Softening Discrete Relaxation

Andrew M. Finch, Richard C. Wilson, Edwin R. Hancock

NeurIPS 1996 pp. 438-444

/neurips/1996/finch1996neurips-softening/

Abstract

This paper describes a new framework for relational graph match(cid:173) ing. The starting point is a recently reported Bayesian consistency measure which gauges structural differences using Hamming dis(cid:173) tance. The main contributions of the work are threefold. Firstly, we demonstrate how the discrete components of the cost func(cid:173) tion can be softened. The second contribution is to show how the softened cost function can be used to locate matches using continuous non-linear optimisation. Finally, we show how the res(cid:173) ulting graph matching algorithm relates to the standard quadratic assignment problem.

PDF NeurIPS Semantic Scholar

Cite

Text

Finch et al. "Softening Discrete Relaxation." Neural Information Processing Systems, 1996.

Markdown

[Finch et al. "Softening Discrete Relaxation." Neural Information Processing Systems, 1996.](https://mlanthology.org/neurips/1996/finch1996neurips-softening/)

BibTeX

@inproceedings{finch1996neurips-softening,
  title     = {{Softening Discrete Relaxation}},
  author    = {Finch, Andrew M. and Wilson, Richard C. and Hancock, Edwin R.},
  booktitle = {Neural Information Processing Systems},
  year      = {1996},
  pages     = {438-444},
  url       = {https://mlanthology.org/neurips/1996/finch1996neurips-softening/}
}