Better Language Models with Model Merging

1996/04/17 by Thorsten Brants, Brants, Thorsten
Computer Science · #Computation and Language (cs.CL) #FOS: Computer and information sciences #Natural Language Processing Techniques #Speech Recognition and Synthesis #Speech and dialogue systems #cmp-lg #cs.CL

paper · pdf · doi:10.48550/arxiv.cmp-lg/9604005

LaTeX, 9 pages. In Proceedings of EMNLP-96, Philadelphia, PA

arxiv created 1996/04/17 · openalex publication_date 1996/04/17 · arxiv updated 2009/11/30 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

This paper investigates model merging, a technique for deriving Markov models from text or speech corpora. Models are derived by starting with a large and specific model and by successively combining states to build smaller and more general models. We present methods to reduce the time complexity of the algorithm and report on experiments on deriving language models for a speech recognition task. The experiments show the advantage of model merging over the standard bigram approach. The merged model assigns a lower perplexity to the test set and uses considerably fewer states.

Related

• A Systematic Study of In-the-Wild Model Merging for Large Language Models
• LoRE-Merging: Exploring Low-Rank Estimation For Large Language Model Merging
• Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities
• Activation-Informed Merging of Large Language Models
• Best-first Model Merging for Hidden Markov Model Induction