A Note On Lattice Definition of Gauge Theories With Chiral Fermions

arXiv:2611.05051 · hep-lat, hep-ph, hep-th · Submitted 2026-11-03 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: High Energy Physics: Lattice. The day in lattice field theory: what the simulations put on the table.

Kai: We're looking at the paper: "A Note On Lattice Definition of Gauge Theories With Chiral Fermions".

Mira: Here's the gist of the paper. Lattice gauge theory with chiral fermions is described by interpolating a lattice gauge field over the continuum and coupling it to continuum chiral fermions to define a regularization scheme that can be implemented in principle on a computer <ref:2611.05051#pg33> The paper details an approach based on interpolating an admissible lattice gauge field over the continuum and coupling it to continuum chiral fermions, which provides a method for defining the fermion path integral in four dimensions for any anomaly-free representation <ref:2611.05051#pg8> This technique relies on two key facts: the well-definedness of the fermion path integral in continuum Yang-Mills theory and its ability to be implemented numerically by coupling it to an interpolated lattice gauge field <ref:2611.05051#pg4> The analysis generalizes previous results, applying to any gauge group and anomaly-free fermion representation in any dimension <ref:2611.05051#pg4> The Well-Definedness of the Continuum Path Integral The first key fact is that in continuum Yang-Mills theory with any gauge group G, the path integral for fermions in any anomaly-free representation, chiral or not, coupled to any classical background field A (and any background spacetime metric), is always well-defined <ref:2611.05051#pg4> This statement is supported by two points: 1.

Kai: In learning how to define the fermion path integral for anomaly-free theories, one also learns to determine all possible nonperturbative anomalies <ref:2611.05051#pg5>.

Mira: For example, in four-dimensional chiral gauge theories with connected and simply-connected gauge group, one learns that the only anomaly beyond the usual triangle anomalies is the familiar mod 2 anomaly involving the number of fermion zero-modes in an instanton field <ref:2611.05051#pg5>.

Title and authors: Kai: We’ve been talking about the core idea behind "A Note On Lattice Definition of Gauge Theories With Chiral Fermions," specifically how they regularize chiral gauge theories on a lattice using interpolated fields and continuum fermions.

Mira: They are really tackling the fermion doubling problem head-on by using interpolation, which is crucial because it allows them to couple lattice gauge fields to continuum chiral fermions without getting that problem in the first place <ref:2611.05051#pg4>.

Lev: From an error correction standpoint, I see the value of that interpolation method being in how much overhead we’d need for those iterative Landau gauge problems <ref:2611.05051#pg3>.

Kai: The paper also points out that they can make improvements on this approach by using a specific formula derived from anomaly inflow and eta-invariants to define the lattice path integral, which is something they call equation one point two <ref:2611.05051#pg4>.

Mira: That formula is important because it allows them to incorporate the discrete and continuous theta-angles into the lattice path integrals by relating them to homomorphisms from the cobordism group to U(one) <ref:2611.05051#pg9>.

Lev: So, for us, it means that if we can successfully implement that on hardware, we’d be dealing with an integral over a compact space of finite dimension <ref:2611.05051#pg4>.

Kai: Right, and the paper also suggests improvements in how they compute the absolute value of the fermion determinant by using a Laplace-like operator D2Y with an inherited APS boundary condition to avoid doubling in numerical study <ref:2611.05051#pg4>.

Mira: That sounds like a neat trick for handling the absolute value of the determinant, which is usually hard because it avoids the positivity issue in Monte Carlo methods <ref:2611.05051#pg4>.

Lev: If we can compute that spectral information from D2Y without doubling, that drastically reduces what we have to worry about when translating this to a real system <ref:2611.05051#pg4>.

Kai: So the overall message is they are providing a path from the continuum definition of the fermion path integral to an implementable lattice scheme, even if it’s not suitable for standard Monte Carlo methods <ref:2611.05051#pg1>.

Mira: It shifts the focus from being stuck on the doubling problem to defining how to handle these nonperturbative aspects through this regularization scheme <ref:2611.05051#pg4>.

Lev: It changes the overhead landscape because it moves us away from needing a full positive measure, but we still have that finite dimension integral to manage <ref:2611.05051#pg4>.

The paper's summary: Kai: To summarize what they found, the paper is building on previous work by Craig and Hsu to describe a lattice regularization of chiral gauge theories that applies in any dimension and for any anomaly-free representation <ref:2611.05051#pg3>.

Mira: They use two key facts: the first is the well-definedness of the fermion path integral in continuum Yang-Mills theory and its ability to be implemented numerically by coupling it to an interpolated lattice gauge field <ref:2611.05051#pg4>.

Lev: So they are essentially saying that because we know how to do it in the continuum, we can use that knowledge as a basis for defining the path integral on the lattice <ref:2611.05051#pg4>.

Kai: The second fact is that this approach works because it uses no special properties of the gauge group or fermion representation, making it applicable to any anomaly-free representation in four dimensions <ref:2611.05051#pg3>.

Mira: And they also explain the path integral formula for an anomaly-free system, chiral or not, can be described by explicit formulas that could be evaluated numerically on a computer <ref:2611.05051#pg5>.

Lev: So the main result here is that they provide an explicit formula for the fermion partition function involving the absolute value of a determinant and an exponential term related to the eta-invariant <ref:2611.05051#pg4>.

Kai: That formula is valid in any anomaly-free representation and depends only on the spacetime W and the gauge field <ref:2611.05051#pg5>.

Mira: It connects everything together by showing how to define the path integral on an arbitrary manifold W by considering its cobordism structure and using quantum renormalized theta-angles <ref:2611.05051#pg9>.

Lev: So they are giving us a general definition that doesn't depend on picking a specific five-manifold Y, just the manifold W <ref:2611.05051#pg9>.

The paper's improvements: Kai: The paper suggests improvements center around using the interpolation of admissible lattice gauge fields to construct a smooth continuum gauge field by acting on that interpolated field with the Yang-Mills gradient flow for a short time <ref:2611.05051#pg4>.

Mira: That smoothing step is important because it ensures the continuum gauge field is of order epsilon in lattice units, which helps ensure we have a good continuum limit <ref:2611.05051#pg4>.

Lev: From an engineering view, that gradient flow step means we’re solving finitely many Landau gauge Dirichlet problems which can be solved numerically <ref:2611.05051#pg4>.

Kai: And then they also discuss how to define smooth continuum gauge fields by acting on the interpolated A with that gradient flow for a short time <ref:2611.05051#pg4>.

Mira: This is a concrete way to get a smooth continuum gauge field in the same topological class as the original lattice configuration <ref:2611.05051#pg4>.

Lev: It sounds like we’re proposing an iterative numerical procedure that builds the field piece by piece, which is something I can actually build for real hardware <ref:2611.05051#pg4>.

Kai: And then they also have this idea about constructing smooth continuum interpolations of admissible nonabelian lattice gauge fields using a sequence of increasingly fine lattices where the top dimensional cells have bounded flatness to ensure a continuum limit <ref:2611.05051#pg4>.

Mira: That bottom part is what ensures we can actually get that continuum limit you need for your experiments, which is essential for any experimental setup <ref:2611.05051#pg4>.

Conclusion: Kai: So to wrap up this paper on "A Note On Lattice Definition of Gauge Theories With Chiral Fermions," they are showing that the main result is that the fermion partition function on any manifold W with a boundary condition L along W equals partial Y is det D/ plus W exp of negative i pi eta <ref:2611.05051#pg4>.

Mira: And they emphasize that this formula is independent of your choice of the five-manifold Y and depends only on the manifold W and the fields on it <ref:2611.05051#pg9>.

Lev: This is a big deal because it means we’s not stuck needing to pick a specific five-manifold Y to define the integral <ref:2611.05051#pg9>.

Kai: It means you can use this general definition for any W in terms of the quantum renormalized theta-angles <ref:2611.05051#pg9>.

Mira: And they provide a general formula for the partition function involving those theta-angles, which shifts them, which is how you get the effect <ref:2611.05051#pg9>.

Lev: So it’s a way to define this path integral without running into those topological obstructions that obstruct your construction <ref:2611.05051#pg9>.

Kai: It’s a nonperturbative version of anomaly inflow where the inflow from the bulk compensates for all perturbative and nonperturbative anomalies along the boundary <ref:2611.05051#pg9>.

Mira: So for our next paper, we'll be looking at how they apply this framework to define observables on manifolds, which will give us a clearer picture of what these new tools can actually do <ref:2611.05051#pg9>.

Lev: I just want to say that the final result for the partition function, ZPsi,APS equals one when m equals minus M <ref:2611.05051#pg9>.

Kai: That's a key piece of information about how the phase behaves, showing Arg of ZPsi,APS can be written as X k Arg minus i lambda k plus m minus i lambda k plus M <ref:2611.05051#pg9>.

Mira: So for anyone listening, this paper is giving us a really solid mathematical framework to define the fermion path integral on complicated spaces without relying on a specific geometry <ref:2611.05051#pg9>.

Edward Witten

School of Natural Sciences, Institute for Advanced Study

hep-lat, hep-ph, hep-th

Submitted: 2026-11-03

Updated: 2026-11-03

Comments: 33 pp

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

Importance score: 67/100

The gist: Lattice gauge theory with chiral fermions is described by interpolating a lattice gauge field over the continuum and coupling it to continuum chiral fermions to define a regularization scheme that…

Key concepts

Admissible Lattice Gauge Fields
These are lattice gauge fields whose plaquette holonomies are close to the identity (i.e., |UP - 1| is small). They allow for a smooth interpolation from the discrete lattice structure to a continuum gauge field, which is crucial for coupling with continuum chiral fermions without introducing fermion doubling problems.
Fermion Path Integral Formula
The partition function is characterized by $| ext{det } D/ + W | ext{exp}(-i ilde{ heta})$, where $D$ is the Dirac operator, $W$ is the gauge field, and $ ilde{ heta}$ relates to the eta-invariant. This formula remains valid for any anomaly-free representation and depends only on the spacetime manifold W and the gauge field.
Quantum Renormalized Theta-Angles
These angles are associated with homomorphisms from four-dimensional cobordism groups to U(1). They are essential for defining the chiral fermion path integral on an arbitrary manifold W, allowing for a general definition that does not depend on the specific five-manifold Y used in the construction.
Eta-Invariant ($\eta(D_Y)$)
The eta-invariant is computed using the spectrum of a self-adjoint operator $D_{2Y}$ (a Laplace-like operator) with generalized APS boundary conditions. This method avoids direct lattice regularization issues and allows for the computation of $\eta(D_Y)$, which is then used in the partition function formula.

Terminology

Summary

Lattice gauge theory with chiral fermions is described by interpolating a lattice gauge field over the continuum and coupling it to continuum chiral fermions to define a regularization scheme that can be implemented in principle on a computer <ref:2611. p. 705051#pg33> The paper details an approach based on interpolating an admissible lattice gauge field over the continuum and coupling it to continuum chiral fermions, which provides a method for defining the fermion path integral in four dimensions for any anomaly-free representation <ref:2611. p. 705051#pg8> This technique relies on two key facts: the well-definedness of the fermion path integral in continuum Yang-Mills theory and its ability to be implemented numerically by coupling it to an interpolated lattice gauge field <ref:2611. p. 205051#pg4> The analysis generalizes previous results, applying to any gauge group and anomaly-free fermion representation in any dimension <ref:2611. p. 205051#pg4>

The Well-Definedness of the Continuum Path Integral

The first key fact is that in continuum Yang-Mills theory with any gauge group G, the path integral for fermions in any anomaly-free representation, chiral or not, coupled to any classical background field A (and any background spacetime metric), is always well-defined <ref:2611.05051#pg4> This statement is supported by two points:

  1. In learning how to define the fermion path integral for anomaly-free theories, one also learns to determine all possible nonperturbative anomalies <ref:2611. p. 205051#pg5>. For example, in four-dimensional chiral gauge theories with connected and simply-connected gauge group, one learns that the only anomaly beyond the usual triangle anomalies is the familiar mod 2 anomaly involving the number of fermion zero-modes in an instanton field <ref:2611. p. 205051#pg5>.

  2. The path integral of an anomaly-free fermion system, chiral or not, can be described by explicit formulas that could be evaluated numerically on a computer <ref:2611. p. 305051#pg5>.

Interpolating Lattice Gauge Fields to the Continuum

The second key fact is that understanding the path integral for anomaly-free chiral fermions in the continuum helps construct a nonperturbative regularization of chiral gauge theories by coupling admissible lattice gauge fields to continuum chiral fermions <ref:2611.05051#pg4>. Admissible lattice gauge fields are those whose plaquette holonomies are close to the identity, i. p. 5e., UP − 1. These admissible lattice gauge fields can be interpolated over the continuum, allowing them to be coupled to continuum chiral fermions without introducing a fermion doubling problem <ref:2611. p. 605051#pg4>. The resulting lattice path integral is defined by an integral over admissible families of link variables weighted by the exponential of the lattice action times the output of the fermion path integral <ref:2611. p. 605051#pg4>.

The Fermion Path Integral Formula

The fermion partition function is characterized by a formula involving the absolute value of a determinant and an exponential term related to the eta-invariant <ref:2611.05051#pg5>. Specifically, for theories defined on oriented spacetimes W of dimension d = 4 with a spin structure, the formula is det D/ + W exp(−iπη(DY)) <ref:2611. p. 305051#pg4>. This formula is valid in any anomaly-free representation and depends only on W and the gauge field <ref:2611. p. 405051#pg5>.

Numerical Implementation and Zero-Mode Handling

While a straightforward lattice Monte Carlo implementation is impractical because the chiral fermion path integral lacks positivity, it is possible in principle because the lattice path integral is an integral over a compact space of finite dimension <ref:2611.05051#pg4>. The computation involves approximating det D/ + W and exp(−iπη(DY)) on a computer without running into the fermion doubling problem, as the spectrum of the self-adjoint operator D2Y can be computed numerically <ref:2611. p. 305051#pg4>. The eta-invariant of the Dirac operator DY can be extracted from its spectrum, and APS boundary conditions can be approximated numerically without running into a doubling problem <ref:2611. p. 305051#pg4>.

Generalization to Arbitrary Manifolds

The analysis generalizes previous results by applying it to any spacetime dimension and any anomaly-free fermion representation, including exotic cases like chiral fermions in two dimensions <ref:2611.05051#pg4>. The discussion also applies on an arbitrary compact Euclidean spin manifold, which is consistent with the possibility of coupling anomaly-free chiral gauge theories to gravity <ref:2611. p. 205051#pg4>. The approach allows for the definition of the chiral fermion path integral on any W by considering its cobordism structure and using quantum renormalized theta-angles <ref:2611. p. 405051#pg9>.

Obstructions to Existence

The obstructions to defining the partition function are related to topological invariants, such as gauge and gravitational instantons, which obstruct the existence of a five-manifold Y over which the structures extend <ref:2611.05051#pg9>. In four dimensions, gauge and gravitational instantons are the only obstructions of infinite order in the cobordism problem <ref:2611.05051#pg9>. For connected and simply-connected gauge groups, there are no obstructions of finite order, but this is not so for other gauge groups <ref:2611. p. 205051#pg9>. The general cobordism group that classifies obstructions to the existence of Y is a finitely generated abelian group <ref:2611. p. 705051#pg9>.

Defining the Partition Function on Cobordant Manifolds

The chiral fermion path integral on an arbitrary manifold W is defined by relating it to the partition functions on its cobordant components W = ∂Y, using a formula involving quantum renormalized theta-angles <ref:2611.05051#pg9>. This allows for a general definition of the chiral fermion path integral for any W in terms of the quantum renormalized theta-angles <ref:2611. p. 405051#pg9>. The discrete and continuous theta-angles are associated with homomorphisms from the relevant four-dimensional cobordism group to U(1) <ref:2611. p. 705051#pg9>.

Defining Observables in Topological Sectors

When the fermion partition function vanishes identically due to zero-modes predicted by an index theorem, a formula generalizing (2.18) is needed for correlation functions of operators with non-zero expectation values <ref:2611. p. 405051#pg9>. This involves defining a minimal observable O on any given manifold to generically absorb zero-modes on that manifold. p. 5 The resulting formula determines the correlation functions uniquely once all theta-angles are fixed. p. 4

Continuum Interpolations of Admissible Lattice Gauge Fields

Continuum interpolations of admissible nonabelian lattice gauge fields have been constructed in several papers, including [37–40] and recently [16]. These interpolations are gauge-invariant and ensure that the continuum gauge field is of order ϵ in lattice units. p. 5 The construction can be carried out on any d-manifold using a sequence of increasingly fine lattices, where the top dimensional cells have bounded flatness to ensure a continuum limit.

Smooth Continuum Gauge Fields

A smooth continuum gauge field can be produced by acting on the interpolated A with the Yang-Mills gradient flow for a short time. Gradient flow for a time s0 will smooth out A on a length scale √s0, yielding a smooth continuum gauge field in the same topological class as A. p. 3

Implementation on Computers

The regularization can be implemented in principle because the lattice path integral is an integral over a compact space of finite dimension. The interpolation and extension over Y are not issues, as they involve solving finitely many Landau gauge Dirichlet problems which can be solved numerically. p. 3 The computation of the eta-invariant of the Dirac operator DY is possible by computing the spectrum of D2Y without running into a doubling problem. p. 3

Consistency Checks and Zero-Mode Parity

A consistency check shows that for an anomaly-free representation, the number ζ of fermion zero-modes on any manifold W is always even, which avoids a troublesome minus sign in defining Qα Oα. This parity argument is based on the fact that nonzero eigenvalues of DY come in pairs (λ, −λ). The resulting formula for the observable O⟩Y,L is gauge-invariant and real analytic. p. 2

Final Determination of Observables

By fixing the undetermined constants, which correspond to the renormalized theta-angles—both discrete and continuous—the formula (2.28) contains enough information to define the expectation value ⟨O⟩W on each fourmanifold W. p. 6 Arbitrary chiral fermion correlation functions on W are then uniquely determined by using this information plus the fermion propagator defined in a space orthogonal to the zero-modes. p. 4

The Role of Zero-Modes in Observables

The zero-mode parts of the fermion fields contribute to the expectation value, and their contribution is determined by computing a product involving determinants of these zero-modes. The resulting expression for ⟨O⟩Y,L includes factors related to the number of zero-modes on each component Wα. p. 3 The product of 1/2m over all modes disappears after regularization and renormalization. p. 1

The Final Result for the Partition Function

The general formula for the partition function of the fermion field Ψ on Y with boundary condition L along W = ∂Y is Z(Y, L) = det D/ + W exp(−iπη). This formula describes how anomaly inflow from Y compensates for the anomaly of the chiral fermion that propagates along the boundary of Y.

The Role of Theta-Angles

The discrete and continuous theta-angles are associated with homomorphisms from the relevant four-dimensional cobordism group to U(1). The partition function Z(W) can be multiplied by an additional factor Υ(W), which shifts the theta-angles. p. 1

The Final Definition of the Path Integral

By definition of the generators, W is cobordant to some disjoint union P i niWi, and the chiral fermion path integral on W is defined as Z(W) Q i Z(-Wi) ni = Z(W) Q i Z(Wi) ni. This gives a general definition of the chiral fermion path integral for any W, in terms of the quantum renormalized theta-angles. p. 4

The Role of Zero-Modes in Correlation Functions

The expectation value ⟨O⟩Y,L is calculated by replacing O with a product of elementary fermion fields that generically suffices to absorb zero-modes on that manifold. The resulting expression for the expectation value includes terms involving determinants of the zero-modes. p. 6

The Final Formula for Correlation Functions

The final formula for the observable is ⟨O⟩Y,L = (−1)∆+N (N +1)/2 Yα det′ D/ + Wα exp(−iπηY) det[ψi(xj)] det[ψ′i(yj)]. This formula determines ⟨O⟩W on each fourmanifold W after fixing the renormalized theta-angles. p. 4

The final result is that ZΨ,APS = 1 when m = -M, and the phase of ZΨ,APS can be written as Arg(ZΨ,APS) = X k Arg −iλk + m −iλk + M. This is a general formula for the partition function of the fermion field Ψ on Y with the boundary condition L along W = ∂Y. p. 3

The Final Physical Interpretation

The final physical interpretation is that anomaly inflow from Y, described by the factor exp(−iπη), is compensating for the anomaly of the chiral fermion that propagates along the boundary of Y. This is a nonperturbative version of anomaly inflow; the inflow from the bulk compensates for all perturbative and nonperturbative anomalies along the boundary. p. 7

The Final Formula for Z(Y, L)

The general formula for the partition function of the fermion field Ψ on Y with boundary condition L along W = ∂Y is Z(Y, L) = det D/ + W exp(−iπη). This formula is independent of the choice of Y and depends only on W and the fields on W. p. 4 p. 4

The Role of APS Boundary Conditions

The APS boundary condition, when applied to DY with no zero-modes, defines a boundary state APS⟩ which is a multiple of the state omegam=0⟩ that would be the fermion ground state if m = 0. This state APS⟩ has no phase ambiguity. p. 3

The Final Result for the Determinant

For free fermions in a background field, the ratio ⟨Lomega⟩⟨omegaAPS2 can be evaluated by expanding the field Ψ in eigenmodes of the Dirac operator DW and evaluating the matrix elements on a mode by mode basis, leading to an answer that is a regularized version of the absolute value of the naive path integral det D/ + W.

The Final Formula for Z(Y, L) in Terms of Determinants

The final formula for Z(Y, L) is det D/ + W exp(−iπη). This is a well-defined absolute value det D/ + W, not the ill-defined det D/ + W. p. 3

The Role of Zero-Modes in the Determinant

The behavior of det D/ + W along the locus where a zero-mode appears is possible because the APS boundary condition fails to be self-adjoint, and this breakdown does not affect the validity of the formula. The product Z(Y, L) is completely gauge-invariant and real analytic but its factorization breaks down along the locus where Z(Y, L) = 0. p. 6

The Final Definition of the Eta-Invariant

The eta-invariant, of course, is now computed using the eigenvalues of DY with generalized APS boundary conditions. The final result is a general formula for the partition function of the fermion field Ψ on Y with boundary condition L along W = ∂Y is Z(Y, L) = det D/ + W exp(−iπη). p. 3 This formula is independent of the choice of Y and depends only on W and the fields on W.

The Final Formula for Z(Y, L) in Terms of Eta-Invariant

The final result is that ZΨ,APS = 1 when m = -M, and the phase of ZΨ,APS can be written as Arg(ZΨ,APS) = X k Arg −iλk + m −iλk + M. This is a general formula for the partition function of the fermion field Ψ on Y with boundary condition L along W = ∂Y. p. 3 p. 3

The final result is that ZΨ,APS = 1 when m = -M, and the phase of ZΨ,APS can be written as Arg(ZΨ,APS) = X k Arg −iλk + m −iλk + M. This is a general formula for the partition function of the fermion field Ψ on Y with boundary condition L along W = ∂Y.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved ability to define lattice path integrals for chiral gauge theories by coupling admissible lattice gauge fields to continuum chiral fermions using a formula derived from anomaly inflow and eta-invariants (equation 1.2).

  2. Improved ability to compute the absolute value of the fermion determinant | det D/ + W | by using a Laplace-like operator D2Y with an inherited APS boundary condition, which avoids the fermion doubling problem in numerical study.

  3. Ability to define and incorporate discrete and continuous theta-angles into lattice path integrals by relating them to homomorphisms from the cobordism group to U(1).

  4. Improved ability to characterize the chiral fermion partition function on arbitrary four-manifolds W by using a general definition involving quantum renormalized theta-angles, which does not depend on the choice of the five-manifold Y.

  5. Ability to compute non-zero expectation values of operators involving fermions on disjoint unions of manifolds by constructing minimal observables O alpha that generically absorb zero-modes, leading to a formula like equation 2.28.

  6. Capability to construct a smooth continuum interpolation procedure for admissible lattice gauge fields over any d-manifold W by solving Landau gauge Dirichlet problems iteratively on simplices and using gradient flow to suppress short-wavelength modes.

  7. Enhanced capability in numerical study of the eta-invariant η(DY) by computing the spectrum of the self-adjoint operator D2Y, which is a Laplace-like operator with no doubling problem, instead of attempting direct lattice regularization of DY.

Sources

Related papers