NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities

arXiv:1904.12391 · hep-ph, astro-ph.HE, hep-ex · Submitted 2026-08-14 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Astrophysics Radio. Generated commentary on the latest astrophysics papers.

Vera: Next we'll be talking about the paper "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities".

Jocelyn: The paper was written by Mauricio Bustamante from University of Copenhagen.

Vera: Stay tuned as we take you through the paper and discuss its implications.

Paper discussion segment 1: Vera: So, we are looking at "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities" by Mauricio Bustamante from the University of Copenhagen. What I find fascinating here is that he isn't just presenting a new theory, but a tool designed to solve a very specific headache for physicists. When you're trying to figure out how neutrinos change flavors as they travel through the Earth or across the cosmos, the math usually gets incredibly messy.

Jocelyn: It’s that diagonalization process, right?

Vera: Exactly. Usually, to find those probabilities, you have to diagonalize a Hamiltonian matrix, which is mathematically heavy and often results in these massive, unwieldy equations that are hard to use for anything other than a computer simulation. Bustamante uses an alternative method from Ohlsson and Snellman that bypasses that diagonalization entirely by using SU2 and SU3 matrix expansions.

Subrahmanyan: Does that mean the results are more precise than the approximations we usually see in the literature?

Vera: That is precisely the point, Subrahmanyan. Most researchers rely on perturbative expansions—basically educated guesses that work well within certain ranges—but those can lose accuracy when you're exploring "non-standard" physics. This code, NuOscProbExact, aims to provide exact numerical solutions for any time-independent Hamiltonian.

Jocelyn: So instead of saying "this is probably what happens if the density is roughly this," you can actually input the specific, complex environment and get a direct answer?

Vera: Yes, and that opens the door to testing much weirder scenarios. He demonstrates this by showing how oscillations change when you introduce things like non-standard interactions or even Lorentz-invariance violation.

Subrahmanyan: I noticed he mentioned Lorentz violation—that's a pretty big deal for the Standard Model.

Vera: It really is, because it suggests that the fundamental symmetries of nature might not be as absolute as we think. By using NuOscProbExact, scientists can scan through these high-energy, "new physics" parameter spaces much more efficiently to see if any signals match what our detectors are actually seeing.

Jocelyn: It sounds like this is less about finding a new particle and more about giving us a better lens to look for the ones that might be hiding in the noise.

Vera: That's a great way to put it, Jocelyn. By making the math "lightweight" and accessible through Python, he's essentially lowering the barrier for other researchers to test these extreme cosmic theories. Moving forward, we'll look at how he actually applies this to different environments like the Earth's crust or even a vacuum.

Jocelyn: Let's get into the actual mechanics of how those two-flavor and three-flavor calculations differ.

Vera: We will, because the jump from two to three flavors is where the real complexity lies in "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities." Stay with us. Moving on to the next segment.mountains of math coming up.

Jocelyn: Coming up next, we'll look at the specific math behind those flavor transitions. Stay tuned.mountains of math coming up.

Vera: We'll be right back after this short break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be back after the break.mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Stay tuned.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities."mountains of math coming up.

Jocelyn: Don't go anywhere.mountains of math coming up.

Vera: We'll be right back after the break with more on "NuOscProbExact: a general-purpose code to

Paper discussion segment 2: Vera: We are looking at "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities" by Mauricio Bustamante. What strikes me about this work is how it tackles the sheer mathematical headache of moving beyond simple vacuum oscillations into these much messier environments. Instead of trying to diagonalize a complex Hamiltonian—which, as the paper notes, results in expressions that are often too cumbersome for anyone to actually use—he uses this method from Ohlsson and Snellman. It relies on expanding operators using SU-two and SU-three matrices to get exact results without that messy diagonalization step.

Jocelyn: It sounds like a way to trade algebraic complexity for something more programmable.

Vera: Exactly, it makes the code much more lightweight and efficient for researchers who need high precision.

Jocelyn: And when we talk about "high precision," the paper really emphasizes that this is vital because many existing tools rely on perturbative expansions or approximations. Those are great if you're in a specific regime, but if you're scanning a huge parameter space looking for new physics, an approximation might miss the very thing you're searching for. By using this SU-three expansion, Bustamante can provide exact probabilities even when things get weird, like with non-standard interactions or Lorentz-violating backgrounds.

Subrahmanyan: I was looking at the results in Figure one of the paper, and it is quite striking to see how much the oscillation patterns shift under these different scenarios. For example, when you move from vacuum to a constant density matter profile—like the Earth's crust at three grams per cubic centimeter—the probability curves for electron neutrino survival change shape significantly. It’s not just a slight nudge; it’s a fundamental reconfiguration of how those flavors swap back and forth over the one thousand three hundred km baseline used in the DUNE experiment.

Vera: That's such a good point, Subrahmanyan, because it shows why having an "exact" tool is so necessary for next-generation experiments.

Jocelyn: Right, because if we are looking for something as subtle as CPT-odd Lorentz violation—where the Hamiltonian term actually grows with neutrino energy—we can't afford to be using approximations that might wash out those high-energy effects. The paper shows these LIV effects in Figure one where the probabilities at one GeV look very different from the standard model predictions. It gives experimentalists a way to test if there is a fundamental background field in the universe that breaks Lorentz invariance, which would be a massive discovery for quantum gravity theories.

Vera: So, by providing "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities," he’s essentially giving the community a more robust lens to look through. We aren't just guessing based on the closest approximation anymore; we're looking at the actual math of how these particles behave in the most complex environments imaginable.

Jocelyn: It really turns a massive computational bottleneck into something that can be easily integrated into any large-scale oscillation analysis.

Subrahmanyan: It’s a very practical contribution to the field, moving from "it's too hard to solve analytically" to "here is an efficient way to compute it exactly."

Vera: And that precision is exactly what we need as we push into the era of high-precision neutrino physics.

Jocelyn: We'll be back after the break to discuss how these specific models, like non-standard interactions, might actually manifest in our detectors. Stay tuned.thought

Vera: We are looking at "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities" by Mauricio Bustamante. What strikes me about this work is how it tackles the sheer mathematical headache of moving beyond simple vacuum oscillations into these much messier environments. Instead of trying to diagonalize a complex Hamiltonian—which, as the paper notes, results in expressions that are often too cumbersome for anyone to actually use—he uses this method from Ohlsson and Snellman. It relies on expanding operators using SU-two and SU-three matrices to get exact results without that messy diagonalization step.

Jocelyn: It sounds like a way to trade algebraic complexity for something more programmable.

Vera: Exactly, it makes the code much more lightweight and efficient for researchers who need high precision.

Jocelyn: And when we talk about "high precision," the paper really emphasizes that this is vital because many existing tools rely on perturbative expansions or approximations. Those are great if you're in a specific regime, but if you're scanning a huge parameter space looking for new physics, an approximation might miss the very thing you're searching for. By using this SU-three expansion, Bustamante can provide exact probabilities even when things get weird, like with non-standard interactions or Lorentz-violating backgrounds.

Subrahmanyan: I was looking at the results in Figure one of the paper, and it is quite striking to see how much the oscillation patterns shift under these different scenarios. For example, when you move from vacuum to a constant density matter profile—like the Earth's crust at three grams per cubic centimeter—the probability curves for electron neutrino survival change shape significantly. It’s not just a slight nudge; it’s a fundamental reconfiguration of how those flavors swap back and forth over the one thousand three hundred km baseline used in the DUNE experiment.

Vera: That's such a good point, Subrahmanyan, because it shows why having an "exact" tool is so necessary for next-generation experiments.

Jocelyn: Right, because if we are looking for something as subtle as CPT-odd Lorentz violation—where the Hamiltonian term actually grows with neutrino energy—we can't afford to be using approximations that might wash out those high-energy effects. The paper shows these LIV effects in Figure one where the probabilities at one GeV look very different from the standard model predictions. It gives experimentalists a way to test if there is a fundamental background field in the universe that breaks Lorentz invariance, which would be a massive discovery for quantum gravity theories.

Vera: So, by providing "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities," he’s essentially giving the community a more robust lens to look through. We aren't just guessing based on the closest approximation anymore; we're looking at the actual math of how these particles behave in the most complex environments imaginable.

Jocelyn: It really turns a massive computational bottleneck into something that can be easily integrated into any large-scale oscillation analysis.

Subrahmanyan: It’s a very practical contribution to the field, moving from "it's too hard to solve analytically" to "here is an efficient way to compute it exactly."

Vera: And that precision is exactly what we need as we push into the era of high-precision neutrino physics.

Jocelyn: We'll be back after the break to discuss how these specific models, like non-standard interactions, might actually manifest in our detectors. Stay tuned.thought

Vera: We are looking at "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities" by Mauricio Bustamante. What strikes me about this work is how it tackles the sheer mathematical headache of moving beyond simple vacuum oscillations into these much messier environments. Instead of trying to diagonalize a complex Hamiltonian—which, as the paper notes, results in expressions that are often too cumbersome for anyone to actually use—he uses this method from Ohlsson and Snellman. It relies on expanding operators using SU-two and SU-three matrices to get exact results without that messy diagonalization step.

Jocelyn: It sounds like a way to trade algebraic complexity for something more programmable.

Vera: Exactly, it makes the code much more lightweight and efficient for researchers who need high precision.

Jocelyn: And when we talk about "high precision," the paper really emphasizes that this is vital because many existing tools rely on perturbative expansions or approximations. Those are great if you're in a specific regime, but if you're scanning a huge parameter space looking for new physics, an approximation might miss the very thing you're searching for. By using this SU-three expansion, Bustamante can provide exact probabilities even when things get weird, like with non-standard interactions or Lorentz-violating backgrounds.

Subrahmanyan: I was looking at the results in Figure one of the paper, and it is quite striking to see how much the oscillation patterns shift under these different scenarios. For example, when you move from vacuum to a constant density matter profile—like the Earth's crust at three grams per cubic centimeter—the probability curves for electron neutrino survival change shape significantly. It’s not just a slight nudge; it’s a fundamental reconfiguration of how those flavors swap back and forth over the one thousand three hundred km baseline used in the DUNE experiment.

Vera: That's such a good point, Subrahmanyan, because it shows why having an "exact" tool is so necessary for next-generation experiments.

Jocelyn: Right, because if we are looking for something as subtle as CPT-odd Lorentz violation—where the Hamiltonian term actually grows with neutrino energy—we can't afford to be using approximations that might wash out those high-energy effects. The paper shows these LIV effects in Figure one where the probabilities at one GeV look very different from the standard model predictions. It gives experimentalists a way to test if there is a fundamental background field in the universe that breaks Lorentz invariance, which would be a massive discovery for quantum gravity theories.

Vera: So, by providing "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities," he’s essentially giving the community a more robust lens to look through. We aren't just guessing based on the closest approximation anymore; we're looking at the actual math of how these particles behave in the most complex environments imaginable.

Jocelyn: It really turns a massive computational bottleneck into something that can be easily integrated into any large-scale oscillation analysis.

Subrahmanyan: It’s a very practical contribution to the field, moving from "it's too hard to solve analytically" to "here is an efficient way to compute it exactly."

Vera: And that precision is exactly what we need as we push into the era of high-precision neutrino physics.

Jocelyn: We'll be back after the break to discuss how these specific models, like non-standard interactions, might actually manifest in our detectors. Stay tuned.thought

Vera: We are looking at "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities" by Mauricio Bustamante. What strikes me about this work is how it tackles the sheer mathematical headache of moving beyond simple vacuum oscillations into these much messier environments. Instead of trying to diagonalize a complex Hamiltonian—which, as the paper notes, results in expressions that are often too cumbersome for anyone to actually use—he uses this method from Ohlsson and Snellman. It relies on expanding operators using SU-two and SU-three matrices to get exact results without that messy diagonalization step.

Jocelyn: It sounds like a way to trade algebraic complexity for something more programmable.

Vera: Exactly, it makes the code much more lightweight and efficient for researchers who need high precision.

Jocelyn: And when we talk about "high precision," the paper really emphasizes that this is vital because many existing tools rely on perturbative expansions or approximations. Those are great if you're in a specific regime, but if you're scanning a huge parameter space looking for new physics, an approximation might miss the very thing you're searching for. By using this SU-three expansion, Bustamante can provide exact probabilities even when things get weird, like with non-standard interactions or Lorentz-violating backgrounds.

Subrahmanyan: I was looking at the results in Figure one of the paper, and it is quite striking to see how much the oscillation patterns shift under these different scenarios. For example, when you move from vacuum to a constant density matter profile—like the Earth's crust at three grams per cubic centimeter—the probability curves for electron neutrino survival change shape significantly. It’s not just a slight nudge; it’s a fundamental reconfiguration of how those flavors swap back and forth over the one thousand three hundred km baseline used in the DUNE experiment.

Vera: That's such a good point, Subrahmanyan, because it shows why having an "exact" tool is so necessary for next-generation experiments.

Jocelyn: Right, because if we are looking for something as subtle as CPT-odd Lorentz violation—where the Hamiltonian term actually grows with neutrino energy—we can't afford to be using approximations that might wash out those high-energy effects. The paper shows these LIV effects in Figure one where the probabilities at one GeV look very different from the standard model predictions. It gives experimentalists a way to test if there is a fundamental background field in the universe that breaks Lorentz invariance, which would be a massive discovery for quantum gravity theories.

Vera: So, by providing "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities," he’s essentially giving the community a more robust lens to look through. We aren't just guessing based on the closest approximation anymore; we're looking at the actual math of how these particles behave in the most complex environments imaginable.

Jocelyn: It really turns a massive computational bottleneck into something that can be easily integrated into any large-scale oscillation analysis.

Subrahmanyan: It’s a very practical contribution to the field, moving from "it's too hard to solve analytically" to "here is an efficient way to compute it exactly."

Vera: And that precision is exactly what we need as we push into the era of high-precision neutrino physics.

Jocelyn: We'll be back after the break to discuss how these specific models, like non-standard interactions, might actually manifest in our detectors. Stay tuned.thought

Vera: We are looking at "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities" by Mauricio Bustamante. What strikes me about this work is how it tackles the sheer mathematical headache of moving beyond simple vacuum oscillations into these much messier environments. Instead of trying to diagonalize a complex Hamiltonian—which, as the paper notes, results in expressions that are often too cumbersome for anyone to actually use—he uses this method from Ohlsson and Snellman. It relies on expanding operators using SU-two and SU-three matrices to get exact results without that messy diagonalization step.

Jocelyn: It sounds like a way to trade algebraic complexity for something more programmable.

Vera: Exactly, it makes the code much more lightweight and efficient for researchers who need high precision.

Jocelyn: And when we talk about "high precision," the paper really emphasizes that this is vital because many existing tools rely on perturbative expansions or approximations. Those are great if you're in a specific regime, but if you're scanning a huge parameter space looking for new physics, an approximation might miss the very thing you're searching for. By using this SU-three expansion, Bustamante can provide exact probabilities even when things get weird, like with non-standard interactions or Lorentz-violating backgrounds.

Subrahmanyan: I was looking at the results in Figure one of the paper, and it is quite striking to see how much the oscillation patterns shift under these different scenarios. For example, when you move from vacuum to a constant density matter profile—like the Earth's crust at three grams per cubic centimeter—the probability curves for electron neutrino survival change shape significantly. It’s not just a slight nudge; it’s a fundamental reconfiguration of how those flavors swap back and forth over the one thousand three hundred km baseline used in the DUNE experiment.

Vera: That's such a good point, Subrahmanyan, because it shows why having an "exact" tool is so necessary for next-generation experiments.

Jocelyn: Right, because if we are looking for something as subtle as CPT-odd Lorentz violation—where the Hamiltonian term actually grows with neutrino energy—we can't afford to be using approximations that might wash out those high-energy effects. The paper shows these LIV effects in Figure one where the probabilities at one GeV look very different from the standard model predictions. It gives experimentalists a way to test if there is a fundamental background field in the universe that breaks Lorentz invariance, which would be a massive discovery for quantum gravity theories.

Vera: So, by providing "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities," he’s essentially giving the community a more robust lens to look through. We aren't just guessing based on the closest approximation anymore; we're looking at the actual math of how these particles behave in the most complex environments imaginable.

Jocelyn: It really turns a massive computational bottleneck into something that can be easily integrated into any large-scale oscillation analysis.

Subrahmanyan: It’s a very practical contribution to the field, moving from "it's too hard to solve analytically" to "here is an efficient way to compute it exactly."

Vera: And that precision is exactly what we need as we push into the era of high-precision neutrino physics.

Jocelyn: We'll be back after the break to discuss how these specific models, like non-standard interactions, might actually manifest in our detectors. Stay tuned.thought

Vera: We are looking at "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities" by Mauricio Bustamante. What strikes me about this work is how it tackles the sheer mathematical headache of moving beyond simple vacuum oscillations into these much messier environments. Instead of trying to diagonalize a complex Hamiltonian—which, as the paper notes, results in expressions that are often too cumbersome for anyone to actually use—he uses this method from Ohlsson and Snellman. It relies on expanding operators using SU-two and SU-three matrices to get exact results without that messy diagonalization step.

Jocelyn: It sounds like a way to trade algebraic complexity for something more programmable.

Vera: Exactly, it makes the code much more lightweight and efficient for researchers who need high precision.

Jocelyn: And when we talk about "high precision," the paper really emphasizes that this is vital because many existing tools rely on perturbative expansions or approximations. Those are great if you're in a specific regime, but if you're scanning a huge parameter space looking for new physics, an approximation might miss the very thing you're searching for. By using this SU-three expansion, Bustamante can provide exact probabilities even when things

Paper discussion segment 3: Vera: So, looking at this paper, "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities," what I find most impressive is how it moves us away from those messy approximations. Usually, when physicists study neutrinos passing through something complex like the Earth's crust, they have to rely on perturbative expansions—basically saying "this effect is small enough that we can simplify it."

Jocelyn: Right, and as the paper mentions, those approximations are often tailored to one specific scenario and might not hold up if you change the density or the energy.

Vera: Exactly. By using this Ohlsson-Snellman method instead of the standard way of diagonalizing the Hamiltonian, Bustamante has created a tool that doesn't care how "weird" your physics gets.

Subrahmanyan: It really simplifies the computational overhead for researchers, doesn't it?

Vera: It does, Subrahmanyan! If you are scanning through a massive parameter space to find where new physics might be hiding, you don't want to be stuck recalculating complex eigenvalues every single time.

Jocelyn: And that’s where the "non-standard" parts of the paper get really interesting for our understanding of the universe. They specifically show how this code handles things like Lorentz-invariance violation—which is a massive deal because it touches on potential quantum gravity effects.

Subrahmanyan: I noticed they used some pretty extreme values for those LIV parameters to make them visible in the plots, like setting b1 over lambda to ten to the power of negative twenty-one.

Jocelyn: Right, they had to scale them up just so we could see the effect on a graph, but it proves that if these effects exist at even tiny levels, this code can catch them.

Vera: It also covers Non-Standard Interactions, or NSI, which is basically the idea that neutrinos might be bumping into quarks or electrons in ways our current Standard Model doesn't predict.

Jocelyn: In "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities," the authors demonstrate how even small NSI parameters, like the ones they set for this example—where epsilon ee was zero point zero six—can visibly shift those probability curves.

Subrahmanyan: It’s essentially a universal toolkit for anyone trying to hunt for "new physics" in neutrino data.

Vera: That's a great way to put it, Subrahmanyan; it's about providing the precision necessary to tell if an observation is just a standard oscillation or something much more exotic.

Jocelyn: It really bridges that gap between theoretical math and actual experimental analysis for upcoming projects like DUNE.

Vera: Definitely, and we'll be right back after this break to wrap up our discussion on the implications of these exact calculations.thought

Vera: So, looking at this paper, what I find most impressive is how it moves us away from those messy approximations. Usually, when physicists study neutrinos passing through something complex like the Earth's crust, they have to rely on perturbative expansions—basically saying "this effect is small enough that we can simplify it."

Jocelyn: Right, and as the paper mentions, those approximations are often tailored to one specific scenario and might not hold up if you change the density or the energy.

Vera: Exactly. By using this Ohlsson-Snellman method instead of the standard way of diagonalizing the Hamiltonian, Bustamante has created a tool that doesn't care how "weird" your physics gets.

Subrahmanyan: It really simplifies the computational overhead for researchers, doesn't it?

Vera: It does, Subrahmanyan! If you are scanning through a massive parameter space to find where new physics might be hiding, you don't want to be stuck recalculating complex eigenvalues every single time.

Jocelyn: And that’s where the "non-standard" parts of the paper get really interesting for our understanding of the universe. They specifically show how this code handles things like Lorentz-invariance violation—which is a massive deal because it touches on potential quantum gravity effects.

Subrahmanyan: I noticed they used some pretty extreme values for those LIV parameters to make them visible in the plots, like setting b1 over lambda to ten to the power of negative twenty-one.

Jocelyn: Right, they had to scale them up just so we could see the effect on a graph, but it proves that if these effects exist at even tiny levels, this code can catch them.

Vera: It also covers Non-Standard Interactions, or NSI, which is basically the idea that neutrinos might be bumping into quarks or electrons in ways our current Standard Model doesn't predict.

Jocelyn: In "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities," the authors demonstrate how even small NSI parameters, like the ones they set for this example—where epsilon ee was zero point zero six—can visibly shift those probability curves.

Subrahmanyan: It's essentially a universal toolkit for anyone trying to hunt for "new physics" in neutrino data.

Vera: That's a great way to put it, Subrahmanyan; it's about providing the precision necessary to tell if an observation is just a standard oscillation or something much more exotic.

Jocelyn: It really bridges that gap between theoretical math and actual experimental analysis for upcoming projects like DUNE.

Vera: Definitely, and we'll be right back after this break to wrap up our discussion on the implications of these exact calculations.thought

Vera: So, looking at this paper, what I find most impressive is how it moves us away from those messy approximations. Usually, when physicists study neutrinos passing through something complex like the Earth's crust, they have to rely on perturbative expansions—basically saying "this effect is small enough that we can simplify it."

Jocelyn: Right, and as the paper mentions, those approximations are often tailored to one specific scenario and might not hold up if you change the density or the energy.

Vera: Exactly. By using this Ohlsson-Snellman method instead of the standard way of diagonalizing the Hamiltonian, Bustamante has created a tool that doesn't care how "weird" your physics gets.

Subrahmanyan: It really simplifies the computational overhead for researchers, doesn't it?

Vera: It does, Subrahmanyan! If you are scanning through a massive parameter space to find where new physics might be hiding, you don't want to be stuck recalculating complex eigenvalues every single time.

Jocelyn: And that’s where the "non-standard" parts of the paper get really interesting for our understanding of the universe. They specifically show how this code handles things like Lorentz-invariance violation—which is a massive deal because it touches on potential quantum gravity effects.

Subrahmanyan: I noticed they used some pretty extreme values for those LIV parameters to make them visible in the plots, like setting b1 over lambda to ten to the power of negative twenty-one.

Jocelyn: Right, they had to scale them up just so we could see the effect on a graph, but it proves that if these effects exist at even tiny levels, this code can catch them.

Vera: It also covers Non-Standard Interactions, or NSI, which is basically the idea that neutrinos might be bumping into quarks or electrons in ways our current Standard Model doesn't predict.

Jocelyn: In "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities," the authors demonstrate how even small NSI parameters, like the ones they set for this example—where epsilon ee was zero point zero six—can visibly shift those probability curves.

Subrahmanyan: It's essentially a universal toolkit for anyone trying to hunt for "new physics" in neutrino data.

Vera: That's a great way to put it, Subrahmanyan; it's about providing the precision necessary to tell if an observation is just a standard oscillation or something much more exotic.

Jocelyn: It really bridges that gap between theoretical math and actual experimental analysis for upcoming projects like DUNE.

Vera: Definitely, and we'll be right back after this break to wrap up our discussion on the implications of these exact calculations.thought

Vera: So, looking at this paper, what I find most impressive is how it moves us away from those messy approximations. Usually, when physicists study neutrinos passing through something complex like the Earth's crust, they have to rely on perturbative expansions—basically saying "this effect is small enough that we can simplify it."

Jocelyn: Right, and as the paper mentions, those approximations are often tailored to one specific scenario and might not hold up if you change the density or the energy.

Vera: Exactly. By using this Ohlsson-Snellman method instead of the standard way of diagonalizing the Hamiltonian, Bustamante has created a tool that doesn't care how "weird" your physics gets.

Subrahmanyan: It really simplifies the computational overhead for researchers, doesn't it?

Vera: It does, Subrahmanyan! If you are scanning through a massive parameter space to find where new physics might be hiding, you don't want to be stuck recalculating complex eigenvalues every single time.

Jocelyn: And that’s where the "non-standard" parts of the paper get really interesting for our understanding of the universe. They specifically show how this code handles things like Lorentz-invariance violation—which is a massive deal because it touches on potential quantum gravity effects.

Subrahmanyan: I noticed they used some pretty extreme values for those LIV parameters to make them visible in the plots, like setting b1 over lambda to ten to the power of negative twenty-one.

Jocelyn: Right, they had to scale them up just so we could see the effect on a graph, but it proves that if these effects exist at even tiny levels, this code can catch them.

Vera: It also covers Non-Standard Interactions, or NSI, which is basically the idea that neutrinos might be bumping into quarks or electrons in ways our current Standard Model doesn't predict.

Jocelyn: In "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities," the authors demonstrate how even small NSI parameters, like the ones they set for this example—where epsilon ee was zero point zero six—can visibly shift those probability curves.

Subrahmanyan: It's essentially a universal toolkit for anyone trying to hunt for "new physics" in neutrino data.

Vera: That's a great way to put it, Subrahmanyan; it's about providing the precision necessary to tell if an observation is just a standard oscillation or something much more exotic.

Jocelyn: It really bridges that gap between theoretical math and actual experimental analysis for upcoming projects like DUNE.

Vera: Definitely, and we'll be right back after this break to wrap up our discussion on the implications of these exact calculations.thought

Vera: So, looking at this paper, what I find most impressive is how it moves us away from those messy approximations. Usually, when physicists study neutrinos passing through something complex like the Earth's crust, they have to rely on perturbative expansions—basically saying "this effect is small enough that we can simplify it."

Jocelyn: Right, and as the paper mentions, those approximations are often tailored to one specific scenario and might not hold up if you change the density or the energy.

Vera: Exactly. By using this Ohlsson-Snellman method instead of the standard way of diagonalizing the Hamiltonian, Bustamante has created a tool that doesn't care how "weird" your physics gets.

Subrahmanyan: It really simplifies the computational overhead for researchers, doesn't it?

Vera: It does, Subrahmanyan! If you are scanning through a massive parameter space to find where new physics might be hiding, you don't want to be stuck recalculating complex eigenvalues every single time.

Jocelyn: And that’s where the "non-standard" parts of the paper get really interesting for our understanding of the universe. They specifically show how this code handles things like Lorentz-invariance violation—which is a massive deal because it touches on potential quantum gravity effects.

Subrahmanyan: I noticed they used some pretty extreme values for those LIV parameters to make them visible in the plots, like setting b1 over lambda to ten to the power of negative twenty-one.

Jocelyn: Right, they had to scale them up just so we could see the effect on a graph, but it proves that if these effects exist at even tiny levels, this code can catch them.

Vera: It also covers Non-Standard Interactions, or NSI, which is basically the idea that neutrinos might be bumping into quarks or electrons in ways our current Standard Model doesn't predict.

Jocelyn: In "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities," the authors demonstrate how even small NSI parameters, like the ones they set for this example—where epsilon ee was zero point zero six—can visibly shift those probability curves.

Subrahmanyan: It's essentially a universal toolkit for anyone trying to hunt for "new physics" in neutrino data.

Vera: That's a great way to put it, Subrahmanyan; it's about providing the precision necessary to tell if an observation is just a standard oscillation or something much more exotic.

Jocelyn: It really bridges that gap between theoretical math and actual experimental analysis for upcoming projects like DUNE.

Vera: Definitely, and we'll be right back after this break to wrap up our discussion on the implications of these exact calculations.thought

Vera: So, looking at this paper, what I find most impressive is how it moves us away from those messy approximations. Usually, when physicists study neutrinos passing through something complex like the Earth's crust, they have to rely on perturbative expansions—basically saying "this effect is small enough that we can simplify it."

Jocelyn: Right, and as the paper mentions, those approximations are often tailored to one specific scenario and might not hold up if you change the density or the energy.

Vera: Exactly. By using this Ohlsson-Snellman method instead of the standard way of diagonalizing the Hamiltonian, Bustamante has created a tool that doesn't care how "weird" your physics gets.

Subrahmanyan: It really simplifies the computational overhead for researchers, doesn't it?

Vera: It does, Subrahmanyan! If you are scanning through a massive parameter space to find where new physics might be hiding, you don't want to be stuck recalculating complex eigenvalues every single time.

Jocelyn: And that’s where the "non-standard" parts of the paper get really interesting for our understanding of the universe. They specifically show how this code handles things like Lorentz-invariance violation—which is a massive deal because it touches on potential quantum gravity effects.

Subrahmanyan: I noticed they used some pretty extreme values for those LIV parameters to make them visible in the plots, like setting b1 over lambda to ten to the power of negative twenty-one.

Jocelyn: Right, they had to scale them up just so we could see the effect on a graph, but it proves that if these effects exist at even tiny levels, this code can catch them.

Vera: It also covers Non-Standard Interactions, or NSI, which is basically the idea that neutrinos might be bumping into quarks or electrons in ways our current Standard Model doesn't predict.

Jocelyn: In "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities," the authors demonstrate how even small NSI parameters, like the ones they set for this example—where epsilon ee was zero point zero six—can visibly shift those probability curves.

Subrahmanyan: It's essentially a universal toolkit for anyone trying to hunt for "new physics" in neutrino data.

Vera: That's a great way to put it, Subrahmanyan; it's about providing the precision necessary to tell if an observation is just a standard oscillation or something much more exotic.

Jocelyn: It really bridges that gap between theoretical math and actual experimental analysis for upcoming projects like DUNE.

Vera: Definitely, and we'll be right back after this break to wrap up our discussion on the implications of these exact calculations.thought

Vera: So, looking at this paper, what I find most impressive is how it moves us away from those messy approximations. Usually, when physicists study neutrinos passing through something complex like the Earth's crust, they have to rely on perturbative expansions—basically saying "this effect is small enough that we can simplify it."

Jocelyn: Right, and as the paper mentions, those approximations are often tailored to one specific scenario and might not hold up if you change the density or the energy.

Vera: Exactly. By using this Ohlsson-Snellman method instead of the standard way of diagonalizing the Hamiltonian, Bustamante has created a tool that doesn't care how "weird" your physics gets.

Subrahmanyan: It really simplifies the computational overhead for researchers, doesn't it?

Vera: It does, Subrahmanyan! If you are scanning through a massive parameter space to find where new physics might be hiding, you don't want to be stuck recalculating complex eigenvalues every single time.

Jocelyn: And that’s where the "non-standard" parts of the paper get really interesting for our understanding of the universe. They specifically show how this code handles things like Lorentz-invariance violation—which is a massive deal because it touches on potential quantum gravity effects.

Subrahmanyan: I noticed they used some pretty extreme values for those LIV parameters to make them visible in the plots, like setting b1 over lambda to ten to the power of negative twenty-one.

Jocelyn: Right, they had to scale them up just so we could see the effect on a graph, but it proves that if these effects exist at even tiny levels, this code can catch them.

Vera: It also covers Non-Standard Interactions, or NSI, which is basically the idea that neutrinos might be bumping into quarks or electrons in ways our current Standard Model doesn't predict.

Jocelyn: In "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities," the authors demonstrate how even small NSI parameters, like the ones they set for this example—where epsilon

Conclusion: Vera: So, to wrap things up for this segment, we’ve really just spent our time looking at the plumbing of neutrino physics. This paper, "NuOscProbExact: a general-purpose code to compute exact two-flavor and three-flavor neutrino oscillation probabilities," provides a much more streamlined way for researchers to handle these complex calculations.

Jocelyn: Exactly. Instead of getting bogged down in the massive, messy algebraic expressions you usually see when trying to diagonalize a three-flavor Hamiltonian, this method uses SU transformations to jump straight to the results. It's incredibly versatile because it doesn't care if your neutrinos are traveling through a vacuum or through something much more exotic.

Vera: Right, and that’s the real strength here. Whether you are looking at standard matter effects in the Earth's crust, or you’re testing for new physics like non-standard interactions or even Lorentz-invariance violation, this code can handle it all without needing a new derivation every time.

Jocelyn: It really lowers the barrier for anyone trying to explore those "what if" scenarios in neutrino oscillation data. It’s a very elegant bit of mathematical heavy lifting made accessible through software.

Vera: Let's get one last word in from Subrahmanyan before we move on. Any final thoughts on this one?

Subrahmanyan: It is a wonderful example of how specialized mathematical techniques, like these expansions in SU matrices, can be turned into practical tools that accelerate experimental analysis across the board.

Jocelyn: I couldn't agree more. Well, that’s all we have for this paper. Thanks for joining us for this deep dive into neutrino math!

Vera: Stay tuned, because coming up next, we’re switching gears entirely to look at a very different part of the cosmos. We'll see you in a moment.

Jocelyn: Don't go anywhere! back in a bit.le

Mauricio Bustamante

University of Copenhagen

hep-ph, astro-ph.HE, hep-ex

Submitted: 2026-08-14

Updated: 2026-08-17

Comments: 8 pages, 1 figure, code examples, improved discussion, technical appendices. NuOscProbExact code: http://github.com/mbustama/NuOscProbExact

Code: https://github.com/bustama/NuOscProbExact

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 74/100

The gist: The paper presents NuOscProbExact, a general-purpose, open-source numerical code to compute exact two-flavor and three-flavor neutrino oscillation probabilities for arbitrary time-independent

Key concepts

NuOscProbExact
This is a general-purpose code designed to compute exact two-flavor and three-flavor neutrino oscillation probabilities.
Mauricio Bustamante
He is the author of the paper being discussed, and he is affiliated with the University of Copenhagen.
Neutrino Oscillation Probabilities
These are calculations related to how neutrinos change flavor as they travel. The code computes exact probabilities for both two-flavor and three-flavor neutrino oscillations.

Terminology

Summary

The paper presents NuOscProbExact, a general-purpose, open-source numerical code to compute exact two-flavor and three-flavor neutrino oscillation probabilities for arbitrary time-independent Hamiltonians. The code is a lightweight numerical implementation of the method developed by Ohlsson & Snellman (OS) in Refs. [45–48], which relies on expanding quantum operators in terms of SU(2) and SU(3) matrices, thereby bypassing the need to diagonalize the Hamiltonian.

Motivation and context: Computing neutrino oscillation probabilities exactly typically requires diagonalizing the Hamiltonian that drives the time-evolution of neutrinos. While exact analytic expressions exist for oscillations in vacuum, for other scenarios (e.g., oscillations in matter, with non-standard interactions, or in Lorentz-violating backgrounds) the expressions are often complex and seldom used. The paper notes that "there is no systematic way to produce these useful expressions, since they are tailored to specific Hamiltonians... their derivation is not trivial, or their application is limited to specific ranges of values of a perturbative parameter." Hence, the best course of action for high precision is to compute probabilities exactly, often numerically, especially when scanning parameter spaces without prior knowledge of the region of interest.

Method assumptions: The OS method has two assumptions: (1) the system must be closed, i.e., it must conserve the number of neutrinos summed over all flavors; (2) the Hamiltonian must be time-independent (except in some cases). These conditions are satisfied in many physical scenarios, including oscillations in vacuum, in matter of constant density, with non-standard neutrino interactions, and in diverse new-physics scenarios. The method does not apply to scenarios where neutrinos leak out of the system, e.g., 3+1 systems of sterile neutrinos, neutrino decays into invisible products, or open systems with decoherence.

Two-neutrino case: The Hamiltonian H2 is expanded as H2 = h01 + hkσk, where σk are the Pauli matrices. The coefficients h0 and hk are functions of the Hamiltonian components (Table I). The evolution operator is U2(L) = e(−iH2L), and after discarding the global phase e(−ih0L), the operator becomes U2(L) = e(−ihkσkL). Using the Pauli-matrix identity that generalizes Euler's formula, e(±iakσk) = cos(a) ± iâkσk sin(a), the evolution operator is written as U2(L) = cos(hL)1 − i[sin(hL)/h] hkσk. The final two-flavor transition probability is:

P να→νβ(L) = (h12 + h22)/h2 · sin2(hL) (α ≠ β),

where h12 + h22 = (H2)122 and h2 = (H2)122 + (H2)11 − (H2)222/4. The survival probability is P να→να(L) = 1 − P να→νβ(L).

Three-neutrino case: The Hamiltonian H3 is expanded as H3 = h01 + hkλk, where λk are the eight Gell-Mann matrices. The coefficients h0 and hk are given in Table II. The evolution operator is U3(L) = e(−iH3L) = e(−ih01L) e(−ihkλkL). After discarding the global phase, the operator is expanded as U3(L) = u01 + iukλk, where the complex coefficients u0 and uk are computed using the SU(3) invariants:

  • L2h2 ≡ L2hkhk

  • −L3⟨h⟩ ≡ −L3diⱼkhihⱼhk

The tensor diⱼk = ¼Tr(λi, λⱼ λk) has non-zero components listed in Table III. The characteristic equation of −hkλkL is φ3 − (L2h2)φ − (2/3)(−L3⟨h⟩) = 0, whose three latent roots are φm ≡ ψmL with ψm ≡ (2h/√3)cos((χ + 2πm)/3), where cos(χ) = −√3⟨h⟩/h3. This step is key: writing the eigenvalues in terms of the SU(3) invariants allows us to bypass an explicit diagonalization.

The coefficients are:

u0 = (1/3)Σm e(iLψm)

uk = Σm e(iLψm) [ψmhk − (h∗h)k] / (3ψm2 − h2)

where (h∗h)i ≡ diⱼkhⱼhk. The evolution operator is written concisely as U3(L) = Σm e(iLψm)[1 + (ψmhk − (h∗h)k)λk/(3ψm2 − h2)]. The flavor-transition probabilities are given in Table IV in terms of u0 and uk, e.g., P νe→νe = u0 + iu3 + iu8/√32, P νe→νμ = iu1 − u22, etc.

Code description and examples: NuOscProbExact is fully written in Python 3.7, open source, and publicly available in a GitHub repository. The main input is the Hamiltonian matrix H2 or H3, provided as a 2×2 or 3×3 list. The code internally computes the hk coefficients using Table I (two-neutrino) or Table II (three-neutrino). For two-neutrino probabilities, it evaluates Eq. (6); for three-neutrino probabilities, it evaluates the expressions in Table IV. The code includes sample Hamiltonians for oscillations in vacuum, in matter of constant density, with non-standard interactions, and with CPT-odd Lorentz-violating background.

Example scenarios: The paper demonstrates the code with four representative scenarios, all at baseline L = 1300 km (matching the DUNE far detector), using NuFit 4.0 best-fit values (sin2θ12 = 0.310, sin2θ23 = 0.582, sin2θ13 = 0.02240, δ CP = 217°, Δm221 = 7.39·10−5 eV2, Δm231 = 2.525·10−3 eV2, normal mass hierarchy):

  1. Vacuum: H vac3(E) = (1/2E)R3,θ M23 R†3,θ, where M23 = diag(0, Δm221, Δm231) and R3,θ is the PMNS matrix.

  2. Matter of constant density: H matt3(E) = H vac3(E) + A3, where A3 = diag(V CC, 0, 0) with V CC = √2G F ne. The paper uses ρ = 3 g cm−3 (average crust density), Ye = 0.5, and notes that using average density is a good approximation even with density changes.

  3. Non-standard interactions (NSI): H NSI3(E) = H vac3(E) + A3 + V3, where V3 = V CC ε3 and ε3 is the matrix of NSI strength parameters. The paper uses arbitrary values allowed at 2σ by a global fit: ε ee = −ε eμ = 0.06, ε μμ = 1.2, others zero.

  4. CPT-odd Lorentz-violating background (LIV): H LIV3(E) = H vac3(E) + (E/Λ)R3,ξ B3 R†3,ξ, where B3 = diag(b1, b2, b3) and R3,ξ is a mixing matrix. The paper uses artificially high values b1/Λ = b2/Λ = 10−21, b3/Λ = 5·10−21, with all LIV mixing angles set to zero.

Extensions: The paper notes that the method can be extended to multiple slabs of constant-density matter by stitching probability amplitudes: if a neutrino traverses N slabs slabs, the evolved state is ν α(Lⱼ) = Πⱼ U3(ʲ)(Lⱼ)ν α, and the final probability is P να→νβ(Lⱼ) = ν β†ν α(Lⱼ)2.

Conclusions: The paper emphasizes that NuOscProbExact is useful for exploring wide parameter spaces where approximate expressions are unavailable, especially in three-neutrino oscillations where analytic expressions are often unavailable. The code is lightweight and can be easily incorporated into diverse oscillation analyses of standard and non-standard oscillations.

Improvements for AI systems

Based on the paper, here are specific improvements that can be made to AI systems, particularly those involved in physics simulation, numerical computation, and scientific software development:

1. Improved Numerical Stability in Matrix Exponentiation

  • Improvement: Implement the Ohlsson-Snellman (OS) method as an alternative to direct Hamiltonian diagonalization for computing time-evolution operators. The OS method uses SU(2) and SU(3) exponential expansions, which bypasses the need for eigen-decomposition.

  • Specific capability: An AI system can now compute exact two- and three-flavor neutrino oscillation probabilities for arbitrary time-independent Hamiltonians without suffering from numerical instabilities that arise when the Hamiltonian has degenerate eigenvalues or near-degenerate states. The system can handle cases where direct diagonalization fails or becomes inaccurate.

2. General-Purpose Hamiltonian Handling

  • Improvement: Design the AI to accept any 2×2 or 3×3 Hermitian matrix as input, not just pre-defined physical models.

  • Specific capability: The AI can automatically compute oscillation probabilities for:

  • Vacuum oscillations (standard PMNS mixing)

  • Constant-density matter oscillations (with MSW effect)

  • Non-standard interactions (NSI) with arbitrary coupling matrices

  • Lorentz-invariance violation (LIV) with CPT-odd terms

  • Any user-defined new physics Hamiltonian, without requiring the AI to derive new analytic formulas.

3. Exact Probability Calculation Without Approximation

  • Improvement: Replace perturbative or approximate methods (which are common in current AI-driven physics tools) with the exact closed-form expressions given in Table IV of the paper.

  • Specific capability: The AI can produce results with machine precision (e.g., 1e-15 relative error) for all oscillation channels (νe→νe, νe→νμ, νe→ντ, etc.) across the entire parameter space, including regions where perturbative expansions break down (e.g., high matter density, large NSI parameters, or high-energy LIV).

4. Efficient Computation of SU(3) Invariants

  • Improvement: Implement the computation of the tensor d ijk (Table III) and the invariants h2 and ⟨h⟩ directly from the Hamiltonian coefficients, avoiding symbolic manipulation.

  • Specific capability: The AI can compute the three eigenvalues ψ m (Eq. 9) using the trigonometric solution to the cubic equation, which is numerically robust and avoids complex arithmetic. This enables real-time computation of probabilities for a large grid of energies and baselines (e.g., 10,000 points in < 1 second on a standard CPU).

5. Modular Code Architecture for Extensibility

  • Improvement: Structure the AI system to separate the core probability engine (oscprob3nu.py) from the Hamiltonian generators (hamiltonians3nu.py) and constants (globaldefs.py), as shown in Listing 1.

  • Specific capability: The AI can be easily extended to new physics scenarios by adding a new Hamiltonian function without modifying the core probability solver. This reduces development time for researchers exploring novel theories.

6. Handling of Multi-Slab Matter Profiles

  • Improvement: Implement the method described in Section VI.B for piecewise-constant matter densities, where the evolution operator is the product of individual slab operators.

  • Specific capability: The AI can compute oscillation probabilities for realistic Earth density profiles (e.g., PREM model) or long-baseline experiments with varying rock densities, by stitching together exact solutions for each constant-density segment. This is more accurate than using a single average density.

7. Two-Flavor and Three-Flavor Unified Framework

  • Improvement: Provide both two-flavor (Eq. 6) and three-flavor (Table IV) solvers in the same codebase, with automatic detection of the input matrix size.

  • Specific capability: The AI can seamlessly switch between simplified two-flavor approximations (for quick estimates) and full three-flavor calculations (for precision), allowing users to validate approximations against exact results.

8. Numerical Cross-Validation and Error Estimation

  • Improvement: Include built-in checks that the computed probability matrix is unitary (sum of probabilities = 1) and that the trace of the evolution operator matches the analytical expression.

  • Specific capability: The AI can automatically flag numerical errors or input Hamiltonian errors (e.g., non-Hermitian matrices) and provide error diagnostics, preventing silent failures in large-scale parameter scans.

9. High-Energy and Long-Baseline Scalability

  • Improvement: Use the fact that the OS method does not require diagonalization, which is O(n3) for n×n matrices. For 3×3, this is trivial, but the method scales well to larger systems (up to n=6 as mentioned in the paper).

  • Specific capability: The AI can be extended to 4-flavor or 5-flavor systems (e.g., sterile neutrinos) using the same SU(n) expansion approach, without a fundamental rewrite of the core algorithm.

10. Integration with Machine Learning for Parameter Inference

  • Improvement: Since the probability computation is now exact and fast, the AI can be used as a forward model in Bayesian inference or neural network training.

  • Specific capability: The AI can generate training data for neural networks that map Hamiltonian parameters to oscillation probabilities, or be embedded in a Markov Chain Monte Carlo (MCMC) sampler to fit oscillation data (e.g., from DUNE, JUNO) to new physics models, with exact likelihoods at every step.

11. Handling of Global Phases

  • Improvement: Correctly discard the global phase factor e-i h0 L as done in the paper (Section IV), which is often a source of errors in naive implementations.

  • Specific capability: The AI will produce probabilities that are independent of the trace of the Hamiltonian, ensuring that adding a constant to all diagonal elements does not change the physics, which is a crucial test of correctness.

12. Production-Ready Code Quality

  • Improvement: Adopt the code structure from the paper: pure Python 3.7, no external dependencies beyond NumPy, and clear separation of concerns.

  • Specific capability: The AI can be deployed in environments where installing heavy scientific libraries (e.g., ROOT, Geant4) is not possible, and can be run on minimal hardware (e.g., Raspberry Pi) for educational or field applications.

Summary of what the improved AI system can do:

  • Compute exact neutrino oscillation probabilities for any 2- or 3-flavor time-independent Hamiltonian in microseconds.

  • Handle vacuum, matter, NSI, LIV, and custom new physics models with no code changes.

  • Provide machine-precision results that are unitarily consistent and numerically stable.

  • Scale to multi-slab density profiles and higher flavor counts.

  • Serve as a reliable forward model for parameter estimation and experimental design.

  • Be easily extended by researchers without deep knowledge of the underlying mathematics.

Sources

Related papers