{"id":"affab995-53e3-48c0-85a2-93b7094e1afc","arxiv_id":"2412.14641","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Three new infinite families of permutation pentanomials over F_{q^2}, q=2^m, are proved to be linearly equivalent to power maps, unifying 14 known sporadic examples.","lead":"This paper shows that 14 of 17 recently discovered five-term permutation polynomials over even-characteristic finite fields are secretly power maps in disguise, conjugated by linear maps. It also folds these 14 families into three new infinite classes of permutation pentanomials with explicit if-and-only-if conditions.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The r_A table in Theorem 2 is self-contradictory for i≡1,j≡0 and the r_C table omits i≡j≡0, so the central even-m classification is undefined for exactly the parameter ranges that include the paper's f_C examples.","rationale":"The paper's contribution is the claim that Theorems 2–4 give exact conditions for the three pentanomial classes and explain 14 of the 17 families. For that claim to hold, the quantities r_A,r_B,r_C in the even-m case must be well-defined and correctly computed. The printed (10) is internally inconsistent in the i≡1,j≡0 case, and (14) is incomplete in the i≡j≡0 case; both are exactly the kind of statement-level defect that prevents the central classification from being applied. The reader's verdict identified the unproved identities (23)–(24) as the weakest assumption; I agree that those identities are load-bearing, but the r-table contradiction is more decisive because it makes the theorem false as a mathematical statement, independently of whether the identities are true. Since the underlying construction may be correctable and the reader already made the verdict conditional, I do not change the verdict, but I would emphasize that the r-tables must be fixed before the claim can be accepted. The proposed CAS check directly settles the contradiction.","tokens_in":16796,"tokens_out":18989,"duration_ms":146386,"concrete_test":"Use a CAS over F2 to compute, for several small i,j in each parity class, gcd(N_•,H_•) and read off the exponent of Q(x)=x^2+x+1; compare with (10), (12), (14). In particular, take (i,j)=(1,2) for A: the table gives both r_A=0 and r_A=3, so compute whether gcd(N_A,H_A)=1 or Q(x)^3; and take (i,j)=(2,2) for C to see whether r_C=0, as the missing even-even entry would require. If these checks show the tables are wrong, regenerate the correct r_• tables and re-run Lemma 9's no-root analysis.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Detailed concern: the central theorem statements are not well-defined. In (10), for i≡1 (mod 2), j≡0 (mod 2), the first line sets r_A=0 while the last two lines set r_A=Q1+1 (if Q1<Q2) or r_A=Q2 (if Q1≥Q2); since Q1,Q2 are powers of 2, one of those two cases always holds, so r_A would have to be both 0 and a positive integer. The r_C table (14) has no entry for i≡j≡0 (mod 2), yet the three f_C families in Table 1 (families 6, 14, 16) all have i,j even, so the condition gcd(Q1+Q2+1−2r_C,q+1)=1 is undefined for them. This is not a cosmetic issue: Lemma 8 claims gcd(N_•,H_•)=Q(x)^{r_•} with r_• given by these tables, and Lemma 9 uses the same values to decide whether H_• has roots in µ_{q+1}; the proof of Theorems 2–4 therefore cannot be correct as written. A secondary concern is that identities (23) and (24), on which Lemma 10's ramification classification rests, are stated with 'Here it is easy to check' and no derivation; they are load-bearing and should be verified independently.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies permutation pentanomials over F_{q^2}, q=2^m, of the form f(x)=x^t H(x^{q-1}). Its main results are Theorems 2–4, which give necessary and sufficient conditions for three general families f_A, f_B, f_C to permute F_{q^2}, and which explain 14 of the 17 families previously found by Zhang et al. The proof strategy is to write f_•(x)=x^{Q_1+Q_2+1}H_•(x), introduce the rational function g_•=N_•/H_• on the unit circle, use two asserted polynomial identities (23) and (24) to control ramification, and then deduce that g_• is a composition of degree-one rational maps with a power map. Lemma 11 then lifts this to a linear equivalence of f_• to either a bivariate power map (m odd) or a univariate power map (m even). Theorem 1 gives an explicit self-contained proof for one of the 14 families using linear permutations.","tokens_in":17041,"tokens_out":11495,"duration_ms":93659,"significance":"If the main results are correct, the paper provides a genuinely simpler and more conceptual explanation for a substantial part of the existing classification of permutation pentanomials, and it extends those families to three infinite classes with clean permutation criteria. The reduction to linear equivalence with power maps is a useful observation, and the paper is honest in crediting the technique to [9]. However, the current version is not reliable as written: the r_A and r_C tables in the central theorems are defective, making the even-m classification undefined for some parameter ranges that include the paper's own examples, and the load-bearing identities (23) and (24) are asserted without proof. Because these issues are local and likely fixable, the paper warrants a major revision rather than rejection.","major_comments":[{"comment":"The table defining r_A is not well-defined. For i≡1, j≡0 (mod 2), the first line assigns r_A=0, while the last two lines assign r_A=Q_1+1 if Q_1<Q_2 and r_A=Q_2 if Q_1≥Q_2. Since Q_1 and Q_2 are powers of 2, one of those inequalities always holds, so r_A is assigned two different values. Moreover, there is no row for i≡j≡1 (mod 2), even though this is precisely the parity pattern of all f_A families appearing in Table 1 (families 2, 9, 11). Lemma 8's claim that gcd(N_A,H_A)=Q(x)^{r_A}, and Lemma 9's use of r_A to determine roots in µ_{q+1}, therefore do not make sense for these parameters. The proof of Case A in Lemma 8 says 'the other cases are similar' but does not resolve this contradiction. The table must be corrected and every parity case checked.","section":"§3, Eq. (10), Theorem 2"},{"comment":"The table for r_C has no entry for i≡j≡0 (mod 2). This is not a cosmetic omission: the three f_C families in Table 1 (families 6, 14, 16) all have both i and j even. For those families, the condition gcd(Q_1+Q_2+1−2r_C, q+1)=1 in Theorem 4(ii) is undefined. The proof of Case C in Lemma 8 itself derives HC(ω)≠0 when i≡j≡0, so r_C=0 in that case, which suggests the intended entry is 0; but the published table omits it, and Lemma 8 does not state the resulting r_C value. This must be fixed before the theorem can be evaluated.","section":"§3, Eq. (14), Theorem 4"},{"comment":"The two identities N'_•(x)H_•(x)+N_•(x)H'_•(x)=Q(x)^{Q_1+Q_2} and Q(g_•(x))H_•(x)^2=Q(x)^{Q_1+Q_2+1} are the foundation of the entire ramification argument. They imply Eqs. (31) and (32), which are what make Lemma 10 and hence Lemma 11 possible. They are introduced with the phrase 'Here it is easy to check' and no derivation. These identities are not self-evident from (16)–(21), and a failure for some (i,j) would invalidate the classification of g_• and the linear equivalence of f_•. The authors should provide a full verification, preferably as a separate lemma, for all three families; this is a load-bearing step, not a routine detail that can be left to the reader.","section":"§3, Eqs. (23) and (24)"}],"minor_comments":[{"comment":"The displayed expansion of H_A(x+ω) omits the constant term H_A(ω). This omission is likely responsible for the inconsistency in the r_A table, since for i≡j≡1 the constant term is nonzero and gives r_A=0, while the displayed expansion without the constant term suggests a positive valuation. The expansion should be written in full, including the constant term.","section":"§3, proof of Lemma 8, Case A"},{"comment":"The sentence 'Case 1: t is odd' should read 'Case 1: m is odd', since t=Q_1+Q_2+1 is always odd for powers of 2. This is a typo but could confuse the reader.","section":"§4, Lemma 11, Case 1"},{"comment":"In the even-m case, the text writes P_2(x)=x^{Q_1+Q_2+1+m_•(q−1)}; here m_• is undefined and should be r_• from the relevant theorem. The proof also jumps from the permutation condition of this power map to the two displayed gcds without showing the standard equivalence gcd(E,q^2−1)=1 ⇔ gcd(Q_1+Q_2+1,q−1)=gcd(Q_1+Q_2+1−2r_•,q+1)=1. Adding this one-line derivation would make the proof complete.","section":"§4, Proofs of Theorems 2–4"},{"comment":"The identity N_•(x)=x^{Q_1+Q_2+1}H_•(x^{-1}) is stated as 'easy to check'; it is in fact immediate from (16)–(21), but a one-line verification for each family would remove any doubt and would also make the subsequent use of this identity clearer.","section":"§3, Eq. (22)"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of a finite-fields / permutation-polynomial journal. I saw no sign of circularity: the permutation properties of the target power maps are independently known, and the linear equivalence is meant to be derived from the ramification analysis. The main problems are concentrated in the r_• tables and the unproved identities (23)–(24). I would be willing to consider a revised version in which the tables are corrected, all parity cases are verified, and the two identities are fully proved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The core observation is good: 14 of the 17 permutation pentanomial families from Zhang et al. are disguised power maps, and the paper turns them into three infinite classes with explicit permutation conditions. That is a real reorganization of a small but active subfield, and it gives people a recipe for spotting linear equivalence to power maps. The paper is also honest that the technique comes from Ding and Zieve [9]; it does not oversell its own originality.\n\nWhat is actually new: Theorems 2–4, which state if-and-only-if permutation conditions for the three classes f_A, f_B, f_C and identify their power-map equivalents. Theorem 1, the worked example for family 17, is clean and checkable. I spot-checked the factorization in Theorem 1 and the permutation behavior in one family; the arithmetic works.\n\nThe soft spots are serious, but they look correctable rather than fatal. The r_A table in (10) is internally inconsistent: for i odd and j even it sets r_A = 0 in the first line and also sets r_A to a positive value in the later lines, with one of those later cases always applying since Q1 and Q2 are powers of 2. The r_C table in (14) omits the case i and j both even, which is exactly the case for the f_C families in Table 1. Table 1 also reproduces conditions for families 6 and 14 that conflict with the necessary gcd(t, q−1) = 1 condition. Because Lemma 8, Lemma 9, and the final theorems all depend on these r values, the even-m part of the main theorem is not well-defined as printed.\n\nOn top of that, equations (23) and (24) are load-bearing: the entire ramification argument in Lemma 10 rests on them. They are asserted with “Here it is easy to check” and no derivation. I have not found a counterexample, and the spot-checks I did are consistent, but these identities need to be verified independently, not left to the reader.\n\nIf the tables are fixed and (23)–(24) check out, the paper is a useful contribution: the argument via ramification and linear equivalence is coherent, and the credit to [9] is appropriate. As it stands, the printed theorem statements are not reliable.\n\nThis paper deserves a serious referee, but the referee should demand a major revision. The authors need to correct the r_A and r_C tables, reconcile Table 1 with their own conditions, and either prove or explicitly verify (23)–(24). Once those are done, I would be comfortable citing it.","headline":"A genuinely useful structural observation about permutation pentanomials, but the central theorem statements as printed are broken and need a thorough revision before the paper can be trusted.","tokens_in":17652,"tokens_out":2347,"would_cite":false,"duration_ms":21834,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11T06"],"pacs":[],"model":"deepseek-v4-flash","headline":"One gcd condition decides whether each of three five-term polynomial families permutes the field.","keywords":["permutation polynomials","pentanomials","finite fields of characteristic 2","linear equivalence","rational functions over finite fields","ramification","gcd conditions","power maps"],"falsifier":"Expand equations (23) and (24) symbolically for several values of $i,j$: any failure of either identity for the stated conditions would break Lemma 10 and hence Theorems 2–4. Alternatively, evaluate $f_A,f_B,f_C$ on all elements of $\\mathbb{F}_{q^2}$ for small $q=2^m$ and small $i,j$; a single instance satisfying the gcd conditions that is not a bijection would refute the 'if' direction, and a permutation outside the conditions would refute the 'only if' direction.","tokens_in":16456,"feed_emoji":"🔁","tokens_out":9636,"duration_ms":67236,"temperature":0.7,"pith_summary":"This paper proves that three general families of five-term polynomials over $\\mathbb{F}_{q^2}$, with $q=2^m$, are permutation polynomials exactly when a small set of gcd conditions holds. Each pentanomial is written as $x^{Q_1+Q_2+1}H(x^{q-1})$ with $Q_1=2^i$, $Q_2=2^j$, and the proof shows that the accompanying rational function $g=N/H$ is, up to fractional linear changes, a power map. When $m$ is odd, each family is linearly equivalent to the bivariate map $(u^{Q_1+Q_2+1}, v^{Q_1+Q_2+1})$; when $m$ is even, it is linearly equivalent to $x^{Q_1+Q_2+1+r(q-1)}$ for an explicit correction $r$ depending on the parities and sizes of $Q_1,Q_2$. This gives a uniform reason why 14 of the 17 pentanomial families found by a recent computer search are permutations, and it extends those 14 families to arbitrary parameter values satisfying the conditions.","feed_headline":"One gcd condition decides these pentanomial permutations","feed_subtitle":"Three five-term families over F_{q^2} become one power map, explaining 14 earlier permutation families.","key_machinery":"The load-bearing object is the rational function $g_\\bullet(x)=N_\\bullet(x)/H_\\bullet(x)$ attached to $f_\\bullet(x)=x^{Q_1+Q_2+1}H_\\bullet(x^{q-1})$: the permutation criterion reduces $f_\\bullet$ to the requirements $\\gcd(Q_1+Q_2+1,q-1)=1$, $H_\\bullet$ has no roots on the unit circle $\\mu_{q+1}$, and $g_\\bullet$ permutes $\\mu_{q+1}$. The proof shows that $g_\\bullet$ obeys two identities involving $Q(x)=x^2+x+1$: $N_\\bullet'(x)H_\\bullet(x)+N_\\bullet(x)H_\\bullet'(x)=Q(x)^{Q_1+Q_2}$ and $Q(g_\\bullet(x))H_\\bullet(x)^2=Q(x)^{Q_1+Q_2+1}$. These identities force the only ramification points of $g_\\bullet$ to be the two roots of $Q(x)$, and its only branch points to be the images of those roots, so a classification of rational functions with two totally ramified points gives $g_\\bullet=\\eta^{-1}\\circ x^{Q_1+Q_2+1}\\circ\\sigma$ for odd $m$, and $g_\\bullet=\\rho^{-1}\\circ x^{Q_1+Q_2+1-2r_\\bullet}\\circ\\sigma$ for even $m$, with $\\eta,\\sigma,\\rho$ fractional linear maps of the appropriate type on $\\mu_{q+1}$. Expanding these compositions yields the claimed linear equivalence of $f_\\bullet$ to a power map.","core_discovery":"The central discovery is that the permutation behaviour of the three pentanomial classes is controlled by the single integer $Q_1+Q_2+1$ together with a correction term $r_\\bullet$ determined by the parities of $i,j$ and the relative sizes of $Q_1,Q_2$. For $q=2^m$: if $m$ is odd, $f_\\bullet$ permutes $\\mathbb{F}_{q^2}$ if and only if $\\gcd(Q_1+Q_2+1,q-1)=1$ plus a parity condition on $i,j$ that depends on the family, and then $f_\\bullet$ is linearly equivalent to coordinatewise exponentiation on $\\mathbb{F}_q^2$; if $m$ is even, $f_\\bullet$ permutes if and only if $\\gcd(Q_1+Q_2+1,q-1)=\\gcd(Q_1+Q_2+1-2r_\\bullet,q+1)=1$, and then $f_\\bullet$ is linearly equivalent to the single power map $x^{Q_1+Q_2+1+r_\\bullet(q-1)}$. Because linear equivalence preserves the permutation property, these equivalences yield necessary and sufficient conditions, not merely sufficient ones, and the three classes include the 14 starred families from the earlier search as special cases.","pith_inferences":["If the two asserted identities hold for all $i,j$, the same two-ramification-point strategy may also accommodate the three remaining families from the earlier search, should they fit a similar shape after a suitable correction term.","The argument is specific to characteristic two because it uses $Q(x)=x^2+x+1$ and the structure of $\\mu_{q+1}$; an odd-characteristic analogue would need a different quadratic and a different normalizing family of fractional linear maps.","A direct computational test over small $q$ and small $i,j$ could check the 'only if' direction exhaustively, which the earlier paper did not claim."],"forward_implications":["For each of the three families, testing whether a pentanomial is a permutation is reduced to one or two gcd computations, and the conditions are both necessary and sufficient.","Fourteen of the seventeen pentanomial families from the earlier search are recovered as special cases, so their individual case-by-case proofs can be replaced by the uniform argument.","Every polynomial in these families is linearly equivalent to a monomial map, so any function-theoretic property invariant under linear equivalence is shared with the corresponding power map.","The gcd conditions are satisfied for infinitely many exponent pairs $(i,j)$, so the three classes go well beyond the original finite search range $4\\le t<100$."],"supporting_citations":[{"why":"supplies the 17 pentanomial families whose 14 starred cases this paper explains and extends.","marker":"[30]"},{"why":"provides the classification of rational functions with two totally ramified points used in Lemma 10, and originates the linear-equivalence idea.","marker":"[9]"},{"why":"gives the criterion that $x^r h(x^{(q-1)/d})$ permutes the field, used to reduce $f_\\bullet$ to conditions on $H_\\bullet$ and $g_\\bullet$.","marker":"[32]"},{"why":"supplies Lemmas 6 and 7 on fractional linear maps that permute or map $\\mu_{q+1}$ to the projective line, used to normalize $\\eta,\\sigma,\\rho$.","marker":"[33]"},{"why":"provides the preliminary lemma used to pass from permutation of $f_\\bullet$ to permutation of the associated rational function on $\\mu_{q+1}$.","marker":"[8]"},{"why":"carries out the same kind of linear-equivalence expansion in Lemma 15, which the proof of Lemma 11 follows.","marker":"[12]"}],"fun_headline_variants":["Three pentanomial classes distilled to one gcd test","One gcd condition unlocks 14 permutation families","Pentanomial permutations reduce to a power map","Single gcd criterion explains 14 pentanomial families","Even and odd cases: one gcd condition decides"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole reduction to power maps rests on two polynomial identities (equations 23 and 24) that are asserted with the phrase 'here it is easy to check' but not derived; if either identity failed for some pair $(i,j)$, the claimed linear equivalence and the theorems would not follow.","fun_headline_variants_meta":{"raw":{"variants":["Three pentanomial classes distilled to one gcd test","One gcd condition unlocks 14 permutation families","Pentanomial permutations reduce to a power map","Single gcd criterion explains 14 pentanomial families","Even and odd cases: one gcd condition decides"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000479,"raw_usage":{"total_tokens":2364,"prompt_tokens":930,"completion_tokens":1434,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":546,"completion_tokens_details":{"reasoning_tokens":1362}},"tokens_in":546,"tokens_out":1434,"duration_ms":9306,"temperature":1.0,"reasoning_tokens":1362,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T12:08:04.061590+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Expand equations (23) and (24) symbolically for several values of $i,j$: any failure of either identity for the stated conditions would break Lemma 10 and hence Theorems 2–4. Alternatively, evaluate $f_A,f_B,f_C$ on all elements of $\\mathbb{F}_{q^2}$ for small $q=2^m$ and small $i,j$; a single instance satisfying the gcd conditions that is not a bijection would refute the 'if' direction, and a permutation outside the conditions would refute the 'only if' direction.","supporting_citations":[{"cited_title":"Zhang, L","cited_arxiv_id":null,"evidence_quote":"supplies the 17 pentanomial families whose 14 starred cases this paper explains and extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the classification of rational functions with two totally ramified points used in Lemma 10, and originates the linear-equivalence idea."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"gives the criterion that $x^r h(x^{(q-1)/d})$ permutes the field, used to reduce $f_\\bullet$ to conditions on $H_\\bullet$ and $g_\\bullet$."},{"cited_title":"Constructing permutation polynomials using generalized Redei functions","cited_arxiv_id":"2305.06322","evidence_quote":"provides the preliminary lemma used to pass from permutation of $f_\\bullet$ to permutation of the associated rational function on $\\mu_{q+1}$."}],"review_version":1}