{"id":"6f6af06a-86aa-4426-8dbf-3bc76941307f","arxiv_id":"1909.02655","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":3.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Residue calculus and partial fraction expansion are applied to invert the nabla Laplace transform, reducing the inverse contour integral to sums over poles or to transform-pair tables.","lead":"This paper gives two ways to undo the nabla Laplace transform, a tool used in discrete fractional calculus: one based on complex residues, the other on splitting functions into simple fractions. The methods turn an integral that is hard to compute into sums or table lookups, with worked examples for ordinary and fractional-order functions.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equation (7) is false as stated: it omits the residue at infinity and fails for F(s)=1, an admissible nabla Laplace transform.","rationale":"The paper's central contribution is the pair of residue formulas (4) and (7) plus the partial-fraction table. Formula (4) is sound for rational F(s) with a contour inside the ROC, since the only pole of the integrand inside c is s=1. Equation (7) is the advertised shortcut that avoids taking high-order derivatives at s=1 by using the poles of F(s), and that is where the risk is concentrated. The fractional-order limitations in Section 3.3 are acknowledged by the authors, so they are not a hidden flaw. The missing residue at infinity is not acknowledged anywhere. The counterexample F(s)=1 is not exotic; it is row 1 of the paper's own Table 1. A reader following Eq. (7) would obtain 0 instead of the delta sequence. This does not destroy the paper, because adding Res∞ or a degree hypothesis fixes the formula and the worked examples are unaffected; however, it means the central claim of 'two alternative formulae' is not correct as written. The reader's conditional verdict remains appropriate, but the specific justification should be updated from 'unproved exterior-residue formula' to a concrete missing residue-at-infinity term.","tokens_in":9309,"tokens_out":13584,"duration_ms":146592,"concrete_test":"Evaluate Eq. (7) with F(s)=1, a=0, k=1: the right-hand side has no finite outside poles and gives 0, whereas direct evaluation of the defining contour integral Eq. (3) gives 1. This single check settles the concern. Independently, re-derive Eq. (7) from the residue theorem including the point at infinity; the derivation should produce f(k)=Σ_outside Res + Res∞, confirming the missing term.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Equation (7) is printed as an unconditional alternative to Eq. (4): when the finite poles of F(s)(1-s)^{-k+a} outside c are {s_n}, then f(k)=Σ_n Res. This is not correct in general. By the residue theorem applied on the Riemann sphere, the clockwise contour integral in Eq. (3) evaluates to 2πj(Σ_outside Res + Res∞), where Res∞ is the residue at infinity. Thus Eq. (7) is valid only when Res∞=0. A concrete counterexample is F(s)=1, which is Table 1 row 1 and corresponds to the delta sequence f(a+1)=1, f(k)=0 for k>a+1. For k=a+1, H(s)=F(s)(1-s)^{-k+a}=1/(1-s), whose only finite pole is s=1 inside c; there are no finite poles outside. Equation (7) therefore predicts f(a+1)=0, while direct evaluation of Eq. (3) gives 1. The paper's Example 1 avoids this problem because its F(s) decays like O(s^{-3}), so Res∞=0; but the formula as written fails for a basic admissible transform. The fix is to add Res∞ to the right-hand side of Eq. (7), or to state a hypothesis such as F(s)=O(s^{-1}) for k=a+1 (and F(s)=O(1) for k≥a+2) so that Res∞ vanishes. This is an internal inconsistency with Eq. (3), not merely a limitation of fractional-order functions, and Section 3.3 does not acknowledge it.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops two analytical methods for inverting the nabla Laplace transform defined in Eq. (1) and (3). The first method is a residue calculation: Eq. (4) expresses f(k) as the negative sum of residues of F(s)(1-s)^{-k+a} at finite poles inside the integration contour, and Eq. (7) claims to express f(k) as the sum of residues at finite poles outside the contour. The second method is partial fraction expansion: Eq. (9) and Eq. (11) give the inverse transform for rational F(s), and Eq. (12)-(13) extend this to certain fractional-order terms. A table of 16 transform pairs is provided, and two examples (one rational, one fractional-order) are worked out. Section 3.3 discusses limitations for fractional-order functions.","tokens_in":9598,"tokens_out":7560,"duration_ms":72964,"significance":"If corrected, the paper would offer a convenient exact inversion toolkit for nabla Laplace transforms, with the table and worked examples being useful for practitioners. The formula in Eq. (4) and the partial fraction formulas in Eq. (9) and Eq. (11) are consistent with the defining contour integral for the proper rational examples in Section 4, and the paper explicitly checks both residue formulas and the partial fraction method against the same result in Example 1. However, Eq. (7) is false as stated because it omits the residue at infinity, and this is an internal inconsistency with Eq. (3), not merely a fractional-order limitation. The fractional-order discussion in Section 3.3 is candid, but the residue-at-infinity gap is not acknowledged there. The novelty claim is modest: the methods are classical residue and partial fraction techniques adapted to a transform that is closely related to the Z-transform.","major_comments":[{"comment":"The formula f(k)=sum_n Res[F(s)(1-s)^{-k+a}, s_n] is not correct as an unconditional statement because it omits the residue at infinity of the integrand H(s)=F(s)(1-s)^{-k+a}. By the residue theorem on the Riemann sphere, the contour integral in Eq. (3) equals 2*pi*j*(sum of residues outside c plus the residue at infinity), so Eq. (7) needs an additional term Res[H(s), infinity] on the right-hand side, or a decay hypothesis on F(s) that makes that residue vanish. A concrete counterexample is F(s)=1, which is Table 1, row 1 and corresponds to the delta sequence f(a+1)=1, f(k)=0 for k>a+1. For k=a+1, H(s)=1/(1-s), whose only finite pole is s=1 inside the contour; there are no finite poles outside, so Eq. (7) predicts f(a+1)=0, whereas direct evaluation of Eq. (3) gives 1. The paper's note that s=1 cannot be a pole of F(s) for finite f(k) does not fix the problem, because the pole of H(s) at s=1 comes from the factor (1-s)^{-k+a}, not from F(s).","section":"Section 3.1, Eq. (7)"},{"comment":"The limitations discussion correctly identifies difficulties with multi-valued fractional-order functions such as F(s)=1/(s^alpha-lambda) for irrational alpha, but it does not mention the residue-at-infinity problem in Eq. (7). This problem occurs even for elementary rational transforms, so it is an internal inconsistency with the defining integral (3) rather than a limitation of fractional calculus. The section should state explicitly that Eq. (7) is valid only when the residue at infinity of F(s)(1-s)^{-k+a} is zero, or it should be amended to include the missing residue-at-infinity term.","section":"Section 3.3"}],"minor_comments":[{"comment":"The phrase 'residual calculation method' should be 'residue calculation method' throughout the paper, including the title, abstract, and Section 3.1.","section":"General"},{"comment":"The 16 transform pairs in Table 1 are asserted without derivation; many can be verified directly from the definition in Eq. (1), and a short explanation or a reference to the property list in [14] would make the table more self-contained and easier to check.","section":"Table 1"},{"comment":"There are several incomplete or garbled sentences in Section 3.3, including 'some inverse transform of fractional order polynomial' and 'The proposed two methods In other words, it is an opportunity and challenge to handle with such complicated irrational F(s)'; these need to be rewritten into complete sentences.","section":"Section 3.3"},{"comment":"The expression 's2 = 0.3^{10/7} e^{j20*pi*i/7}' is ambiguous because the index i is not defined and the statement that the number of such poles is infinite is not obviously consistent with the displayed formula, which has finite periodicity in i; please clarify the branch structure.","section":"Example 2"},{"comment":"The inversion formula in Eq. (3) should explicitly state the required analyticity assumptions on F(s) on the contour c and the precise sense in which the contour 'locates in the convergent region', since these conditions are inherited by all subsequent formulas.","section":"Section 2, Eq. (3)"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear Colleague,\n\nThe short version: the paper has a useful table and a correct inside-pole residue formula, but the outside-pole formula (7) is false as stated because it drops the residue at infinity. That is not a nitpick: take F(s)=1, which is Table 1 row 1; for k=a+1, the integrand is 1/(1-s), there are no finite poles outside the clockwise contour, yet the inverse transform is 1. Equation (7) gives 0. The missing term is Res∞, and the formula works only when that residue vanishes (e.g., F(s)=O(1/s) as s→∞).\n\nWhat the paper does well: the inside-pole formula (4) is correct, and Table 1's 16 transform pairs look right—they are standard Z-transform pairs under z^{-1}=1-s, which the authors explicitly state in Section 3.2. Example 1 is worked out correctly by both methods. The partial-fraction extension (11) is standard but cleanly presented. The discussion of fractional-order limitations in Section 3.3 is honest.\n\nThe soft spots: aside from (7), the claimed novelty is inflated. Because of the explicit mapping g(k)=f(k+1), z^{-1}=1-s, the inverse nabla transform is just the inverse Z transform; residue and partial-fraction methods for that are classical. The 'first time' statement in the conclusions won't hold up. Some table entries are asserted without derivation, though they are verifiable. The paper also leans heavily on self-cited prior work [14], [16] for the contour integral; that is acceptable background but worth noting.\n\nIn sum: a methods/pairs paper with a real error in one of its two headline formulas. The error is fixable—add the residue at infinity or state the hypothesis—and the table alone might justify a revised version. As it stands, I would not cite it for Eq. (7), but I would send it to a referee because the rest is correct and the table is practical. After revision, it could be a serviceable reference.\n\nHope this helps.","headline":"Useful table and correct inside-pole formula, but Eq. (7) is false as stated (missing residue at infinity) and the novelty claim is overstated.","tokens_in":10133,"tokens_out":4635,"would_cite":false,"duration_ms":44584,"reading_group":"no","serious_thinker":"no","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["39A12","44A10","26A33"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper derives two exact analytic routes for inverting the nabla Laplace transform: residue summation and partial-fraction table lookup, with a sign rule tied to the clockwise contour.","keywords":["nabla Laplace transform","inverse transform","residue method","partial fraction expansion","discrete fractional calculus","Mittag-Leffler function","transform-pair table"],"falsifier":"Take a rational $F(s)$ whose inverse is known independently, for example $F(s)=1/(s-2)$, whose inverse should be $(-1)^{k-a}$; evaluate the defining contour integral (3) numerically on a small clockwise circle around $s=1$ for $k-a=1,2,3$ and compare signs—any sign slip in the residue formulas appears immediately. For a fractional-order case, compute the series definition of the claimed Mittag-Leffler output for the first few $k$ and compare against direct numerical evaluation of the inverse transform; a wrong table pair produces a finite discrepancy.","tokens_in":9109,"feed_emoji":"🧮","tokens_out":10625,"duration_ms":95107,"temperature":0.7,"pith_summary":"This paper establishes that the inverse nabla Laplace transform—the operation that recovers a discrete-time causal sequence $f(k)$ from its transform $F(s)$—can be computed exactly without evaluating the defining contour integral. The first route is a residue method: $f(k)$ is the negative of the sum of residues of $F(s)(1-s)^{-k+a}$ at poles inside a clockwise contour, or the plain sum at poles outside it. The second route is partial-fraction expansion: once $F(s)$ is decomposed into simple rational pieces, each piece maps to a power of $(1-s_i)^{-1}$, and a table of sixteen transform pairs supplies exact inverses for common sequences, including fractional-order Mittag-Leffler forms. The methods give exact inverses for rational $F(s)$ and for fractional-order functions that can be rearranged into table entries, and the paper states explicitly where they stop working, namely for branch-cut or multi-valued cases such as $1/(s^{\\alpha}-\\lambda)$ with irrational $\\alpha$.","feed_headline":"Two exact methods invert the nabla Laplace transform","feed_subtitle":"Residue sums and partial fractions replace hard contour integrals for discrete-time signals.","key_machinery":"The load-bearing object is the kernel $s \\leftrightarrow (1-s)^{-k+a}$ built into the inversion integral (3). In the residue route, the residue theorem applied to the meromorphic integrand $F(s)(1-s)^{-k+a}$ separates poles inside the clockwise contour (giving a minus sign) from poles outside (giving no sign), with higher-order poles handled by the standard derivative formula. In the partial-fraction route, the same kernel becomes the elementary inverse $N_a^{-1}\\{1/(s-\\lambda)\\}=1/(1-\\lambda)^{k-a}$, and multiple poles produce polynomial prefactors $(k-a)^{i-1}$. The table of sixteen transform pairs is the practical machine: it packages the algebra so that a user only decomposes $F(s)$ and reads off the sequence, including fractional-order cases via the discrete Mittag-Leffler function.","core_discovery":"The central claim is that the inversion formula $N_a^{-1}\\{F(s)\\} = \\frac{1}{2\\pi j}\\oint_c F(s)(1-s)^{-k+a}\\,ds$ can be evaluated analytically by residue calculus. Because the contour $c$ winds clockwise about $s=1$, the inside-pole formula carries a minus sign, $f(k)=-\\sum_m \\operatorname{Res}[F(s)(1-s)^{-k+a},s_m]$, while the outside-pole formula carries a plus sign, $f(k)=\\sum_n \\operatorname{Res}[\\cdots,s_n]$. Equivalently, when $F(s)=\\sum_i r_i/(s-s_i)$, with multiple-pole terms included, the inverse is $f(k)=\\sum_i r_i/(1-s_i)^{k-a}$ plus related multiple-pole terms involving $(k-a)^{i-1}$. This turns inversion into algebra plus table lookup: the sixteen transform pairs in Table 1 cover powers, exponentials, sinusoids, and discrete Mittag-Leffler functions, and the examples show a rational case and a fractional-order case solved exactly.","pith_inferences":["Because $z^{-1}=1-s$ links the nabla and Z transforms, the table approach can likely port any finite-order rational Z-domain inversion table into nabla form; the residue identities here should coincide with inverse-Z-transform results up to orientation.","For rational $\\alpha=p/q$, the function $1/(s^{p/q}-\\lambda)$ has finitely many pole branches on the appropriate Riemann surface, so collecting residues over all branches may yield a closed-form discrete Mittag-Leffler inverse even though the irrational case is left open.","A direct numerical test of the branch-cut case $1/(\\sqrt{s}-\\lambda)$ by numerical integration of the inversion integral around $s=1$ would reveal whether purely polar residue formulas miss a branch-cut contribution.","The paper's own construction suggests a generative recipe: applying convolution, time-scaling, and frequency-differentiation properties to the existing sixteen pairs produces new pairs, so the table is expandable rather than closed."],"forward_implications":["For rational $F(s)$ with known poles, exact inversion becomes a residue count plus table lookup, removing the approximation error of numerical inversion.","The clockwise contour flips the usual inverse-Z-transform sign pattern: inside-pole residues enter with a minus sign and outside-pole residues with a plus sign.","The sixteen transform pairs give closed-form inverses for common causal sequences and for fractional-order Mittag-Leffler entries, usable in solving nabla fractional difference equations.","Multiple poles are handled by explicit polynomial factors, so repeated-root transfer functions remain exactly invertible by the same formulas."],"supporting_citations":[{"why":"defines the nabla Laplace transform of a causal sequence, the object being inverted","marker":"[9]"},{"why":"supplies the inverse contour integral (3) that both methods evaluate, and the $z^{-1}=1-s$ variable relation","marker":"[16]"},{"why":"provides the residue theorem used to turn the contour integral into sums over poles","marker":"[17]"},{"why":"gives the fundamental transform properties and initial-value theorem used to build the table and exclude $s=1$ as a pole","marker":"[14]"},{"why":"supplies the region-of-convergence framework and discrete fractional calculus background the inversion relies on","marker":"[12]"}],"fun_headline_variants":["Nabla Laplace inversion: exact via residues and partial fractions","Closed-form inverse nabla Laplace: two exact methods","Exact inverse nabla Laplace via residue sums and tables","Residue calculus gives exact inverse nabla Laplace","Inverting nabla Laplace: exact residue and partial fraction methods"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the inverse contour integral (3) is a valid inversion formula for the chosen $F(s)$ and that all singularities of $F(s)(1-s)^{-k+a}$ are isolated poles cleanly separated inside or outside the contour; branch-point or multi-valued cases such as $1/(s^{\\alpha}-\\lambda)$ with irrational $\\alpha$ violate this and are acknowledged as not covered.","fun_headline_variants_meta":{"raw":{"variants":["Nabla Laplace inversion: exact via residues and partial fractions","Closed-form inverse nabla Laplace: two exact methods","Exact inverse nabla Laplace via residue sums and tables","Residue calculus gives exact inverse nabla Laplace","Inverting nabla Laplace: exact residue and partial fraction methods"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000542,"raw_usage":{"total_tokens":2547,"prompt_tokens":843,"completion_tokens":1704,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":459,"completion_tokens_details":{"reasoning_tokens":1623}},"tokens_in":459,"tokens_out":1704,"duration_ms":12002,"temperature":1.0,"reasoning_tokens":1623,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:34:10.905641+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a rational $F(s)$ whose inverse is known independently, for example $F(s)=1/(s-2)$, whose inverse should be $(-1)^{k-a}$; evaluate the defining contour integral (3) numerically on a small clockwise circle around $s=1$ for $k-a=1,2,3$ and compare signs—any sign slip in the residue formulas appears immediately. For a fractional-order case, compute the series definition of the claimed Mittag-Leffler output for the first few $k$ and compare against direct numerical evaluation of the inverse transform; a wrong table pair produces a finite discrepancy.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"defines the nabla Laplace transform of a causal sequence, the object being inverted"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the inverse contour integral (3) that both methods evaluate, and the $z^{-1}=1-s$ variable relation"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the residue theorem used to turn the contour integral into sums over poles"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"gives the fundamental transform properties and initial-value theorem used to build the table and exclude $s=1$ as a pole"},{"cited_title":"Goodrich, and A","cited_arxiv_id":null,"evidence_quote":"supplies the region-of-convergence framework and discrete fractional calculus background the inversion relies on"}],"review_version":1}