursty events, robustness to outliers/level shifts	—

3. TSMixup-style mixup family: convex combinations across families (Chronos showed mixing improves zero-shot generalisation).

4. Scale/offset/quantisation diversity in post-processing: log-uniform scales over five decades, offsets, integer quantisation (count-like series), softplus positivity — the from-scratch model must learn scale-robustness from the corpus alone.

5. Length mixture biased long (55 % of crops in [768, 2048]) — more signal per series while keeping short-series coverage.

Determinism

The corpus is a pure function of (seed, n_series). Every RNG is a np.random.default_rng(SeedSequence([seed, tag, index])); there is no global RNG, no torch, no wall-clock, no hash(). Verified by cascade verify (two full draws, digest-compared).

Verify / test

# from the cascade repo root
python local_validator/validate.py --repo my_generator          # full local gauntlet
python -m pytest my_generator/tests -q                          # contract tests
cascade verify ./my_generator                                   # the real pre-deploy check

Deploy (when ready)

cascade deploy ./my_generator --hub-repo <namespace/aurora-mix> \
    --wallet-name <coldkey> --wallet-hotkey <hotkey>

Remember: one hotkey = one submission, for life. Deploy only after the local gauntlet and (ideally) a rented-GPU A/B train against the current king's public generator.