cd /news/natural-language-processing/visibility-graph-representations-of-… · home › topics › natural-language-processing › article
[ARTICLE · art-172406] src=machinebrief.com ↗ pub= topic=natural-language-processing verified=true sentiment=· neutral

Visibility Graph Representations of Word Embeddings for Text Classification

A paper on arXiv (2611.22685v1) reports that natural visibility graph representations of averaged word embeddings reached 91.24% accuracy on the four-class AG News benchmark at a 1,000-word vocabulary selected by chi-square, the best result among the configurations tested from 700 selected words onward with a fully connected classifier. The degree-based graph input exceeded a BERT-based baseline by 0.29 percentage points at 1,500 words, and in a separate convolutional comparison the adjacency-matrix input reached 86.16% at 1,000 words versus 83.12% for the strongest static embedding. The authors compared graph inputs against continuous bag-of-words, skip-gram, GloVe, FastText and BERT representations across vocabulary budgets of 100 to 1,500 words, and note that controls are needed to establish the reliability of the small differences observed.

key takeaway

Reading an averaged document embedding as an ordered series and classifying its natural visibility graph's node-degree vector beat a BERT-based baseline by 0.29 points at 1,500 selected words and hit 91.24% at 1,000 words on AG News.

by read1 min views2 publishedNov 19, 2026

arXiv:2611.22685v1 Announce Type: new Abstract: Text classifiers depend on representations that expose the distinctions between classes in a form the learning model can use, and a restricted vocabulary changes which representation is most effective. This paper studies natural visibility graphs as an intermediate representation for news topic classification under controlled vocabulary budgets. Word embeddings are averaged into a fixed-length document vector, the vector is read as an ordered numerical series, and its natural visibility graph is passed to a classifier either as an ordered node-degree vector or as an adjacency matrix. On the four-class AG News benchmark, chi-square selection retains 100 to 1,500 words, and the graph inputs are compared with direct embedding-based representations from continuous bag-of-words, skip-gram, GloVe, FastText and BERT. With a fully connected classifier, the degree representation is the most accurate configuration from 700 selected words onward; it reaches its highest accuracy of 91.24% at 1,000 words and exceeds the BERT-based baseline by 0.29 percentage points at 1,500 words. In a separate convolutional comparison, the adjacency input reaches 86.16% at 1,000 words, against 83.12% for the strongest static embedding. The results indicate that visibility-derived features can be competitive once the selected vocabulary provides adequate coverage. We discuss the roles of graph compression, coordinate ordering and classifier choice, and the controls needed to establish the reliability of the small differences observed.

── more in #natural-language-processing 4 stories · sorted by recency
── more on @arxiv 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/visibility-graph-rep…] indexed:0 read:1min 2026-11-19 · —