We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Approximating Permutations with Neural Network Components for Travelling Photographer Problem

Authors: Sue Sin Chong
Abstract: Most of the current inference techniques rely upon Bayesian inference on Probabilistic Graphical Models of observations and do predictions and classification on observations. However, there is very little literature on the mining of relationships between observations and building models of the relationship between sets of observations or the generating context of the observations. Moreover, event understanding of machines with observation inputs needs to understand the relationship between observations. Thus there is a crucial need to build models and develop effective data structures to accumulate and organize relationships between observations. Given a PGM model, this paper attempts to fit a permutation of states to a sequence of observation tokens (The Travelling Photographer Problem). We have devised a machine learning inspired architecture for randomized approximation of state permutation, facilitating parallelization of heuristic search of permutations. As a result, our algorithm can solve The Travelling Photographer Problem with minimal error. Furthermore, we demonstrate that by mimicking machine learning components such as normalization, dropout, and lambda layer with a randomized algorithm, we can devise an architecture that solves TPP, a permutation NP-Hard problem. Other than TPP, we can also provide a 2-Local improvement heuristic for the Travelling Salesman Problem (TSP) with similar ideas.
Comments: 11 pages
Subjects: Machine Learning (cs.LG)
Journal reference: Published in 2022 Asia Conference on Algorithms, Computing and Machine Learning (CACML 2022)
DOI: 10.1109/CACML55074.2022.00036
Cite as: arXiv:2205.00242 [cs.LG]
  (or arXiv:2205.00242v2 [cs.LG] for this version)

Submission history

From: Sue Sin Chong [view email]
[v1] Sat, 30 Apr 2022 11:24:21 GMT (208kb,D)
[v2] Mon, 23 May 2022 07:22:02 GMT (225kb,D)

Link back to: arXiv, form interface, contact.