Discovering Translation-Worthy Languages with E-Values

Wajdi Ben Saad; Safa Madiouni

cs.CLSep 6, 2026

Discovering Translation-Worthy Languages with E-Values

Authors: Wajdi Ben Saad, Safa Madiouni

Organizations: Carthago Labs Paris, France · IÉSEG Paris, France

Abstract

Choosing when to translate multilingual documents is a central routing problem in text classification: translation can improve predictions for some languages while degrading others or adding unnecessary computation. Uniform translation and heuristic language tiers do not provide statistically controlled route selection. We introduce a language-level router based on paired e-processes that continuously compares direct and translation-assisted classification before freezing a routing policy. A familywise-controlled threshold of 280 bounds the probability of any false route across 14 eligible languages per dataset by 0.05. On SIB-200 and MASSIVE, the router selects translation for 4 of 15 languages and 14 of 15 locales, improving held-out accuracy over direct classification by 8.14 and 16.70 percentage points, respectively. All 28 decisions remain stable across 50 outcome-independent orderings and relative to the per-group threshold. Our results demonstrate that paired e-processes enable statistically controlled, anytime-valid, and auditable multilingual classification routing.

Multilingual Text Classification

See all 20 papers

A Cost-Efficient Routing Pipeline for Multilingual Short-Text Classification Using Small Language Models
Aug 11, 2026Wajdi Ben Saad, Safa MadiouniZero-Shot Text Classification Multilingual Text Classification
Carthago Labs Paris, France · Université Paris Dauphine-PSL Paris, France
No One Fits All: From Fixed Prompting to Learned Routing in Multilingual LLMs
Apr 18, 2026Wei-Chi Wu, Sheng-Lun Wei, Hen-Hsen Huang +1Multilingual Language Model Evaluation LLM Prompting
Department of Computer Science and Information Engineering, National Taiwan University, Taiwan · Institute of Information Science, Academia Sinica, Taiwan · AI Research Center (AINTU), National Taiwan University, Taiwan
Learning When to Translate for Multilingual Reasoning
Jun 1, 2026Deokhyung Kang, Hyounghun KimMultilingual Language Models Cross-Lingual Reasoning in Language Models
Graduate School of Artificial Intelligence · Department of Computer Science & Engineering

Aug 11, 2026J/K move · Enter open · S save

A Cost-Efficient Routing Pipeline for Multilingual Short-Text Classification Using Small Language Models

Wajdi Ben Saad, Safa Madiouni

Carthago Labs Paris, France · Université Paris Dauphine-PSL Paris, France

Zero-Shot Text Classification Multilingual Text Classification Text Classification Low-Resource Language Processing

Open paper AskPDF arXiv

Multilingual short-text classification supports operational systems such as content moderation, customer support routing, and intent recognition, yet aggregate evaluation often hides large differences between high-resource and low-resource languages. Uniform inference policies are simple to deploy, but they assume that all languages are equally well served. In this work, we evaluate a fixed-list routing strategy that keeps stronger languages on a direct multilingual path and selectively sends weaker languages through translation into English before zero-shot classification. The pipeline is fully self-hosted, uses pretrained compact sentence encoders, and requires no task-specific fine-tuning. We test the approach on two benchmarks chosen to differ in scale and label granularity: a 15-language subset of SIB-200 for seven-way topic classification and a 15-locale subset of MASSIVE for intent classification over an official 60-intent inventory. On SIB-200, the best overall configuration is R1, which translates only the low-resource tier: high-tier and mid-tier Macro-F1 remain unchanged, while low-tier Macro-F1 rises from 0.4632 to 0.6828. On the MASSIVE subset, the same low-tier intervention raises low-tier Macro-F1 from 0.2143 to 0.4417, but the best overall result is obtained by full translation, R3, at Macro-F1 0.4647. Across these two benchmarks, selective translation is a reliable intervention for weaker languages, whereas the optimal routing boundary depends on the task. We therefore report routing through tier-level quality gains and tier-level latency rather than a single global efficiency score.

Abstract

Multilingual Text Classification

Explore similar work

A Cost-Efficient Routing Pipeline for Multilingual Short-Text Classification Using Small Language Models

No One Fits All: From Fixed Prompting to Learned Routing in Multilingual LLMs

Learning When to Translate for Multilingual Reasoning