NeatIBP 1.0, a package generating small-size integration-by-parts relations for Feynman integrals

Zihao Wu, Janko Böhm, Rourou Ma, Hefeng Xu, Yang ZhangComput. Phys. Commun 2024 · DOI

TL;DR

This paper introduces NEATIBP, a new software package that automates the generation of "small-size" integration-by-parts (IBP) identities for Feynman integrals. By employing advanced algebraic geometry techniques, NEATIBP significantly reduces the complexity of these identities, making them more manageable for calculations in quantum field theory, particularly for high-energy physics experiments.

What Problem Does This Paper Address?

Calculating scattering amplitudes in quantum field theory often involves dealing with a vast number of Feynman integrals. To make these calculations tractable, these integrals are typically reduced to a smaller set of "master integrals" using techniques like Integration-by-Parts (IBP) reduction. However, the standard IBP reduction methods, such as Laporta's algorithm, can generate an enormous number of IBP identities, leading to computationally prohibitive systems. This paper addresses the challenge of reducing the size of these IBP systems, making the reduction process more efficient and accessible.

Methods

NEATIBP leverages a combination of sophisticated mathematical techniques:

  • Syzygy and Module Intersection: The core of NEATIBP's approach lies in using syzygy equations and module intersection. These algebraic geometry tools allow the generation of IBP relations that control the "propagator degree," effectively keeping the indices of the integrals smaller. This is a departure from traditional methods that can lead to increased indices.
  • Baikov Representation: The package utilizes the Baikov representation of Feynman integrals, which translates the problem of finding IBP relations into module intersection computations.
  • Computer Algebra Systems: NEATIBP is powered by the computer algebra systems Mathematica and Singular, with the SPASM library playing a crucial role in performing row reduction over finite fields.
  • Parallelization: The package is designed for parallel computation, distributing tasks across different Feynman integral sectors to speed up the process.
  • Seeding and Selection Algorithms: Sophisticated algorithms are employed for "seeding" (assigning specific values to symbolic indices) and selecting the most relevant and independent IBP relations, further optimizing the system size.

Datasets

Not discussed in the paper. The paper focuses on a computational method and its implementation, rather than training on specific datasets. The "datasets" in this context would be the Feynman diagrams and their associated integrals, which are generated or defined by the user.

Key Insights and Ideas

  • Size Matters: The paper highlights that the computational complexity of IBP reduction is heavily influenced by the number of initial IBP relations. NEATIBP's primary contribution is to drastically reduce this number.
  • Algebraic Geometry as a Tool: The successful application of advanced algebraic geometry concepts (syzygies, module intersection) to a physics problem demonstrates the power of interdisciplinary approaches in tackling complex computational challenges.
  • Controlled Growth of Indices: By controlling the propagator degree, NEATIBP avoids the proliferation of integrals with unnecessarily high indices, which are often irrelevant to the final reduction.
  • Automated Efficiency: The package automates a complex process that previously required significant manual intervention and expertise, making advanced IBP reduction more accessible.

Key Results

NEATIBP generates IBP systems that are "orders of magnitude smaller" than those produced by standard algorithms like Laporta's. This leads to significantly reduced computation times and memory requirements. The paper demonstrates this with several multi-loop IBP examples, showing substantial improvements in efficiency compared to existing tools like FIRE6. For instance, in one example, NEATIBP generated 14120 IBP identities compared to FIRE6's 11207942 for a similar task.

Limitations and Future Work

  • Finite Field Risks: The paper acknowledges that using row reduction over finite fields can carry risks if the chosen numeric points are not "generic" enough, potentially leading to incorrect results. NEATIBP v1.0 includes a check process to mitigate this, but more robust methods are planned.
  • Propagator Cuts: While NEATIBP v1.0 can handle propagator cuts, it currently has limitations on the complexity of the integrand for cuts. Future versions aim to address this by automatically finding all spanning cuts.
  • Open-Source Migration: A future goal is to migrate the code to rely solely on open-source software.
  • Advanced Parallelization: The authors plan to introduce more sophisticated parallelization strategies, including within individual sectors and across different cuts.
  • Alternative Methods: Future work will explore alternative approaches for module intersection and syzygy computation, such as linear algebra techniques and the dual conformal symmetry method.

Summary updated

BibTeX
@article{wu2024neatibp,
  author = {Zihao Wu and Janko Böhm and Rourou Ma and Hefeng Xu and Yang Zhang},
  title = {{NeatIBP 1.0, a package generating small-size integration-by-parts relations for Feynman integrals}},
  journal = {Comput. Phys. Commun.},
  year = {2024},
  doi = {10.1016/j.cpc.2023.108999},
}
Figure