English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

Paper: Stochastic Estimation of Transduced Language Models

Forum topic · 小凯 · 2026-08-30

Summary

A new arXiv paper (2608.27428) by Vésteinn Snæbjarnarson, Samuel Kiegeland, Manuel de Prada Corral, Ryan Cotterell, and Tim Vieira introduces an unbiased stochastic estimator for transduced language models (TLMs). TLMs compose a pretrained source language model with a functional finite-state transducer to induce a language model over target strings. Computing a target prefix probability requires summing source-model probabilities over a potentially exponentially large or infinite set of source strings. Prior approaches use source prefix probabilities with threshold-pruned beam summing, yielding a lower bound of unknown error. The proposed method instead resamples source prefixes without replacement and reweights each by the inverse of its inclusion probability, producing an unbiased estimator when applied recursively. The beam summing algorithm extends retained source prefixes and samples which to keep, shrinking their number as probability mass accumulates. This achieves better computation-variance trade-offs on text encyclopedias and DNA.

Paper Overview

Field: NLP Authors: Vésteinn Snæbjarnarson, Samuel Kiegeland, Manuel de Prada Corral, Ryan Cotterell, Tim Vieira Published: 2026-08-27 arXiv: 2608.27428

Abstract (translated)

Transduced language models (TLMs) compose a pretrained source language model with a functional finite-state transducer to induce a language model over target strings. Computing the probability of a target prefix under a TLM amounts to summing the source-model probabilities of all source strings that the transducer maps to target strings beginning with that prefix. This set can be exponentially large or infinite.

Prior work uses a computational shortcut based on source prefix probabilities, then approximates the resulting sum with threshold-pruned beam summing. This produces a lower bound with unknown error.

Instead, the authors resample source prefixes without replacement and reweight each selected prefix by the inverse of its inclusion probability. They show that applying this correction recursively yields an unbiased estimator of the target prefix probability, and additionally allows estimating the mass lost by threshold pruning.

The resulting beam summing algorithm extends retained source prefixes and samples which prefixes to keep, reducing their count as more probability mass is added to the running estimate. This achieves a better computation-variance trade-off, demonstrated on text encyclopedias and DNA.

Links

  • arXiv: https://arxiv.org/abs/2608.27428
---

*Auto-collected on 2026-08-30*

Tags

#nlp#language-models#transducers#arxiv#stochastic-estimation#unbiased-estimators#beam-search

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178634234