{
 "artifacts": null,
 "category": "architecture",
 "description": "Replace an independently learned n-by-n pairwise score tensor with coordinates in the paper's n-1 dimensional consistent subspace. The neural network predicts only basis coefficients, and a fixed reconstruction produces all pairwise logits, reducing the comparison representation from O(n^2) degrees of freedom to O(n) while guaranteeing transitivity.",
 "formulas_latex": [
  "$$S=\\sum_{k=1}^{n-1}\\alpha_kB_k,\\qquad \\alpha_k=\\frac{\\langle S,B_k\\rangle}{\\langle B_k,B_k\\rangle}$$",
  "$$S_{ij}=\\sum_{k=1}^{n-1}\\alpha_k b^{(k)}_{ij}=u_i-u_j$$",
  "$$\\dim(\\mathcal A_n)=n-1\\ll\\frac{n(n-1)}{2}=\\dim(\\mathfrak{so}(n))$$",
  "$$S_{ij}+S_{jk}+S_{ki}=0\\qquad\\text{for all }i,j,k$$"
 ],
 "id": 189,
 "implementation": "(1) Use this in a ranking, preference, retrieval, or set-comparison module whose current final output is an n by n score matrix. Replace its final pairwise layer with a linear or MLP head that outputs `alpha` of shape [batch,n-1]. Store a fixed tensor B of shape [n-1,n,n], normalized so `(B[k]*B[k]).sum()` is one. (2) Reconstruct with `S=(alpha[..., :, None, None]*B).sum(dim=-3)`, set the diagonal to zero, and use `sigmoid(S[i,j])` for pairwise probabilities or sort a utility vector recovered from the basis coordinates. The simplest equivalent implementation predicts utilities u of shape [batch,n] and computes `S=u[..., :,None]-u[...,None,:]`; the basis form is useful when the paper's sparse or windowed basis is desired. (3) The fixed basis, dimensionality reduction, and cycle guarantee come from the paper; alpha is learned by ordinary backpropagation. For variable-size sets, use a maximum-size basis with a validity mask, or generate an orthonormal centered basis for each n at runtime. (4) First run a synthetic n=32 Bradley-Terry benchmark and a MovieLens pairwise-preference benchmark. Compare the O(n^2) unconstrained pairwise head with the O(n) transitive head at matched hidden width and training steps. Measure parameter count, activation memory, forward latency, pairwise accuracy, ranking regret, and triangle-cycle count. The expected result is zero predicted cycles, substantially lower memory and latency, and comparable or better held-out ranking accuracy when labels contain contradictions. A negative result would be a large clean-data accuracy gap caused by the transitivity assumption.",
 "math_summary": "The space of skew-symmetric n by n matrices has dimension n(n-1)/2, whereas the additive-consistent subspace A_n has dimension n-1. Let B_1,...,B_{n-1} be the paper's orthogonal basis, where B_k=[b^{(k)}_{ij}]. Every consistent logit matrix has the expansion S=sum_{k=1}^{n-1} alpha_k B_k, with alpha_k=\u003cS,B_k\u003e/\u003cB_k,B_k\u003e. Equivalently, S_{ij}=u_i-u_j for item utilities u. Therefore all triangle identities S_{ij}+S_{jk}+S_{ki}=0 hold exactly. The neural head learns alpha in R^{n-1} and reconstructs S using the fixed basis; unlike an unconstrained pairwise tensor, it cannot represent contradictory preference cycles.",
 "math_tags": [
  "linear-algebra",
  "geometry",
  "combinatorics"
 ],
 "ml_areas": [
  "embedding",
  "graph-nn",
  "inference-speedup"
 ],
 "paper": {
  "arxiv_id": "2608.25923",
  "arxiv_url": "https://arxiv.org/abs/2608.25923",
  "summary_what_math_gives_to_ml": "The paper gives an explicit orthogonal-coordinate construction for the additive-consistency subspace of skew-symmetric pairwise-comparison matrices, together with closed-form projection and multiplicative reconstruction formulas. The transferable asset is a differentiable mechanism that removes cyclic and contradictory components from arbitrary pairwise logits while retaining the closest globally transitive signal. This can be used as a preference-learning head, a pairwise-ranking regularizer, or a structured comparison layer with only O(n) consistent degrees of freedom instead of O(n^2). The most practical tests are robustness to noisy comparisons and parameter reduction at matched ranking accuracy.",
  "title": "Efficient tensor bases for pairwise comparisons",
  "year": "2026"
 },
 "ratings": {
  "difficulty": 4,
  "novelty": 5,
  "usefulness": 5
 },
 "solves": [
  "memory",
  "scalability",
  "generalization"
 ],
 "title": "O(n)-parameter transitive pairwise head",
 "url": "https://synthcore.org/idea/189/o-n-parameter-transitive-pairwise-head",
 "verification": {
  "peer_reviewed": false,
  "status": "unverified",
  "status_label": "Unverified",
  "verdict_source": "deterministic test code (paired-seed permutation statistics)",
  "verification_axes": {
   "benchmark_mechanism": {
    "confirmed": null,
    "tested": false
   },
   "practical_benchmark": {
    "beats_baseline": null,
    "tested": false
   },
   "toy_mechanism_gate": {
    "confirmed": null,
    "tested": false
   }
  }
 }
}