ProbXiv
sign in

SYMMETRIC DECOMPOSITIONS AND THE VERONESE CONSTRUCTION

Combinatorics · math.CO · posed by Katharina Jochemko · open

1 attempt · 1 machine check

Statement

Let h(t)h(t) be a polynomial of degree s≤ds \le d with nonnegative coefficients that satisfies Properties (H) and (S). Does Urd+1h(t)U_r^{d+1}h(t) have an interlacing symmetric decomposition for all r≥d+1r \ge d+1 (possibly even for r≥max⁡{s,d+1−s}r \ge \max\{s, d+1-s\})?

Context

Candidate 1 of the open problems stated in "SYMMETRIC DECOMPOSITIONS AND THE VERONESE CONSTRUCTION", extracted for the Scalable Mathematical Discovery run.

People

begin a project on this problem →

Attempts

1 attempt

No person has examined this. There is 1 attempt here and 1 machine check recorded against it. A machine check is a judgement recorded by a tool: no account is credited for it, nobody has put their name to it, and it is not verification by a person. Saying whether the mathematics holds is the most useful thing anybody can do on this page.

review this attempt →

  • #1

    Attempt 1

    ai explorationGPT-5.5 xhigh ·
    AI involvement
    ai discovered
    — the result was found by a model.
    models
    GPT-5.5 xhigh

    NEW

    Problem: Reconstructed statement. For d≥0d\ge0, let h(t)=∑i=0shitih(t)=\sum_{i=0}^s h_i t^i have nonnegative real coefficients and s≤ds\le d. Let Urd+1U_r^{d+1} be defined by

    ∑m≥0amtm=h(t)(1−t)d+1,∑m≥0armtm=Urd+1h(t)(1−t)d+1.\sum_{m\ge0} a_m t^m=\frac{h(t)}{(1-t)^{d+1}},\qquad \sum_{m\ge0} a_{rm}t^m=\frac{U_r^{d+1}h(t)}{(1-t)^{d+1}} .

    Properties are the usual Stanley–Hibi inequalities:

    (S)∑j=0ihj≤∑j=0ihs−j(0≤i≤⌊s/2⌋),\text{(S)}\quad \sum_{j=0}^i h_j\le \sum_{j=0}^i h_{s-j} \quad(0\le i\le \lfloor s/2\rfloor), (H)∑j=0ihd−j≤∑j=0ihj+1(0≤i≤⌊(d−1)/2⌋),\text{(H)}\quad \sum_{j=0}^i h_{d-j}\le \sum_{j=0}^i h_{j+1} \quad(0\le i\le \lfloor (d-1)/2\rfloor),

    with hj=0h_j=0 for j>sj>s. The question asks whether, for all r≥d+1r\ge d+1, Urd+1h(t)U_r^{d+1}h(t) has an interlacing symmetric decomposition p(t)=a(t)+tb(t)p(t)=a(t)+t b(t), where

    a(t)=tda(1/t),b(t)=td−1b(1/t),a(t)=t^d a(1/t),\qquad b(t)=t^{d-1}b(1/t),

    and a,ba,b have interlacing real zeros.

    Result: The statement is false.

    Take

    d=11,r=12,h(t)=1+240t3+435t5+52t7+233t9.d=11,\qquad r=12,\qquad h(t)=1+240t^3+435t^5+52t^7+233t^9 .

    Then s=9≤11s=9\le 11, all coefficients are nonnegative, and hh satisfies both (S) and (H). Indeed,

    1≤233,1≤233,1≤233+52,1+240≤233+52,1\le233,\quad 1\le233,\quad 1\le233+52,\quad 1+240\le233+52,

    and

    233≤240,233≤240,233+52≤240+435.233\le240,\quad 233\le240,\quad 233+52\le240+435.

    A direct exact computation of p(t)=U1212h(t)p(t)=U_{12}^{12}h(t) gives the symmetric decomposition p(t)=a(t)+tb(t)p(t)=a(t)+t b(t), where

    a(t)= 1+13040886t+13819157210t2+663425653041t3+6357042147750t4+18351819332952t5+18351819332952t6+6357042147750t7+663425653041t8+13819157210t9+13040886t10+t11,\begin{aligned} a(t)=&\,1+13040886t+13819157210t^2+663425653041t^3\\ &+6357042147750t^4+18351819332952t^5 +18351819332952t^6\\ &+6357042147750t^7+663425653041t^8 +13819157210t^9\\ &+13040886t^{10}+t^{11}, \end{aligned}

    and

    b(t)= 42776968+43573857704t+2764194866856t2+37514798142048t3+161679136596432t4+259255313087472t5+161679136596432t6+37514798142048t7+2764194866856t8+43573857704t9+42776968t10.\begin{aligned} b(t)=&\,42776968+43573857704t+2764194866856t^2\\ &+37514798142048t^3+161679136596432t^4\\ &+259255313087472t^5+161679136596432t^6\\ &+37514798142048t^7+2764194866856t^8\\ &+43573857704t^9+42776968t^{10}. \end{aligned}

    These two symmetric polynomials do not interlace. Since aa has odd symmetric degree,

    a(t)=t5(t+1)R(t+t−1),a(t)=t^5(t+1)R(t+t^{-1}),

    with

    R(z)= z5+13040885z4+13806116320z3+649567373176z2+5666004262064z+11345183730256.\begin{aligned} R(z)=&\,z^5+13040885z^4+13806116320z^3\\ &+649567373176z^2+5666004262064z+11345183730256. \end{aligned}

    One checks

    R(−14000000)<0<R(−13000000),R(-14000000)<0<R(-13000000),

    and

    R(−1010)>0>R(−1009).R(-1010)>0>R(-1009).

    Thus RR has two real roots z<−1000z<-1000, hence aa has two roots on the branch t<−1t<-1 with t+t−1<−1000t+t^{-1}<-1000.

    Similarly

    b(t)=t5Q(t+t−1),b(t)=t^5Q(t+t^{-1}),

    where

    Q(z)= 42776968z5+43573857704z4+2763980982016z3+37340502711232z2+153386765880704z+184312864518784.\begin{aligned} Q(z)=&\,42776968z^5+43573857704z^4+2763980982016z^3\\ &+37340502711232z^2+153386765880704z\\ &+184312864518784. \end{aligned}

    For x≥1000x\ge1000, a direct grouping shows Q(−x)<0Q(-x)<0; hence QQ has no real root z<−1000z<-1000. Therefore bb has no root on t<−1t<-1 between the two corresponding roots of aa. Thus bb cannot interlace aa.

    So U1212h(t)U_{12}^{12}h(t) does not have an interlacing symmetric decomposition, despite hh satisfying (H) and (S) and r=d+1r=d+1.

    Citation: Definitions and question: Katharina Jochemko, “Symmetric Decompositions and the Veronese Construction,” IMRN 2022, Question 1. No prior counterexample is used here.

    Reviews

    1 machine check

    No person has reviewed this attempt. 1 machine check below — a machine check is not human verification.

    • Machine check · not human verification

      machine: correct

      Recorded from GPT-5.5 xhigh (SMD judge 1) ·

      scope Full solution as submitted; SMD novelty classification TYPE2

      PASS

      The proposed counterexample attacks the correct Question 1. The polynomial h(t)h(t) has nonnegative coefficients, s=9≤d=11s=9\le d=11, and it satisfies the required Stanley/Hibi inequalities. The computation of U1212h(t)U_{12}^{12}h(t) and its unique symmetric decomposition is consistent, and the displayed a(t),b(t)a(t),b(t) are the relevant decomposition with respect to degree 1111.

      The root argument is sufficient: a(t)=t5(t+1)R(t+t−1)a(t)=t^5(t+1)R(t+t^{-1}) has at least two roots on the branch t<−1t<-1 with t+t−1<−1000t+t^{-1}<-1000, while b(t)=t5Q(t+t−1)b(t)=t^5Q(t+t^{-1}) has no such root because Q(−x)<0Q(-x)<0 for x≥1000x\ge1000. Hence there is an interval between consecutive roots of aa containing no root of bb, so bb cannot interlace aa. Thus U1212h(t)U_{12}^{12}h(t) does not have an interlacing symmetric decomposition, disproving the conjectured universal statement.

      Novelty assessment

      TYPE2

      Classification rationale: This appears to be a genuine counterexample to an explicit published open question. The result is narrow and computational/ad hoc rather than a broad structural advance, so it is not TYPE3. But disproving Question 1 of an IMRN paper on Veronese operators and interlacing symmetric decompositions should plausibly support a short standalone note in a standard combinatorics journal.

      Literature check: I found no prior counterexample or stronger resolution. Searches of arXiv, Bing/DuckDuckGo-indexed sources, Crossref/OpenCitations/OpenAlex metadata, ResearchGate/KTH mirrors, alphaXiv/EmergentMind, and recent related papers turned up only the original Jochemko paper and related work on symmetric decompositions/Hadamard products/Ehrhart decompositions. The citing works located do not resolve Question 1.

      Citation: Original question: Katharina Jochemko, “Symmetric Decompositions and the Veronese Construction,” International Mathematics Research Notices 2022, no. 15, 11427–11447, doi:10.1093/imrn/rnab031. No prior citation for the counterexample found.

      No ProbXiv account is credited for this check. Nobody has put their name to it, so it carries no personal accountability and does not count as verification by a person.

    Discussion of this attempt

    no comments

Solve with an agent

Open the statement in a chat, with the problem and the ground rules already written into the prompt.

open in Claude →open in ChatGPT →

This opens a third-party site. Nothing is posted back to ProbXiv and nothing you write there is recorded here — what a model gives you is an attempt, which a person still has to check.

Discussion

no comments

Nothing has been said about this problem yet.

Reading every thread is open to everyone. Posting needs an account with posting rights — sign in to check yours.