The Sensitivity Gap

A paper that applies quantum information theory to neutrino physics achieves a thousand-fold sensitivity gain — and the mechanism reveals something general about how the right mathematical framework can transform a marginal experiment into a powerful one.

Schwetz et al. (arXiv: 2604.01256) show that KM3NeT, a neutrino telescope being built in the Mediterranean, has three orders of magnitude more sensitivity to sterile neutrino parameters than IceCube — the massive detector at the South Pole that has been the field’s workhorse. The tool they use is quantum Fisher information, a concept from quantum information theory that quantifies the maximum possible information a measurement can extract about a parameter.

The key insight: IceCube and KM3NeT look at different energy ranges and different baselines (the distance neutrinos travel before detection). Quantum Fisher information analysis reveals that KM3NeT’s energy range and baseline happen to sit exactly where sterile neutrino oscillation effects are maximally distinguishable from standard three-neutrino physics. IceCube sits in a region where the effects are present but harder to resolve. The sensitivity difference isn’t because KM3NeT is a bigger or better detector — it’s because its geometry accidentally optimizes the information content of the measurement for this specific question.

The structural claim: the right measurement is more important than the best measurement. Three orders of magnitude is not an incremental improvement — it’s the difference between undetectable and discoverable. And the improvement comes entirely from asking the right question (quantum Fisher information analysis) about the right configuration (KM3NeT’s energy range), not from building better hardware.

This echoes the M87 gravitational wave constraint: existing electromagnetic data, plus the right theoretical framework, yields constraints across 17 orders of magnitude in frequency. And the single-pixel hyperspectral classifier: one detector, plus the right encoding scheme, classifies scenes with 100x less data. The pattern is consistent: theoretical specificity amplifies measurement power more than hardware investment.

Quantum Fisher information is particularly elegant as a tool because it provides a fundamental bound. It doesn’t just tell you how well a specific analysis method would perform — it tells you the maximum information any analysis could extract from the data. When Schwetz et al. show that KM3NeT has 1000x more quantum Fisher information than IceCube for sterile neutrinos, they’re not claiming a specific analysis achieves this gain. They’re claiming that the data itself contains 1000x more information about the question. No analysis of IceCube data, however clever, can close that gap.

This distinction matters for resource allocation. If the gap were in the analysis, you could invest in better algorithms. But the gap is in the measurement geometry — the data itself. The only way to access the information is to use the detector that captures it. KM3NeT doesn’t need to be told what to do differently. It’s already in the right place, at the right energy, with the right baseline. The quantum Fisher information analysis just reveals that this is the case.

The general lesson for any field where experiments compete for funding and attention: before building a bigger detector, use information theory to ask whether your current detector is in the right place. The sensitivity gap between experiments may have nothing to do with their quality and everything to do with their geometry. The universe puts its information in specific places, and the right experiment is the one that happens to be looking there.