# MLFont: Machine Learning-Driven Font Subsetting and Range-Compact Unicode-Range Slicing for Low-Latency Web Typography

DevFeed: [MLFont: Machine Learning-Driven Font Subsetting and Range-Compact Unicode-Range Slicing for Low-Latency Web Typography](<https://devfeed.tech/articles/mlfont-machine-learning-driven-font-subsetting-and-range-compact-unicode-range-slicing-for-low-latency-web-typography-93387.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.24937>)

Author: Quan Wang, Yiran Mao

Published: 2026-10-21T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [Deep learning](<https://devfeed.tech/topics/deep-learning.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [3](<https://devfeed.tech/tags/3.md>), [abstract](<https://devfeed.tech/tags/abstract.md>), [algorithm](<https://devfeed.tech/tags/algorithm.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [bandwidth](<https://devfeed.tech/tags/bandwidth.md>), [bandwidth-savings](<https://devfeed.tech/tags/bandwidth-savings.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [bloat](<https://devfeed.tech/tags/bloat.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [font](<https://devfeed.tech/tags/font.md>), [latency](<https://devfeed.tech/tags/latency.md>), [learning](<https://devfeed.tech/tags/learning.md>), [low-latency](<https://devfeed.tech/tags/low-latency.md>), [machine](<https://devfeed.tech/tags/machine.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [range](<https://devfeed.tech/tags/range.md>), [typography](<https://devfeed.tech/tags/typography.md>), [unicode](<https://devfeed.tech/tags/unicode.md>), [web](<https://devfeed.tech/tags/web.md>), [web-typography](<https://devfeed.tech/tags/web-typography.md>)

## AI overview

MLFont learns Unicode font slices from text corpora to reduce web font downloads while retaining full glyph coverage. Its three stages isolate high-frequency codepoints, group co-occurring characters with Online LDA or DNNTM, then use a range-compact genetic algorithm with Potts-Viterbi dynamic programming and island bridging to balance topic fit and CSS fragmentation. On five Google Noto fonts and 20 websites, the reported benchmarks show 56.7%-75.3% lower 3G per-glyph latency than single-file delivery and 46.1%-66.3% lower than equal-bin baselines on CJK, with 100% coverage. CSS intervals fell 17.9%-41.7%; cold-cache speedups were 1.56x-6.09x, with reported bandwidth savings of 65.4%-84.9% for CJK and 44.7% for LGC. Results are from simulator and Headless Chrome benchmarks; the excerpt does not establish performance beyond those tested settings.

## Source excerpt

arXiv:2610.24937v1 Announce Type: new Abstract: CSS3 @font-face rules with unicode-range descriptors partition web fonts into downloadable slices. While Latin fonts need ~20 KB in WOFF2, Pan-European (LGC) and East Asian (CJK) fonts contain 2,297-30,759 codepoints (158 KB-5.9 MB in WOFF2), where heuristic partitioning causes unused-glyph downloads, high HTTP RTT, CSS bloat, or missing-glyph fallbacks. We present MLFont, a machine learning and optimization framework that learns font slicing strategies from corpora via three stages: (1) a Zipfian High-Frequency Tier (most) isolating top-frequency codepoints; (2) a Latent Co-Occurrence Topic Tier using Online Variational Bayes LDA (Online LDA) or a Deep Neural Network Topic Model (DNNTM) with Shannon Entropy Difference Loss and Sinkhorn-Knopp normalization; and (3) a Memetic Range-Compact Genetic Algorithm (GA) with Potts-Viterbi dynamic programming and island bridging that minimizes topic discrepancy and CSS fragmentation. Across five Google Noto fonts (LGC, KR, JP, TC, SC), MLFont cuts 3G per-glyph latency by 56.7%-75.3% over single-file delivery and 46.1%-66.3% over equal-bin baselines on CJK with 100% coverage, reduces CSS intervals by 17.9%-41.7%, and achieves 1.56x-6.09x cold-cache speedups (65.4%-84.9% CJK and 44.7% LGC bandwidth savings) across 20 websites in simulator and Headless Chrome benchmarks (45/45 exact slice-set parity). Code: https://github.com/wq2012/mlfont Models: https://huggingface.co/collections/wq2012/mlfont