WorksPhysics and Astronomy
Fast, training-free simplification of spinor-helicity amplitudes
Abstract
We present a training-free method that simplifies massless spinor-helicity expressions by a greedy best-first search over on-shell identities, and test it at four, five, and six points. Applied to the color-ordered five-gluon MHV tree-level amplitude, the search derives the Parke-Taylor formula from all 1024 Feynman-diagram expressions with different polarization reference spinors and up to 228 numerator terms. On public test sets used by previous machine-learning methods, the search solves every case and generally outperforms these methods in solve rate. Its mean wall-clock time per case is 0.02-0.3 s on one CPU core. With longer searches and an alternative ranking of candidate expressions, it also solves all 21,600 cases of a new suite of larger test expressions. Each result comes with an explicit derivation and is independently checked over finite fields at on-shell kinematics. For the classes studied here, the method effectively compresses long expressions produced by identity scrambling or Feynman-diagram calculations into compact forms. This is a first step toward extracting physics from amplitudes by relating distinct compact expressions, selecting representations with preferred factors and poles, and guiding the discovery of new amplitude structures.
Introduction
Scattering amplitudes in gauge and gravity theories admit remarkably compact representations in spinor-helicity variables [1, 2, 3, 4, 5, 6, 7, 8]. A Parke–Taylor -gluon MHV amplitude, for example, is a single rational monomial in angle brackets. Diagrammatic and algebraic calculations rarely produce such forms directly, and they scale poorly with multiplicity unless color, helicity, and analytic structure are exploited to keep intermediate expressions under control [3, 4]. On-shell methods [2, 9, 10, 11] reduce this growth but still leave the task of recognizing a compact form once a large equivalent expression has been produced, whether by Feynman diagrams, on-shell recursion, numerical reconstruction [12, 13, 14, 15], or generic computer algebra. Schouten identities, momentum conservation, and momentum-squared relations connect these equivalent forms, but each identity replaces a bracket with a sum of bracket monomials and can therefore enlarge an expression as well as shrink it.
Existing spinor-helicity packages provide tools for symbolic manipulation and simplification [16, 17]. Refs. [18, 19] approach expression simplification as a machine-learning problem. The public test sets of Ref. [18], also used in Ref. [19], are generated by scrambling compact targets through random identity rewrites, and the task is to recover equivalent expressions no larger than those targets. Ref. [19] trains a supervised step-wise policy to imitate sequences of identity rewrites (hereafter SFT), while Ref. [18] trains a one-shot sequence-to-sequence simplifier that maps a scrambled expression directly to a compact form (hereafter CDS). Both solve most cases of these sets. SFT returns an explicit derivation, whereas CDS returns only the final expression.
A fixed measure of expression size can also guide a search over identities without training. In dilogarithm simplification, Dersy, Schwartz, and Zhang reported solve rates of 91% for a modified best-first search and 89% for their reinforcement-learning agent with beam search [20]. A complementary line of research explores programmatic simplifications using coding agents driven by fixed numerical objectives [21].
Denominators pose a further difficulty. Multi-particle invariants with and Dirac chains are sums of bracket monomials rather than single brackets. Such a factor can give rise to a physical pole or a spurious pole. Canceling a spurious denominator factor may require several identity rewrites.
In this paper we apply a training-free greedy best-first search (GBFS) to spinor-helicity simplification, using the same multiplicity-dependent identity menu as SFT. The search ranks candidate expressions by a bracket-count objective. It keeps multi-particle invariants and Dirac chains in the denominator unexpanded and cancels each such factor once the numerator becomes divisible by it. Each output comes with an explicit derivation from the input.
We evaluate GBFS on the public four-, five-, and six-point test sets of Ref. [18] and on a sizedepth suite of larger inputs with more scrambling rewrites, generated by our reimplementation of its generator. GBFS solves every case in the public sets, with a mean wall-clock time of 0.02–0.3 s per case on one CPU core. With longer searches and an alternative ranking of candidate expressions, it also solves the full sizedepth suite. On the five-point Yang–Mills set of Ref. [19], whose 1024 Feynman-diagram expressions differ only in the choice of polarization reference spinors and contain 8 to 228 numerator terms, GBFS reduces every expression to the Parke–Taylor form. A synthetic example with a Dirac chain in the denominator shows how the search cancels a spurious pole by building the matching factor in the numerator.
The remainder of this paper is organized as follows. § 2 establishes the notation, defines the identity menu and bracket-count objective, and introduces the test sets and success criteria. § 3 describes the search algorithm, the alternative ranking, the treatment of sum-denominators, and the finite-field check. § 4 reports solve rates, wall-clock times, and search costs on the public sets, the sizedepth suite, and the five-point Yang–Mills set, and closes with the limits of the search. § 5 summarizes the results and discusses extensions of the search.
Basic setup
We first specify the spinor-helicity variables and identities used by the search, then define the measures of expression size and the benchmark success criteria.
We follow the conventions of Ref. [6]. For each massless external momentum , we introduce two-component spinors and satisfying . We define the Lorentz-invariant angle and square brackets,
Both brackets are antisymmetric. The amplitudes and intermediate expressions considered here are rational functions of these brackets. For symbolic manipulation, we write each bracket in canonical order and absorb the signs from antisymmetry into the coefficients.
The identity menu consists of three families of massless identities, each solved for a chosen bracket. The Schouten identity expresses the linear dependence of any three two-component spinors. For four labels it reads
and the same holds for square brackets. A Schouten move solves Eq. (2) for the chosen bracket , with the spectator legs and drawn from the remaining legs. The other two families are momentum conservation,
and the momentum-squared relations , where is the sum of the momenta in a subset of the legs and is its complement. For each canonical bracket, the menu contains Schouten, momentum-conservation and momentum-squared variants. This gives 11, 19, and 32 variants at , 5, and 6, respectively. A move applies one variant either to every occurrence of the chosen bracket in the numerator or to one occurrence in a single numerator term.
An on-shell expression admits many algebraically equivalent forms. A physically simple form has a compact numerator and makes its poles and factorization channels manifest, as in the Parke–Taylor formula [1]. As a computable measure of size, we use the factored bracket count, the number of brackets in the numerator counted with multiplicity, with the monomial factor common to all numerator terms counted only once. The raw numerator bracket count instead counts the brackets of each numerator term separately. For example, has a factored count of 4 and a raw count of 6. The objective of GBFS is the ordered pair
where the entries are the factored numerator bracket count and the number of additive numerator terms. We first compare the factored numerator bracket counts. If these are equal, we prefer the expression with fewer numerator terms. The raw count is used by the alternative frontier ordering in § 3.1 and by the success criteria below. Denominator factors that are sums of bracket monomials, such as multi-particle invariants and Dirac chains, are kept unexpanded, as described in § 3.2.
Test sets and success criteria: In the public sets and the sizedepth suite, each input is generated from a compact target by random scrambling moves, where is the generation depth. The task is to recover an expression representing the same amplitude as the input and no larger than the target under the criteria below. The public sets are the four-, five- and six-point test sets released by Ref. [18] and also used for evaluation in Ref. [19]. The sizedepth suite of § 4.2 is generated by our reimplementation of the same generator. Besides the three identity families above, the generator can multiply a term by a unity factor such as , where is a rewriting of by one identity of the menu, and add a multiple of the vanishing combination . Such unity factors can introduce spurious poles. Starting from the target 0, the generator must first add a vanishing combination. The public sets include cases generated in this way, which are labeled as vanishing.
On the public sets, a case is solved when both the raw numerator bracket count and the additive term count of the output are at most those of the target. On the sizedepth suite, a case is solved when the raw numerator bracket count of the output is at most that of the target.
Method
We describe how the search explores identity rewrites, handles denominators containing sums, and checks that the resulting expressions preserve the input amplitude.
Best-first identity search
GBFS is a greedy best-first search [22, 23] over the identity menu of § 2, where each node represents an expression. After every move, the child is brought to a normal form by expanding the numerator into bracket monomials to collect like terms and canceling bracket powers common to the numerator and the monomial denominator, with sum-denominator factors handled as in § 3.2. All bracket counts refer to this normal form, allowing a single move to drastically shorten an expression when its terms collapse. At any given node, the algorithm enumerates all legal moves, scores each child by in Eq. (4), and retains new expressions on the frontier. The frontier contains all discovered expressions that have not yet been expanded. At each iteration, the node with the smallest is expanded and leaves the frontier, while the best expression found so far is retained separately. An expression whose normal form has already been reached is not added to the frontier again, which prevents cycles. Each node records the move that produced it, so the returned expression comes with an explicit derivation from the input. Every move preserves the expression’s value.
A move can enlarge the expression by replacing a bracket with a sum of bracket monomials divided by a single bracket. Cancellations during normalization can reduce it, either in the same move or after further moves. Children with a higher than their parent remain on the frontier, allowing the search to follow derivations that first enlarge the expression and then reduce it.
When the target is known, the search stops once a node satisfies the success criterion of § 2. Otherwise it returns the smallest node found when the move budget or the wall-clock time cap is exhausted (Figure 1). The public sets and the sizedepth suite are evaluated in this target-stopped mode. In contrast, target-free runs are used for the Yang–Mills set of § 4.3 and the sample of App. § 8, leaving the comparison against the target as a separate, post-run evaluation.

Figure 1. Best-first identity search, with schematic factored numerator bracket counts. Solid tree arrows are identity rewrites. Dashed links from 45 to 28 and from 42 to 40 indicate selection of the next waiting node. Neither 45 nor 42 is expanded. The separate sequence below gives the expansion order for the depicted candidates, assuming no earlier target stop. Expanding 12 generates 45, so the next waiting node is 28; expanding 21 generates 42, so the next is 40. The best expression found is retained even after its node has been expanded. The highlighted returned derivation 35 40 18 3 therefore differs from the expansion order and includes temporary growth. Edge labels denote Schouten (S), momentum conservation (), and momentum-squared () identities. The output includes the derivation and an independent, probabilistic finite-field check performed after search. A finite budget need not reach a compact target or a globally minimal expression.
The search is bounded by a move budget, which counts every evaluated move, and a wall-clock time cap. On the public sets and the sizedepth suite we use moves and a 300 s time cap per case. Because a target-stopped run ends as soon as the success criterion is met, it typically uses only a small fraction of this budget. Cases left unsolved at these limits are searched with a move budget of and, if still unsolved, with moves and a longer time cap. No run at these two move budgets reaches its time cap. Further extensions are described in § 4.4. We refer to these searches as using increased budgets.
GBFS ranks the frontier by . We denote by the variant that ranks it by the raw numerator bracket count alone. The identity menu and success criteria are unchanged. The two variants differ only in the order in which nodes are expanded, and § 4.4 shows that this order can substantially change the search cost of individual cases.
Sum-denominators
Physical amplitudes often contain multi-particle invariants with and Dirac-chain factors in their denominators. Each such factor is a sum of bracket monomials, so the denominator is in general a bracket monomial multiplied by one or more such sums . GBFS expands Dirac chains and multi-particle invariants in the numerator into brackets but keeps each denominator sum unexpanded. A move rewrites only numerator brackets and leaves these denominator sums unchanged. During normalization after each move, any that divides the numerator as a polynomial in the brackets is canceled, and the quotient becomes the new numerator.
Removing a spurious factor therefore requires a derivation that makes the numerator divisible by . Figure 2 illustrates this with a synthetic input containing a single sum-denominator factor, the six-point Dirac chain . Up to parity and an overall sign, this is the spurious denominator factor in the two-term BCFW representation of the split-helicity six-gluon NMHV amplitude [11, 24].
Figure 2. Cancellation of the Dirac-chain factor . GBFS makes the numerator divisible by and cancels the common factors, reducing the synthetic input to the simplified test expression . The full derivation is given in App. § 7.§ 1.
Starting from an input with 60 raw numerator brackets, GBFS applies a momentum-squared move followed by a momentum-conservation move, after which the numerator is divisible by . Canceling and the remaining common bracket factors reduces it to the simplified test expression . This constructed example tests cancellation of the denominator factor. Keeping sum-denominators unexpanded allows GBFS to simplify expressions with multi-particle invariants or Dirac chains in their denominators.
Randomized equality checks over finite fields
Two expressions are equal if they coincide at every on-shell kinematic point where both are defined. We test this equality at random on-shell kinematic points over finite fields, a standard technique in amplitude computations [25, 26, 27, 28, 29]. Following the polynomial view of Ref. [14], an expression is a rational function of the spinor components , , with each bracket of Eq. (1) a minor. The on-shell kinematic points form the variety defined by momentum conservation,
We sample each kinematic point directly on this variety over . The components of and are drawn independently and uniformly from . For each value of , Eq. (5) gives a linear system for and , with determinant . We discard points at which this determinant vanishes modulo and otherwise solve the two linear equations for the remaining spinor components. Since the brackets are constructed from explicit two-component spinors, they automatically satisfy the Schouten identity. Momentum conservation is imposed exactly, so the momentum-conservation and momentum-squared relations of § 2 hold as well. Two equal expressions therefore take the same value at every sampled point where both are defined modulo .
To compare and , we write each expression as a single fraction, and . Here and are the full denominators, including any sum-denominator factors of § 3.2. The cross-multiplied congruence
is evaluated in , with each rational coefficient interpreted as . We use 16 independent kinematic points for each of the two primes and . Points at which or vanishes modulo are discarded and resampled. The expressions and pass the finite-field check if Eq. (6) holds at all 32 valid points. If, for either prime, the resampling limit is reached before 16 valid points are obtained, the comparison is inconclusive. Testing whether an expression vanishes identically amounts to setting .
The probability of a false positive can be bounded using the Schwartz–Zippel lemma [30, 31]. After substituting the solutions of these linear systems and clearing denominators, the difference becomes a polynomial over in the free spinor components. For an unequal pair, suppose that is nonzero for both primes. Let be its total degree and the probability that a sampled point is valid. Conditioned on validity, the probability that the point is a root of is at most . Since the points are independent, the probability that the pair passes the finite-field check is bounded by
We use the finite-field check to test whether the outputs of GBFS, SFT, and CDS are equivalent to their respective inputs. We also checked every step of the GBFS derivations behind Table 1, Table 2, and Table 8, and all steps passed.
| public set | GBFS ( moves) solve rate | GBFS ( moves) mean s/case | SFT solve rate | SFT mean s/case | CDS () solve rate | CDS () mean s/case | |
| 4-pt | 10020 | 100.00 | 0.021 | 99.90 | 1.56 | 98.24 | 0.167 |
| 5-pt | 10270 | 100.00 | 0.088 | 99.63 | 3.77 | 95.99 | 0.164 |
| 6-pt | 9934 | 100.00 | 0.317 | 99.39 | 6.56 | 96.94 | 0.165 |
Table 1. Solve rate (%) and mean wall-clock time per case on the shallow public test sets () [18, 19]. is the number of cases in each set. All methods use the target-relative success criterion of § 2, which requires the raw numerator bracket count and the additive term count of the output to be at most those of the target. We run GBFS and SFT on a single CPU core and CDS on a single A800 GPU. Times are averaged over all cases. For SFT, the mean times over solved cases are 1.31, 2.83, and 5.14 s at four, five, and six points, respectively. GBFS stops as soon as the success criterion is met. For CDS, timing covers the full call from the input expression to the returned result.
| GBFS/+ | SFT | GBFS/+ | SFT | GBFS/+ | SFT | GBFS/+ | SFT | ||
| 4 | 15 | 100 | 98 | 100 | 97 | 100 | 97† | 99/100 | 90† |
| 4 | 20 | 100 | 99 | 100 | 99 | 99/100 | 94† | 99/100 | 89† |
| 4 | 25 | 100 | 99 | 100 | 97 | 99/100 | 96† | 99/99‡ | 85† |
| 4 | 30 | 100 | 98 | 99/100 | 98 | 100 | 93† | 100 | 90† |
| 5 | 15 | 100 | 100 | 100 | 99 | 100 | 97 | 100 | 89 |
| 5 | 20 | 100 | 99 | 100 | 99 | 100 | 96 | 100 | 90 |
| 5 | 25 | 100 | 99 | 100 | 98 | 100 | 96 | 99/100 | 92 |
| 5 | 30 | 100 | 100 | 100 | 99 | 100 | 96 | 100 | 90 |
| 6 | 15 | 100 | 100 | 100 | 99 | 100 | 97 | 100 | 95 |
| 6 | 20 | 100 | 100 | 100 | 99 | 100 | 97 | 100 | 96 |
| 6 | 25 | 100 | 99 | 100 | 99 | 100 | 99 | 100 | 96 |
| 6 | 30 | 100 | 99 | 100 | 99 | 100 | 97 | 100 | 96 |
Table 2. Solve rate (%) on the sizedepth suite, per cell. Columns are the bracket limit and rows are the generation depth . CDS was not evaluated on the full suite. The full suite, including and 250, is given in Table 8. Rates are rounded to the nearest percent, except that a value below 100 never rounds up to it. SFT uses the published checkpoints with its original inference configuration [19]. GBFS entries report the solve rates at moves and with increased budgets (GBFS+), respectively. A single value denotes identical rates. The ‡ marks the case unsolved by GBFS+ but solved by GBFSraw (§ 4.4). Both methods are scored by the raw numerator bracket count relative to the target. † The four-point SFT checkpoint encodes at most 8 distinct brackets per term and can rewrite at most 10 terms individually. The fractions of four-point inputs within both limits are 87%, 74%, 53%, and 33% for , 90, 120, and 150, respectively. The marked entries therefore reflect both SFT’s simplification procedure and the limitations of its published checkpoint.
| GBFS/+ | SFT | GBFS/+ | SFT | GBFS/+ | SFT | GBFS/+ | SFT | GBFS/+ | SFT | GBFS/+ | SFT | ||
| 15 | 4 | 100 | 98.33 | 100 | 97.33 | 100 | 96.67† | 99.67/100 | 89.67† | 99.67/100 | 79.33† | 99.67/100 | 72.33† |
| 20 | 4 | 100 | 99.00 | 100 | 97.67 | 99.67/100 | 94.33† | 99.67/100 | 88.67† | 99.67/100 | 79.00† | 99.67/100 | 63.33† |
| 25 | 4 | 100 | 98.67 | 100 | 96.67 | 99.33/100 | 96.00† | 99.00/99.67‡ | 85.00† | 99.67/100 | 77.00† | 99.33/100 | 64.00† |
| 30 | 4 | 100 | 98.33 | 99.67/100 | 98.00 | 100 | 92.67† | 100 | 90.00† | 100 | 75.33† | 100 | 59.67† |
| 15 | 5 | 100 | 100 | 100 | 99.67 | 100 | 96.67 | 100 | 88.67 | 100 | 84.67 | 99.33/100 | 71.00 |
| 20 | 5 | 100 | 99.33 | 100 | 99.00 | 100 | 95.67 | 100 | 90.00 | 100 | 77.33 | 99.00/100 | 62.00 |
| 25 | 5 | 100 | 99.67 | 100 | 97.67 | 100 | 96.33 | 99.67/100 | 91.67 | 99.33/100 | 75.67 | 98.67/100 | 59.67 |
| 30 | 5 | 100 | 100 | 100 | 99.67 | 100 | 96.33 | 100 | 89.67 | 99.67/100 | 74.67 | 98.33/100 | 58.33 |
| 15 | 6 | 100 | 100 | 100 | 99.00 | 100 | 96.67 | 100 | 95.33 | 100 | 89.33 | 99.33/100 | 79.33 |
| 20 | 6 | 100 | 100 | 100 | 99.00 | 100 | 97.33 | 100 | 96.00 | 99.67/100 | 84.67 | 98.00/100 | 74.67 |
| 25 | 6 | 100 | 99.67 | 100 | 99.33 | 100 | 99.67 | 100 | 95.67 | 99.00/100 | 86.67 | 97.00/100 | 68.33 |
| 30 | 6 | 100 | 99.00 | 100 | 99.00 | 100 | 97.00 | 100 | 96.33 | 99.00/100 | 85.33 | 93.67/100 | 69.00 |
Table 8. Solve rate (%) on the full size×depth suite. The table extends Table 2 to and 250 and reports rates to two decimal places. The evaluation protocol and notation are as in Table 2. † At and 250, the fractions of four-point inputs within both checkpoint limits are 14% and 9%, respectively.
Results
We evaluate GBFS on the public sets, the sizedepth suite, and the five-point Yang–Mills set. On the public sets, we compare it with SFT, a supervised step-wise policy over the same identity menu [19], and CDS, a sequence-to-sequence simplifier [18]. On the sizedepth suite, we compare it with SFT. Each comparison uses identical inputs and the same success criterion.
Public sets
Table 1 reports the solve rate and the mean wall-clock time per case on the shallow public test sets (; § 2). At a budget of moves per case, GBFS solves all four-, five-, and six-point expressions, with no case reaching the 300 s time cap. On a single CPU core, the mean wall-clock time per case is 0.021, 0.088, and 0.317 s, respectively.
We use the inference settings of Ref. [19] for SFT and CDS, including their original stopping rules. SFT uses a 30 s wall-clock limit for each backtracking attempt, so its solve rate can vary with machine speed. In our runs, SFT leaves 10, 38, and 61 cases unsolved at four, five, and six points, compared with 11, 36, and 63 in Ref. [19]. After inference, we bring the saved SFT outputs and CDS candidates to the same normal form as the outputs of GBFS and apply the common numerator-based success criterion of § 2. A CDS case is solved if the greedy output matches the target exactly or an equivalent candidate from beam search with meets this criterion.
The target is not a proven minimum, so an expression that meets the success criterion may still be simplified further. In a target-free run on 296 cases labeled as non-vanishing in the five-point public set, GBFS found expressions with fewer numerator brackets than their targets in 32 cases. Appendix § 8 lists these cases.
The sizedepth suite
The inputs in the public sets are generated with at most three scrambling moves, [18]. To study larger and more deeply scrambled inputs, we evaluate GBFS and SFT on the sizedepth suite, which varies the generation depth and the bracket limit . Each input is obtained from its target by random scrambling moves. During generation, a move is rejected if the resulting expression has a raw numerator bracket count greater than .
The suite contains cases in each cell. SFT uses its publicly released checkpoints, and the GBFS move budget and time cap are those of § 3.
At the budget of moves, GBFS achieves a 100% solve rate in 46 out of 72 cells and 93.7–99.7% in the remaining 26, with no case reaching the time cap. In contrast, at fixed and , the absolute decrease in the SFT solve rate is 21%–42% as increases from 60 to 250. At fixed and , the absolute variation across is at most 13% (Figure 4).
Figure 4. Solve rate versus bracket limit . SFT is shown separately for each generation depth ( cases per point), while GBFS is averaged over the four values of ( cases per point) at a budget of moves.
With increased budgets, GBFS solves 21,599 of the 21,600 cases, and GBFSraw solves the remaining one. We compare the two frontier orderings on this case in § 4.4.
Table 4 and Figure 3 show a steep increase in solve rate with the move budget. Table 3 shows that, marginalized over , the GBFS solve rate is nearly flat in , while the median number of evaluated moves varies by less than 20% from to .
Figure 3. GBFS solve rate versus the move budget. In each panel, colored curves show the sizedepth suite of Table 4, one for each bracket limit . The dashed gray curve shows the public set at the same . The vertical line marks the budget of moves used in Table 1 and Table 2. Open circles mark the budget at which a curve reaches a 100% solve rate, and arrows indicate that a curve continues beyond the plotted range. The four-point curve, marked ‡, does not reach 100% within moves.
| 4 | 99.83 | 99.78 | 99.56 | 99.94 | 1,997 | 2,000 | 1,996 | 1,987 |
| 5 | 99.89 | 99.83 | 99.61 | 99.67 | 7,052 | 7,287 | 7,987 | 8,154 |
| 6 | 99.89 | 99.61 | 99.33 | 98.78 | 10,437 | 11,494 | 11,874 | 12,166 |
Table 3. GBFS solve rate and median search cost versus generation depth , marginalized over the six bracket limits , with cases per cell. The evaluation uses the budget of moves and a 300 s time cap. The search cost is the number of evaluated moves, and its median is computed over cases solved within this budget.
| 4 | 60 | 44.33 | 98.08 | 99.33 | 100 | 100 | 100 | 100 |
| 4 | 90 | 22.25 | 97.33 | 98.92 | 99.83 | 99.92 | 99.92 | 100 |
| 4 | 120 | 12.00 | 95.17 | 97.92 | 99.50 | 99.75 | 99.92 | 100 |
| 4 | 150 | 7.33 | 93.67 | 98.00 | 99.42 | 99.58 | 99.67 | 99.67 |
| 4 | 200 | 3.92 | 93.75 | 98.58 | 99.67 | 99.75 | 99.75 | 99.75 |
| 4 | 250 | 3.42 | 87.67 | 98.42 | 99.50 | 99.67 | 99.67 | 99.75 |
| 4 | all | 15.54 | 94.28 | 98.53 | 99.65 | 99.78 | 99.82 | 99.86 |
| 5 | 60 | 14.67 | 91.83 | 99.25 | 100 | 100 | 100 | 100 |
| 5 | 90 | 2.50 | 62.25 | 95.00 | 99.75 | 100 | 100 | 100 |
| 5 | 120 | 0.83 | 29.67 | 79.58 | 99.50 | 100 | 100 | 100 |
| 5 | 150 | 0.25 | 11.58 | 62.75 | 99.50 | 99.92 | 100 | 100 |
| 5 | 200 | 0.00 | 4.08 | 25.17 | 95.42 | 99.75 | 99.92 | 100 |
| 5 | 250 | 0.00 | 1.50 | 12.50 | 86.08 | 98.83 | 99.75 | 99.83 |
| 5 | all | 3.04 | 33.49 | 62.38 | 96.71 | 99.75 | 99.94 | 99.97 |
| 6 | 60 | 8.92 | 73.00 | 96.92 | 100 | 100 | 100 | 100 |
| 6 | 90 | 3.08 | 35.33 | 84.50 | 100 | 100 | 100 | 100 |
| 6 | 120 | 0.92 | 11.25 | 49.25 | 99.42 | 100 | 100 | 100 |
| 6 | 150 | 0.33 | 4.17 | 26.42 | 97.67 | 100 | 100 | 100 |
| 6 | 200 | 0.08 | 1.08 | 6.75 | 79.75 | 99.42 | 100 | 100 |
| 6 | 250 | 0.00 | 0.00 | 2.08 | 56.25 | 97.00 | 100 | 100 |
| 6 | all | 2.22 | 20.81 | 44.32 | 88.85 | 99.40 | 100 | 100 |
Table 4. GBFS solve rate versus the move budget on the sizedepth suite, marginalized over the four generation depths , with cases in each row and 21,600 cases in total. Each case is run with a budget of moves, stopping once the success criterion is met. The recorded move counts determine the solve rates at the smaller budgets listed here. At moves, the six-point solve rate falls from 96.9% at to 2.1% at , while the four-point rate falls only from 99.3% to 98.4%.
Feynman-diagram gluon amplitudes at five points
The public sets and the sizedepth suite are constructed by scrambling compact target expressions. We next apply GBFS to expressions obtained directly from Feynman diagrams.
Refs. [18, 19] introduced this five-gluon partial-amplitude set, which contains 1024 expressions differing only in the assignment of polarization reference spinors to external legs. The expressions contain between 8 and 228 numerator terms, with a mean of about 90. We use the common-sign release, in which 512 of the original expressions have been multiplied by to align their overall sign conventions. Every expression in this release is equal on shell to the same Parke–Taylor form [1, 7],
a single term of 3 numerator brackets, or 7 factors total including denominator multiplicity. Changing the reference spinors changes the form and size of the expression but not the amplitude, so the set varies the input size at a fixed target.
In target-free mode, GBFS reduces all 1024 cases to Eq. (8) with a common budget of moves per case. Each run exhausts this budget without target-based stopping. The following search costs count moves up to the first discovery of the returned compact form, rather than the full run. The median search cost to reach this target is moves, with a mean of .
The largest case, case 1023, has 228 numerator terms and 2736 brackets. GBFS resolves this expression via an 18-move derivation with 13 momentum-conservation, 4 Schouten, and 1 momentum-squared moves, and the raw numerator bracket count decreases at every step (). Finding this derivation required evaluating moves and took 49 s on a single CPU core or 7.6 s using 24 threads.
Table 5 details this expression alongside two other representative cases. The timing columns come from separate runs that stop once the expression reaches the three numerator brackets of Eq. (8). Stopping early does not change the frontier ordering, so these runs reach the target after the same number of moves as the target-free runs.
| case | terms | derivation | search cost | 1 core | 24 threads |
| 1023 (largest) | 228 | 18 | 49 s | 7.6 s | |
| 990 (most demanding in [19]) | 178 | 34 | 74 s | 12 s | |
| 935 (most demanding here) | 150 | 108 | 400 s | 73 s |
Table 5. Three representative cases from the five-point Yang–Mills set. Derivation length denotes the number of moves in the successful derivation, whereas the search cost reflects the total number of moves evaluated to find that derivation.
The median search cost increases with the initial numerator term count, from moves in the 1–25 term bin to moves in the 201–228 bin (Table 6). Within each bin, however, search costs vary by factors of approximately 7 to 54. Across bins, case 935 (150 terms) requires about 24 times as many moves as the largest input, case 1023 (228 terms), and its derivation is 6 times as long.
| terms | solved | min | median | max | spread | |
| 1–25 | 47 | 47/47 | 2.7k | 8.0k | 19.6k | 7× |
| 26–50 | 138 | 138/138 | 7.4k | 24.5k | 337.6k | 46× |
| 51–75 | 206 | 206/206 | 18.7k | 56.2k | 368.8k | 20× |
| 76–100 | 245 | 245/245 | 29.6k | 97.2k | 977.2k | 33× |
| 101–125 | 184 | 184/184 | 39.5k | 198.8k | 1,390.7k | 35× |
| 126–150 | 117 | 117/117 | 57.9k | 238.4k | 3,112.2k | 54× |
| 151–175 | 50 | 50/50 | 76.0k | 351.8k | 2,072.5k | 27× |
| 176–200 | 31 | 31/31 | 96.1k | 359.6k | 1,455.9k | 15× |
| 201–228 | 6 | 6/6 | 131.1k | 477.6k | 2,394.5k | 18× |
| all | 1024 | 1024/1024 | 2.7k | 96.4k | 3.11M | 1166× |
Table 6. Solve rate and search cost for the 1024 five-point Yang–Mills cases, binned by initial numerator term count. Spread is the ratio of the highest to the lowest search cost within each bin.
Limits of the search
With increased budgets, GBFS leaves one four-point case (, ) unsolved. Its target raw bracket count is 36. GBFS reaches a factored bracket count of 41 after 14,494 moves, but its minimum raw count remains 44 after moves. solves the same case in moves, returning an expression with 33 raw brackets. For another four-point case (, ), GBFS requires moves, compared with 4,738 for .
At a fixed budget of moves, the solve rate of GBFS exceeds that of by approximately 0.06% in absolute terms. With increased budgets, solves the entire suite. The two orderings trade performance across cases. We retain the factored count as the default because it matches the objective in Eq. (4) and solves slightly more cases at the fixed budget. The raw ordering provides an alternative for the hard tail. These results come from separate runs of the two orderings and do not test a rule that switches between them automatically. More generally, different counting rules favor different expression forms. Designing objectives that favor specified factors or pole structures is a direction for future components beyond the compression algorithm studied here.
In the exceptional -move search above, many generated candidates remain waiting for expansion, making the frontier costly to hold in memory. A candidate is generated when an identity move produces it and expanded when it is selected to generate further candidates. The frontier could be stored on disk to reduce RAM use at the cost of additional input/output operations. It could also be bounded in size, which may change which solutions are found within a fixed budget.
Conclusion and outlook
Recovering a compact form from a scrambled spinor-helicity expression can be posed as a search over sequences of identity moves. GBFS solves every case in the public sets at a budget of moves, exceeding the solve rates of SFT and CDS. With increased budgets, it solves 21,599 of the 21,600 cases in the sizedepth suite, and GBFSraw solves the remaining one. In target-free mode, GBFS reduces all 1024 Feynman-diagram expressions of the five-point Yang–Mills set to the Parke–Taylor form. Keeping sum-denominators unexpanded allows a spurious denominator factor to cancel once the numerator becomes divisible by it. Each output comes with an explicit derivation and is compared with its input using the finite-field check. The six-point example in App. § 7.§ 2 gives a 13-move derivation reducing 115 raw numerator brackets to 1, illustrating how the returned trace makes a multi-step simplification explicit.
Target-free search can also improve upon the supplied reference forms. Among 296 sampled five-point cases, it finds 32 outputs with fewer raw numerator brackets than their targets (App. § 8), showing that these targets need not be the simplest representations.
The cost of expanding a node depends on the number of legal moves, which is determined by the identity menu and the brackets and terms in the numerator. The menu is generated for arbitrary and contains 11 identity variants per canonical bracket at four points and 32 at six. Evaluating the search at seven and eight points is a natural next step. Exceptionally long searches can also accumulate a large frontier, as illustrated by the four-point case in § 4.4. Hybrid guidance [32, 33, 34, 35, 36] offers one way to reduce the total search time and memory usage. A learned model could rank frontier nodes to reduce the number of expansions needed to reach a compact expression.
A further extension concerns Dirac chains with momentum sums, such as , and multi-particle invariants, which are currently expanded into brackets when they appear in a numerator. Identities that act on these objects directly would let the search keep their compact form.
In a preliminary six-point NMHV test, target-free search reduces the expanded numerator from 126 to 32 terms while retaining the input denominator. A systematic study of six- to eight-point NMHV tree amplitudes and, subsequently, rational parts of one-loop amplitudes would test the method on more complex compact targets.
A longer-term goal is a single simplifier for massless and massive spinor-helicity amplitudes. Massive spinors carry little-group indices and satisfy mass-dependent identities [6], and on-shell recursion provides compact massive amplitudes against which such a simplifier can be tested [37, 38, 39]. Extending the symbolic representation to massive spinors and adding the corresponding identities to the menu would allow the same search algorithm to be applied to massive amplitudes.
We call the compressor presented here AmpZip, the first component of AmpLens, a developing toolkit for exposing the analytic structure of scattering amplitudes. Planned components will address extraction of prescribed prefactors, residue extraction at specified poles, and machine-learning-informed search between compact expressions for selected high-value targets. These capabilities remain future work beyond the compression results reported here. The code, input data and evaluation commands used in this paper will be released through GitHub with the arXiv submission.
Acknowledgments
We thank David Shih, Lance Dixon, and Jiayin Gu for helpful conversations, and David Shih and Aurélien Dersy for providing the five-point Yang–Mills expressions used in this work. Y.F. is supported by the National Natural Science Foundation of China under Grant No. 12375091 at Fudan University. Z.L. is supported by the Department of Energy under Grant No. DE-SC0011842 and in part by a Sloan Research Fellowship from the Alfred P. Sloan Foundation at the University of Minnesota. This work was assisted, in particular in the programming and in running the computations, by Anthropic’s Claude and OpenAI’s GPT models.
Example derivations
This appendix gives explicit derivations for the sum-denominator input of Figure 2 and for a six-point case from the sizedepth suite. For each derivation, a table lists the identity applied at each step and the resulting bracket and term counts, and the intermediate expressions follow the table.
Cancellation of a Dirac-chain factor
The synthetic input illustrated in Figure 2 is
with
GBFS reaches after evaluating 2,672 moves.
| step | identity | rewrite | raw | factored | terms |
| 0 | — | input | 60 | 49 | 12 |
| momentum-squared | 30 | 25 | 6 | ||
| 2 | momentum conservation | 1 | 1 | 1 |
Table 10.
Starred steps replace one bracket factor in a single numerator term. Unstarred steps replace every occurrence of the indicated bracket in the numerator. Underlined brackets are replaced in the next step.
(R1) with
Step 1
with .
Step 2
A six-point case from the size×depth suite
We consider six-point case 144 from the size×depth suite, with generation depth and bracket limit . Of the 30 scrambling moves, 5 multiply a term by a unity factor and 6 add a vanishing combination. The input has 23 terms and a raw numerator bracket count of 115, whereas the target has a single numerator bracket. GBFS recovers the target in a 13-move derivation, found after evaluating 31,108 moves. Table 7 compares the three methods on this input.
| step | identity | rewrite | raw | factored | terms |
| 0 | — | input | 115 | 115 | 23 |
| 1{sup}*` | Schouten | 100 | 100 | 20 | |
| 2{sup}*` | Schouten | 90 | 90 | 18 | |
| 3 | Schouten | 68 | 68 | 17 | |
| 4 | Schouten | 64 | 64 | 16 | |
| 5{sup}*` | Schouten | 60 | 60 | 15 | |
| 6{sup}*` | Schouten | 56 | 56 | 14 | |
| 7{sup}*` | Schouten | 52 | 52 | 13 | |
| 8 | momentum conservation | 48 | 48 | 12 | |
| 9{sup}*` | momentum conservation | 44 | 44 | 11 | |
| 10 | Schouten | 28 | 28 | 7 | |
| 11 | momentum conservation | 15 | 15 | 5 | |
| 12 | momentum conservation | 9 | 7 | 3 | |
| 13 | Schouten | 1 | 1 | 1 |
Table 7.
The target is
Step 0
with .
Step 1
with .
Step 2
with .
Step 3
with .
Step 4
with .
Step 5
with .
Step 6
with .
Step 7
with .
Step 8
with .
Step 9
with .
Step 10
with .
Step 11
with .
Step 12
Step 13
| method | solved | raw brackets | finite-field check |
| GBFS | yes | 1 | passed |
| SFT | no | 102 | passed |
| CDS () | no | 4 | failed |
Table 7 (PDF p. 21). Comparison of GBFS, SFT, and CDS on six-point case 144 with and . The target has a single numerator bracket. The last column gives the result of the finite-field check of each output against the input. CDS uses the inference settings described in § 4.1, and its row shows the highest-scoring of the 20 beam candidates. None of these candidates passes the finite-field check.
Cases simplified below the target
The target of a case is not a proven minimum. In target-free mode, we ran GBFS on a sample of 296 cases labeled as non-vanishing in the five-point public set, with a budget of moves per case. No case reached the 60 s time cap. For 264 cases, the output has the same raw numerator bracket count as the target. For the other 32, it has 2 to 14 fewer brackets.
In 25 of these 32 cases, the output has the same denominator as the target. In 6 cases, the denominator of the output is a proper divisor of that of the target. In the remaining case 288, the target vanishes by momentum conservation and the output is 0.
The following list gives the target and search output for each case. Case numbers refer to the subset labeled as non-vanishing in the five-point public set. Each heading gives the raw numerator bracket counts of the target and the output, together with the change from the target to the output. Cases are ordered by . In case 4706, the output has a nonzero constant numerator and therefore a bracket count of 0.
case 9042 21 7 brackets ():
target:
GBFS:
case 288 12 0 brackets ():
target:
GBFS: 0 case 4450 30 → 20 brackets ():
target:
GBFS:
case 7144 27 → 18 brackets ():
target:
GBFS:
case 1930 24 → 16 brackets ():
target:
GBFS:
case 1134 12 → 5 brackets ():
target:
GBFS:
case 2282 12 → 5 brackets ():
target:
GBFS:
case 6970 21 → 14 brackets ():
target:
GBFS:
case 814 9 → 3 brackets ():
target:
GBFS:
case 1468 18 → 12 brackets ():
target:
GBFS:
case 3732 12 → 6 brackets ():
target:
GBFS:
case 6124 9 → 3 brackets ():
target:
GBFS:
case 7034 18 → 12 brackets ():
target:
GBFS:
case 7624 9 → 3 brackets ():
target:
GBFS:
case 750 10 → 5 brackets ():
target:
GBFS:
case 1102 15 → 10 brackets ():
target:
GBFS:
case 2584 15 → 10 brackets ():
target:
GBFS:
case 7272 15 → 10 brackets ():
target:
GBFS:
case 7560 15 → 10 brackets ():
target:
GBFS:
case 3988 12 → 8 brackets ():
target:
GBFS:
case 4706 4 → 0 brackets ():
target:
GBFS:
case 4976 12 → 8 brackets ():
target:
GBFS:
case 5470 12 8 brackets ():
target:
GBFS:
case 6444 6 2 brackets ():
target:
GBFS:
case 7240 12 8 brackets ():
target:
GBFS:
case 7464 6 2 brackets ():
target:
GBFS:
case 878 6 3 brackets ():
target:
GBFS:
case 1372 6 3 brackets ():
target:
GBFS:
case 6220 6 3 brackets ():
target:
GBFS:
case 7798 6 3 brackets ():
target:
GBFS:
case 8850 9 6 brackets ():
target:
GBFS:
case 1166 20 18 brackets ():
target:
GBFS:
Full results on the sizedepth suite
Table 8 gives the solve rates on all 72 cells of the size×depth suite, and Table 9 gives the number of moves used per solved case.
| median | p90 | p99 | max | median | p90 | p99 | max | median | p90 | p99 | max | |
| 60 | 1,085 | 1,996 | 6,990 | 29,268 | 2,330 | 4,737 | 9,610 | 19,371 | 3,526 | 7,173 | 12,964 | 16,738 |
| 90 | 1,522 | 2,864 | 11,467 | 165,546 | 4,223 | 8,005 | 15,095 | 34,738 | 6,159 | 11,649 | 19,118 | 26,605 |
| 120 | 1,931 | 3,659 | 15,819 | 150,663 | 6,732 | 12,740 | 25,300 | 46,269 | 10,150 | 17,030 | 28,134 | 38,468 |
| 150 | 2,259 | 4,007 | 15,032 | 558,124 | 8,846 | 16,096 | 25,190 | 77,236 | 13,486 | 22,342 | 33,574 | 41,734 |
| 200 | 2,699 | 4,473 | 11,018 | 2,958,920 | 13,633 | 24,272 | 40,404 | 122,420 | 20,964 | 34,433 | 53,985 | 77,971 |
| 250 | 3,060 | 5,237 | 15,622 | 28,971,674 | 18,268 | 33,576 | 63,393 | 2,892,468 | 28,490 | 46,918 | 71,438 | 89,189 |
| all | 1,996 | 4,127 | 13,010 | 28,971,674 | 7,610 | 20,873 | 39,943 | 2,892,468 | 11,532 | 31,045 | 54,053 | 89,189 |
Table 9. Moves per case solved by GBFS on the size×depth suite, including searches with increased budgets. Statistics are marginalized over the four generation depths , with cases for each . For , the statistics cover 1,199 solved cases. These costs exclude unsuccessful attempts and separate raw-ordering searches.
References
- [1]S.J. Parke and T.R. Taylor, An Amplitude for n Gluon Scattering, Phys. Rev. Lett. 56 (1986) 2459.pubmed.ncbi.nlm.nih.gov/10032998
- [2]R. Britto, F. Cachazo, B. Feng and E. Witten, Direct proof of tree-level recursion relation in Yang-Mills theory, Phys. Rev. Lett. 94 (2005) 181602 [hep-th/0501052].DOI
- [3]L.J. Dixon, Calculating scattering amplitudes efficiently, in QCD and Beyond: Proceedings of TASI ’95, 1996 [hep-ph/9601359].arxiv.org/abs/hep-ph/9601359
- [4]L.J. Dixon, A brief introduction to modern amplitude methods, in Theoretical Advanced Study Institute in Elementary Particle Physics: Particle Physics: The Higgs Boson and Beyond, pp. 31–67, 2014, DOI [1310.5353].DOI
- [5]H. Elvang and Y.-t. Huang, Scattering Amplitudes in Gauge Theory and Gravity, Cambridge University Press (2015).
- [6]N. Arkani-Hamed, T.-C. Huang and Y.-t. Huang, Scattering amplitudes for all masses and spins, JHEP 11 (2021) 070 [1709.04891].
- [7]M.L. Mangano and S.J. Parke, Multiparton Amplitudes in Gauge Theories, Phys. Rept. 200 (1991) 301 [hep-th/0509223].
- [8]C. Cheung, TASI lectures on scattering amplitudes, in Theoretical Advanced Study Institute in Elementary Particle Physics: Anticipating the Next Discoveries in Particle Physics, pp. 571–623, 2018, DOI [1708.03872].arxiv.org/abs/1708.03872
- [9]E. Witten, Perturbative gauge theory as a string theory in twistor space, Commun. Math. Phys. 252 (2004) 189 [hep-th/0312171].
- [10]F. Cachazo, P. Svrcek and E. Witten, MHV vertices and tree amplitudes in gauge theory, JHEP 09 (2004) 006 [hep-th/0403047].DOI
- [11]R. Britto, F. Cachazo and B. Feng, New recursion relations for tree amplitudes of gluons, Nucl. Phys. B 715 (2005) 499 [hep-th/0412308].
- [12]G. De Laurentis and D. Maître, Extracting analytical one-loop amplitudes from numerical evaluations, JHEP 07 (2019) 123 [1904.04067].DOI
- [13]G. De Laurentis and D. Maître, Two-Loop Five-Parton Leading-Colour Finite Remainders in the Spinor-Helicity Formalism, JHEP 02 (2021) 016 [2010.14525].DOI
- [14]G. De Laurentis and B. Page, Ansätze for scattering amplitudes from p-adic numbers and algebraic geometry, JHEP 12 (2022) 140 [2203.04269].DOI
- [15]O. Barrera, A. Dersy, R. Husain, M.D. Schwartz and X. Zhang, Analytic regression of Feynman integrals from high-precision numerical sampling, JHEP 01 (2026) 014 [2507.17815].arxiv.org/abs/2507.17815
- [16]D. Maître and P. Mastrolia, S@M, a Mathematica Implementation of the Spinor-Helicity Formalism, Comput. Phys. Commun. 179 (2008) 501 [0710.5559].DOI
- [17]M. Accettulli Huber, SpinorHelicity4D: a Mathematica toolbox for the four-dimensional spinor-helicity formalism, 2304.01589.
- [18]C. Cheung, A. Dersy and M.D. Schwartz, Learning the simplicity of scattering amplitudes, SciPost Phys. 18 (2025) 040 [2408.04720].DOI
- [19]D. Shih, Learning to Unscramble: Simplifying Symbolic Expressions via Self-Supervised Oracle Trajectories, 2603.11164.arxiv.org/abs/2603.11164
- [20]A. Dersy, M.D. Schwartz and X. Zhang, Simplifying polylogarithms with machine learning, Int. J. Data Sci. Math. Sci. 01 (2023) 135 [2206.04115].arxiv.org/abs/2206.04115
- [21]Y. Gu and S. Krippendorf, Scattering Amplitudes as Programs: Self-Evolving Search for Theory and Event Generation, 2607.21629.arxiv.org/abs/2607.21629
- [22]J.E. Doran and D. Michie, Experiments with the Graph Traverser program, Proc. R. Soc. Lond. A 294 (1966) 235.DOI
- [23]J. Pearl, Heuristics: Intelligent Search Strategies for Computer Problem Solving, Addison-Wesley, Reading, MA (1984).
- [24]A. Hodges, Eliminating spurious poles from gauge-theoretic amplitudes, JHEP 05 (2013) 135 [0905.1473].arxiv.org/abs/0905.1473
- [25]A. von Manteuffel and R.M. Schabinger, A novel approach to integration by parts reduction, Phys. Lett. B 744 (2015) 101 [1406.4513].DOI
- [26]T. Peraro, Scattering amplitudes over finite fields and multivariate functional reconstruction, JHEP 12 (2016) 030 [1608.01902].DOI
- [27]T. Peraro, FiniteFlow: multivariate functional reconstruction using finite fields and dataflow graphs, JHEP 07 (2019) 031 [1905.08019].DOI
- [28]J. Klappert and F. Lange, Reconstructing rational functions with FireFly, Comput. Phys. Commun. 247 (2020) 106951 [1904.00009].DOI
- [29]G. De Laurentis, Lips: p-adic and singular phase space, J. Phys. Conf. Ser. 3206 (2026) 012118 [2305.14075].arxiv.org/abs/2305.14075
- [30]J.T. Schwartz, Fast probabilistic algorithms for verification of polynomial identities, J. ACM 27 (1980) 701.DOI
- [31]R. Zippel, Probabilistic algorithms for sparse polynomials, in Symbolic and Algebraic Computation (EURO SAM ’79), vol. 72 of Lecture Notes in Computer Science, pp. 216–226, Springer, 1979, DOI.DOI
- [32]M. von Hippel and M. Wilhelm, Refining Integration-by-Parts Reduction of Feynman Integrals with Machine Learning, JHEP 05 (2025) 185 [2502.05121].arxiv.org/abs/2502.05121
- [33]Z.-Y. Song, T.-Z. Yang, Q.-H. Cao, M.-x. Luo and H.X. Zhu, Explainable AI-assisted optimization for Feynman integral reduction, JHEP 06 (2026) 225 [2502.09544].arxiv.org/abs/2502.09544
- [34]M. Zeng, Reinforcement learning and metaheuristics for Feynman-integral reduction, Phys. Rev. D 112 (2025) 114051 [2504.16045].arxiv.org/abs/2504.16045
- [35]D. Shih, Learning to Unscramble Feynman Loop Integrals with SAILIR, 2604.05034.arxiv.org/abs/2604.05034
- [36]J. Berman, F. Charton, A. Luna, M. Wilhelm and M. Zeng, Efficient AI-Inspired Reduction of Feynman Integrals via Tube Seeding, 2606.10698.arxiv.org/abs/2606.10698
- [37]Y. Ema, T. Gao, W. Ke, Z. Liu, K.-F. Lyu and I. Mahbub, Momentum shift and on-shell constructible massive amplitudes, Phys. Rev. D 110 (2024) 105003 [2403.15538].arxiv.org/abs/2403.15538
- [38]Y. Ema, T. Gao, W. Ke, Z. Liu, K.-F. Lyu and I. Mahbub, Momentum shift and on-shell recursion relation for electroweak theory, Phys. Rev. D 110 (2024) 105002 [2407.14587].arxiv.org/abs/2407.14587
- [39]Y. Ema, T. Gao, W. Ke, Z. Liu and I. Mahbub, On-shell recursion relations for higher-spin Compton amplitudes, JHEP 01 (2026) 069 [2506.02106].arxiv.org/abs/2506.02106


