Dark Energy
Dark Energy — Cosmic Expansion History w(z) · Cosmology
Measure whether dark energy is a constant or evolving — reconstruct the equation of state w(z) from the expansion history.
📋 The problem
The universe's expansion is accelerating, driven by 'dark energy.' Is it a cosmological constant (w = −1) or evolving? Reconstructing the equation of state w(z) from supernovae, BAO and cosmic chronometers is the decisive test.
🧗 Why it's a grand challenge
w(z) is recovered from integrated distances — a smoothing inverse problem; systematics across probes must be jointly modeled; tiny deviations from −1 matter enormously.
🧮 Governing model
H(z)² = H₀²[Ω_m(1+z)³ + Ω_DE·e^{3∫(1+w)/(1+z)dz}]H(z)²=H₀²[Ω_m(1+z)³+Ω_DE·exp(3∫₀ᶻ(1+w(z'))/(1+z')dz')]; luminosity distance d_L(z) fits SNe Ia + BAO.
Current best: DESI + Pantheon+ w₀wₐ constraints
🧭 Possible approaches
- Gaussian-process / nonparametric w(z) reconstruction
- Joint SNe + BAO + chronometer simulation-based inference
- Model comparison: ΛCDM vs w0wa vs quintessence
🎯 Build the benchmark
Reconstruct w(z) from distance–redshift data with RMS error ≲ 0.08 and H0 within ~1 km/s/Mpc.
Metric: w_recon_err — RMS error on reconstructed w(z) (lower better)
Datasets to start from: Pantheon+ Type-Ia supernovae, DESI BAO distance measurements, Cosmic-chronometer H(z) set
🤖 Build an AI agent to solve it
An agent that ingests new cosmological datasets and updates the w(z) posterior with systematics-aware inference.
Once a benchmark exists, an AI4Science agent can iterate solutions against it — every verified solution earns PWM.
This is a frontier framing page — an open problem, not yet benchmarked or verified, unlike PWM's mature computational-imaging benchmarks.