Directed Wada cascade in systems with discrete scale invariance
summary
The gist
Dissipative systems whose equations are unchanged by a dilation can host an infinite ladder of attractors…
In short
The research investigates dissipative systems with discrete scale invariance (DSI), where equations remain unchanged under magnification by a factor S. This symmetry leads to an infinite ladder of attractors, but instead of a standard Wada property, basins form a directed cascade nesting toward the origin. This structure means uncertainty about the final state is unbounded inward but bounded outward.
Key concepts
- Discrete Scale Invariance (DSI)
- DSI is a symmetry where the equations describing a dissipative system look identical after you magnify them by a fixed factor S and its powers. While this suggests perfect self-similarity, the paper shows this symmetry fails in a specific way, leading to structured complexity rather than simple repetition.
- Directed Wada Cascade
- This is the specific structure of basin boundaries where nesting occurs only in one direction—toward the origin. Points on one shell's boundary lie on all deeper shells' boundaries. This nesting is finite and strict, meaning you can never find a point that lies on the boundary of every single attractor.
- Reach Law
- The reach law describes how uncertainty behaves in these systems. It states that the uncertainty is unbounded as you approach the origin (shrinking r) but remains bounded outward. This law is derived from the interplay between trapping regions and discrete scale invariance, providing a predictive rule for system behavior.
Terminology used across episodes
This episode discusses
- Directed Wada cascade in systems with discrete scale invariance · Paper Radio
- Symmetry Groups of Basins of Attraction in Equivariant Dynamical Systems
The paper
Directed Wada cascade in systems with discrete scale invariance · Read on arXiv
P. Haerter, R. L. Viana, M. A.F. Sanjuáoán, I. L. Caldas
Instituto de Física, Universidade de São Paulo · Departamento de Física, Universidade Federal do Paraná · Centro Interdisciplinar de Ciência, Tecnologia e Inovação, Núcleo de Modelagem e Computação Científica · Nonlinear Dynamics, Chaos and Complex Systems Group at Universidad Rey Juan Carlos · Royal Academy of Sciences of Spain
Dissipative systems whose equations are unchanged by a dilation can host an infinite ladder of attractors, each shell an exact magnified copy of the one before. Since such dynamics cannot tell a magnification from its inverse, the natural expectation is that these basins are maximally entangled, with every boundary point lying on the boundary of all of them, following the Wada property. We find that this expectation fails, and fails in a precise way. Dilation covariance and a single trapping region together bound which basin boundaries a point can lie on, and the bound depends only on its distance to the origin. No point lies on the boundary of all of them. The basins on whose boundary a point does lie are the deeper ones, as we prove for the stable manifolds of the boundary saddles and support numerically for the rest of the boundary, so the boundaries nest in one direction only. We call this structure a directed Wada cascade, and the standard Wada tests, as usually applied, do not detect it. It arises in maps whose nonlinearity is a log-periodic modulation of a homogeneous function, because the same periodicity builds the ladder through mode locking and confines the orbits. We show the cascade in maps that differ in invertibility, nonlinearity, scaling exponent and dimension, and in a flow.
Transcript
Introduction to the show: ident: Nonlinear Sciences. The day in nonlinear science: chaos, patterns, networks and adaptation.
Kai: We're looking at the paper: "Directed Wada cascade in systems with discrete scale invariance".
Mira: Here's the gist of the paper. Dissipative systems whose equations are unchanged by a dilation can host an infinite ladder of attractors, each shell an exact magnified copy of the one before The basin boundaries are not Wada but form a directed Wada cascade where nesting runs in one direction only toward the origin This structure arises in maps whose nonlinearity is a log-periodic modulation of a homogeneous function, and it makes the uncertainty about the final state unbounded toward the origin but bounded outward The Nature of Discrete Scale Invariance Discrete scale invariance (DSI) is a symmetry where dissipative system equations look the same after magnification by a fixed factor S and its powers This symmetry suggests the most entangled arrangement, the Wada property, but it fails in a precise way The symmetry maps an attractor onto another, leading to a ladder of attractors Ak = S − kA0 This differs from discrete translation where basins are copies of one another and passes the symmetry to the basin boundaries in full The Directed Wada Cascade Structure The basins on whose boundary a point does lie are the deeper ones, as proved for the stable manifolds of the boundary saddles and support numerically for the rest of the boundary, so the boundaries nest in one direction only This structure is called a directed Wada cascade It has three properties: (i) Nesting.
Kai: A point on the boundary of the basin of one shell lies on the boundary of the basins of all deeper shells (ii) Finite reach.
Mira: The outermost basin on whose boundary a point can lie is set by its distance to the origin, so no point lies on the boundary of all basins (iii) Strictness.
Kai: The boundaries of consecutive basins differ, ∂Bk ≠ ∂Bk+1, and with (i) the deeper one has points that the outer one lacks Mechanism of Confinement and Reach Law Two ingredients are enough for this to happen: the symmetry and a trapping square The trapping squares are defined by a region R that the map cannot leave once a trajectory has entered it, F(R) ⊆ R Discrete scale invariance extends this to a whole family of balls where the set of trapping radii is invariant under ρ → Sρ The reach law is given by k ≥ logS(ρ∗0/r) This law shows that the uncertainty is unbounded inward and bounded outward Generators of the Cascade The mechanism needs trapping balls, and for the matryoshka map they can be found explicitly The simplest example is a map of the plane called the matryoshka map The recipe is a linear map plus a nonlinear term whose strength is a bounded function of the logarithm of the amplitude, periodic with period ln S This log-periodic modulation builds the ladder of attractors through mode locking and also confines the orbits Experimental Verification and Conclusion The reach law holds in every covariant system with a trapping ball The standard Wada diagnostics cannot see this because they ask only how many basins meet near a point and there are always at least three The uncertainty is unbounded inward and bounded outward The cascade appears where both trapping regions and the ladder exist, not just one ingredient alone This asymmetry in the final-state uncertainty is a consequence of the structure Examples of Systems Exhibiting the Cascade The paper shows the cascade in maps that differ in invertibility, nonlinearity, scaling exponent and dimension, and in a flow Examples include two more planar maps differing from the first, a flow of three variables, and a network of higher dimension The network of coupled matryoshka maps shows that the bound is never violated across various topologies and dimensions In high dimension, the fraction that end on the largest attractor falls from 0.39 at d = 2 to 0.008 at d = 32 and is below 10−3 for d ≥ 64 The directed asymmetry goes the other way as dimension grows Open Questions and Future Directions Several questions remain open, such as proving that the stable manifolds of the boundary saddles make up the whole boundary For the radial log-periodic map only the crossing of manifolds is certified, and for the log-periodic rotation map, the flow and network show numerical evidence Extending these certificates to these cases is a natural next step The last question involves a physical realization of the ladder in an electronic circuit where the reach law could be tested directly Acknowledgments P.
Title and authors: Kai: So we're looking at this paper called "Directed Wada cascade in systems with discrete scale invariance." It’s by Haerter, Viana, Sanjuán, and Caldas. The main idea is that instead of everything being totally entangled like a classic Wada property suggests, the boundaries actually form a specific directed structure where nesting only happens one way.
Mira: That's right. They start by looking at dissipative systems whose equations don't change when you blow them up by some factor S. The natural expectation is that you get this complete entanglement, this Wada property where every boundary point touches every basin boundary, but they show that this fails in a very specific way.
Lev: And what's new here is that the symmetry doesn't map one attractor onto another perfectly in a way that forces total overlap. Instead, it leads to something much more constrained.
Kai: They call this structure a directed Wada cascade, and they show it arises when the nonlinearity in the system is modulated by something log-periodic with respect to S. This modulation does two things at once, which is key.
Mira: It builds that ladder of attractors—the exact magnified copies of each other—and it also acts like a confinement mechanism that keeps the orbits trapped within certain regions. That's the first big concept they introduce here.
Lev: What this means for us in terms of hardware, it suggests we might be able to engineer systems where we can control this nesting. The log-periodic modulation is essentially what builds the ladder through mode locking, which helps confine the dynamics as well as creating those attractors.
Kai: So they introduce these three properties: nesting, finite reach, and strictness. Nesting means a point on one shell's boundary lies on the boundary of all deeper shells. But there's also that finite reach property.
Mira: Finite reach means you can’t just pick any point in the plane and find it on the boundary of every single basin. The outermost basin you can be on is limited by how far away that point is from the origin, which is a concrete bound.
Lev: From an engineering standpoint, this finite reach is important because it sets a limit on how much uncertainty we have about where something will end up, depending on its starting position relative to the origin.
Title and authors: Kai: And strictness means that boundaries between consecutive shells are actually different. The boundary of shell k is not the same as the boundary of shell k plus one.
Mira: That difference, combined with nesting, creates this one-way flow toward deeper attractors—the origin in this context—because you're always moving toward the basin you belong to. This isn't just random mixing; it’s directed.
Lev: It sounds like the mechanism is very robust because it relies on a single trapping region, which they define as a set R that the map cannot leave once it enters it, and that this region is invariant under scaling by S. That invariance is what keeps the structure stable.
Kai: So for someone just listening to this, what does this mean? It means if you’re tracking something in one of these systems, you can’t be totally uncertain about its final state no matter how much you zoom in or out on the dynamics.
Mira: Exactly. The uncertainty about the final state is unbounded when looking inward toward smaller attractors, but it stays bounded outward by a distance from the origin that depends on your initial conditions.
Lev: That distance bound is given by this reach law, which they state as being greater than or equal to log S of the ratio of some reference radius to your current position. It’s a mathematical statement about how uncertainty scales with the magnification factor S.
Kai: The paper shows this reach law holds in every covariant system that has a trapping ball, but standard Wada diagnostics can't see it because they just ask how many basins meet near a point, and there are always at least three nearby.
Mira: They miss this because they don't account for the directed asymmetry in the final state uncertainty, which is a consequence of this cascade structure. It’s not just about counting boundaries; it’s about the direction that those boundaries point.
Lev: The paper uses interval arithmetic to prove that the crossing between neighboring shells is actually certified for systems like the matryoshka map with specific parameters. This gives them confidence in the nesting property.
Kai: So, to wrap up on this paper, it’s about how discrete scale invariance can lead to a directed structure instead of total chaos, thanks to a log-periodic modulation building the ladder and confining the orbits via a single trapping region.
Title and authors: Mira: It gives us a concrete mechanism for why some dissipative systems organize themselves in this highly structured way when they have that specific type of symmetry.
Lev: And for real hardware, it suggests that if we can design our systems with this log-periodic nonlinearity, we can get predictable confinement and bounded uncertainty, which is vital for reliable operation.
Kai: It’s a very specific kind of structure that emerges from scale invariance when you add a bit of controlled complexity through modulation.
Mira: It moves beyond just saying there are many attractors and explains *why* they are arranged in this particular cascading fashion rather than being randomly distributed on the phase space.
Lev: We need to keep checking how these confinement mechanisms scale up as we move to higher dimensions, because the paper does show that at high dimensions, like dimension thirty-two or sixty-four the fraction of orbits ending on the largest attractor drops really fast.
Kai: So for now, it’s a strong theoretical framework showing that this directed cascade isn't just an accident; it’s a predictable consequence of scale invariance coupled with log-periodic forcing and trapping.
Mira: It gives us a way to understand how complexity can be organized into nested structures that aren't perfectly symmetric, which is something we see everywhere in condensed matter systems.
Lev: We'll keep looking at those open questions about proving that the stable manifolds of the boundary saddles truly make up the whole boundary, because that’s where a lot of the numerical support comes from.
Kai: So, to summarize this paper on "Directed Wada cascade in systems with discrete scale invariance," we see a specific way that scale invariance leads to a directed nesting of attractors bounded by distance from the origin.
Mira: The key is the log-periodic modulation building the ladder and trapping regions dictating where those boundaries can lie, leading to an uncertainty profile that goes inward but stays bounded outward.
Lev: This structure is testable in principle, and for engineers, it gives us a blueprint for designing systems where we can control this confinement through careful tuning of the nonlinearity.
Kai: It’s a really cool way to look at how we can get order out of systems that are fundamentally dissipative and scale invariant.
The paper's summary: Kai: So, this paper is about dissipative systems that have discrete scale invariance, which means their equations look the same if you blow them up by some factor S, and they show that this symmetry leads to a specific kind of structure instead of just total chaos.
Mira: Right. They're looking at how that scale invariance creates a directed Wada cascade. The key is that the nesting doesn't happen everywhere; it runs in one direction only toward the origin.
Lev: It’s about confinement too, because this isn't just about attractors popping out randomly; there’s a mechanism that traps the orbits inside certain regions determined by how far they are from the origin.
Kai: So, what does that mean for us? It means we can build systems where we know exactly how the uncertainty of the final state scales with its distance from the center of operation.
Mira: That uncertainty is unbounded when you look toward smaller attractors, but it’s capped by a reach law based on that scaling factor S and your starting radius. It’s not just a random guess about where things go.
Lev: From an error-correction standpoint, that reach law tells us the overhead we need to account for when trying to predict the long-term behavior of the system on real hardware. It sets a physical limit on predictability based on initial conditions.
Kai: The authors use maps like the matryoshka map as their example. They show that this structure is built by a log-periodic modulation in the nonlinearity, which also acts like a mode locking mechanism to build those attractors and confinement regions at the same time.
Mira: That's the recipe they’re using. The log-periodic term is what organizes everything into that ladder of magnified copies, and it dictates how those orbits stay confined by controlling gain across different scales.
Lev: I look at that modulation in terms of control theory; it suggests a way to use periodic gain changes to actively manage the dynamics and keep them locked into these specific attractors without needing perfect external feedback loops everywhere.
Kai: It’s about finding this balance where scale invariance gives us structure, but the log-periodic term provides the necessary asymmetry that makes the cascade directed rather than just symmetrical nesting.
Mira: The result is a system where you have a clear hierarchy of basins, and you know which basin you are in based on its depth relative to others. It’s not all equally likely to end up in any one attractor.
Lev: The paper shows that this works across different dimensions, even going up to thirty-two or sixty-four dimensions; at those higher numbers, the fraction of orbits ending on the biggest attractor drops very fast, which tells us about stability in complex AI models.
Kai: So if we want to design a system with this property, we need that specific log-periodic nonlinearity and a trapping region that respects the scale invariance.
Mira: Exactly. It’s not just about having many attractors; it’s about how they are organized into this specific, one-way funnel structure.
Lev: This is important because it gives us a mathematical way to quantify the degradation of predictability as we increase the complexity or dimension of the system we are trying to model or control.
Kai: We're going to keep digging into those open questions about proving that these stable manifolds from boundary saddles actually fill up the entire basin boundaries, because that’s where much of their confidence in this structure comes from.
The paper's improvements: Tom: So we’re looking at how this paper builds on the original concept of discrete scale invariance to fix some problems in their model of attractors, focusing on improving that directed cascade structure they described earlier.
Kai: Right, they’re taking the initial framework and showing how to make it more robust by adding specific constraints related to the physical realization, like confinement regions defined by trapping balls.
Mira: They're suggesting that we need more than just a simple map; we need those trapping squares where the map physically cannot leave once you enter it. That’s crucial for getting that directed behavior right.
Lev: I see why they push for those trapping balls; on real hardware, you can’t have infinite freedom, and defining a region R that the map cannot leave is a very practical way to enforce those boundaries.
Kai: And they link this back to the reach law again, showing how the scaling of these trapping radii has to be invariant under the magnification factor S. That keeps everything consistent across different scales.
Mira: It basically means that if you scale up your system by S, your confinement region scales in a predictable way that doesn't break the underlying symmetry of the equations.
Lev: That invariance is what makes it runnable; if that scaling relationship isn't there, the error correction overhead blows up because you can’t predict how much energy or state space you need to cover at a given scale.
Kai: So they’re moving from just a mathematical description to something that has some physical grounding by tying the scaling of these regions directly to the system's fundamental symmetry.
Mira: It sharpens the result; instead of just saying it *might* happen, they show exactly what kind of nonlinearity—the log-periodic modulation—is required to generate that specific cascade structure.
Lev: From an error-correction standpoint, knowing that the scaling is invariant helps us design algorithms where we can anticipate how much uncertainty will be introduced when we try to simulate or run the system at a different resolution.
Kai: It’s about making sure that when we build these complex AI models, they don't just have many attractors floating around; they have a clear, one-way path toward certain states based on their starting position relative to the origin.
Mira: That structure is what allows for that predictability in the final state uncertainty; it’s not just random noise; it’s organized uncertainty.
Lev: We need to watch how these improvements hold up when we move into higher dimensions, because even with this better structure, we still have to deal with the computational cost of simulating those larger spaces.
Kai: The next thing they are looking at is extending these results beyond simple maps into flows and networks, checking if that directed cascade structure survives when things get more complex.
Conclusion: Kai: So we've gone through how discrete scale invariance leads to a directed Wada cascade, showing that nesting runs in one direction and uncertainty is bounded outward by distance from the origin in this paper called "Directed Wada cascade in systems with discrete scale invariance."
Mira: Exactly. The main takeaway is that the specific log-periodic nonlinearity we introduced actually creates this asymmetry, which standard diagnostics completely miss. It’s about organization rather than just a bunch of overlapping boundaries.
Lev: I think what's most interesting for error correction is seeing how these confinement properties scale; if the trapping ball scales correctly with S, it means the overhead for simulating that system doesn't explode in a way we can't calculate.
Kai: And we’re still waiting on those final proofs, like showing that the stable manifolds of the boundary saddles actually fill up the whole basin boundary, which is where much of their numerical support comes from.
Mira: It really changes how you think about complexity; it suggests that in systems with this specific type of symmetry, you don't just get chaos; you get a highly structured form of order.
Lev: For real hardware, it means we can start designing controllers that are aware of this scaling law, which would save a ton on the computational resources needed to keep the system stable over long periods.
Kai: We’ll see if we can build something concrete that exhibits this structure, maybe using some of those control methods they proposed for managing the modulation.
Mira: Definitely. It opens up a new avenue for theorists to look at how physical constraints like trapping regions dictate the geometry of attractor landscapes in dissipative systems.
Lev: We should keep an eye on those higher-dimensional results they showed; seeing that fraction drop so fast suggests that complexity can actually be managed or understood better when you move into larger spaces.
Kai: It’s a really cool piece of work because it connects the abstract math of scale invariance to tangible confinement mechanisms, which is exactly what we are trying to achieve with our superconducting qubits.
More episodes
- 2608.20276-Flip rate prediction in the double pendulum
- The day in conversation for 2026-11-18
- Daily Summary for 2026-11-18
- 2603.16051-Gradual emergence of temporal structures depending on the distance between neighboring callers in natural habitat of male treefrogs
- 2603.21801-Diffraction of deep-water solitons
- The day in conversation for 2026-11-17
- Daily Summary for 2026-11-17
- 2204.04335-Measuring and improving community resilience: a Fuzzy Logic approach
- 2604.11624-Prediction of chaotic dynamics from data: An introduction
- 2610.21575-Cooperation Dynamics with Environmental Feedback: Steady States and Transitions