GPTQ is equivalent to Babai's nearest plane algorithm for CVP on the Hessian lattice of layer inputs, yielding geometric interpretation, inherited error bounds, and improved clipping-free quantization with GPU kernels.
Mathematische Annalen , publisher =
8 Pith papers cite this work. Polarity classification is still indexing.
8
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
verdicts
UNVERDICTED 8roles
background 1polarities
background 1representative citing papers
The zeta(3)-proportional contribution to the four-loop twist-two gluon anomalous dimension gamma_gg^(3)(N) is constructed in analytic form from low-N moments via the LLL algorithm together with reciprocity and supersymmetric constraints.
Integer-valued images can be uniquely recovered from a minimal set of DFT coefficients through algebraic reduction to 1D problems and dynamic programming algorithms that use lattice approximation to handle NP-hard subproblems.
Lattice reduction does not change the solution vector for many algorithms in vector perturbation precoding due to a unique lattice structure, and LLL-aided nearest plane methods do not beat conventional THP under mutual information.
Optimized multiseries hypergeometric formulas for single and multiple logarithms, found via integer programming on a lattice, enable efficient high-precision computation with demonstrated results exceeding 10^11 digits and an application to 2*10^12 digits of log(10).
Explicit estimates for the count of integral ideals in number fields are derived with error terms that grow much more slowly with the degree n than the standard n^{n^2} bound.
Presents analogous arguments supporting the Cobham-Edmonds thesis that feasible computation explicates to P.
A layered framework is defined to interpret post-quantum cryptographic security assumptions through complexity models, combinatorial Hodge theory on lattices, and Julia-based lattice reduction experiments.
citing papers explorer
-
The Geometry of LLM Quantization: GPTQ as Babai's Nearest Plane Algorithm
GPTQ is equivalent to Babai's nearest plane algorithm for CVP on the Hessian lattice of layer inputs, yielding geometric interpretation, inherited error bounds, and improved clipping-free quantization with GPU kernels.
-
Four-Loop Gluon Anomalous Dimension of General Lorentz Spin: Transcendental Part
The zeta(3)-proportional contribution to the four-loop twist-two gluon anomalous dimension gamma_gg^(3)(N) is constructed in analytic form from low-N moments via the LLL algorithm together with reciprocity and supersymmetric constraints.
-
Recovery of Integer Images from Minimal DFT Measurements: Uniqueness and Inversion Algorithms
Integer-valued images can be uniquely recovered from a minimal set of DFT coefficients through algebraic reduction to 1D problems and dynamic programming algorithms that use lattice approximation to handle NP-hard subproblems.
-
Is Lattice Reduction Necessary for Vector Perturbation Precoding?
Lattice reduction does not change the solution vector for many algorithms in vector perturbation precoding due to a unique lattice structure, and LLL-aided nearest plane methods do not beat conventional THP under mutual information.
-
Fast Ramanujan--type Series for Logarithms. Part II
Optimized multiseries hypergeometric formulas for single and multiple logarithms, found via integer programming on a lattice, enable efficient high-precision computation with demonstrated results exceeding 10^11 digits and an application to 2*10^12 digits of log(10).
-
Explicit counting of ideals in number fields of arbitrary degree
Explicit estimates for the count of integral ideals in number fields are derived with error terms that grow much more slowly with the degree n than the standard n^{n^2} bound.
-
Feasibilism, Explication, and the Cobham-Edmonds Thesis
Presents analogous arguments supporting the Cobham-Edmonds thesis that feasible computation explicates to P.
-
Explainable PQC: A Layered Interpretive Framework for Post-Quantum Cryptographic Security Assumptions
A layered framework is defined to interpret post-quantum cryptographic security assumptions through complexity models, combinatorial Hodge theory on lattices, and Julia-based lattice reduction experiments.