English
Related papers

Related papers: DoReMi: Optimizing Data Mixtures Speeds Up Languag…

200 papers

Selecting the best data mixture is critical for successful Supervised Fine-Tuning (SFT) of Multimodal Large Language Models. However, determining the optimal mixture weights across multiple domain-specific datasets remains a significant…

Machine Learning · Computer Science 2026-02-06 Davide Berasi , Matteo Farina , Massimiliano Mancini , Elisa Ricci

Constructing a unified 3D scene understanding model has long been hindered by the significant topological discrepancies across different sensor modalities. While applying the Mixture-of-Experts (MoE) architecture is an effective approach to…

Computer Vision and Pattern Recognition · Computer Science 2026-04-14 Mingwei Xing , Xinliang Wang , Yifeng Shi

Fine-tuning of Large Language Models (LLMs) for downstream tasks, performed on domain-specific data has shown significant promise. However, commercial use of such LLMs is limited by the high computational cost required for their deployment…

Computation and Language · Computer Science 2025-03-06 Boris Nazarov , Darya Frolova , Yackov Lubarsky , Alexei Gaissinski , Pavel Kisilev

The data used to pretrain large language models has a decisive impact on a model's downstream performance, which has led to a large body of work on data selection methods that aim to automatically determine the most suitable data to use for…

Computation and Language · Computer Science 2023-12-12 Alon Albalak , Liangming Pan , Colin Raffel , William Yang Wang

Prompt tuning, or the conditioning of a frozen pretrained language model (PLM) with soft prompts learned from data, has demonstrated impressive performance on a wide range of NLP tasks. However, prompt tuning requires a large training…

Computation and Language · Computer Science 2022-10-24 Xu Guo , Boyang Li , Han Yu

Continual pre-training is widely used to adapt LLMs to target languages and domains, yet the mixture ratio of training data remains a sensitive hyperparameter that is expensive to tune: they must be fixed before training begins, and a…

Computation and Language · Computer Science 2026-04-07 Haiyue Song , Masao Utiyama

Pre-trained Language Models (PLMs) can be accurately fine-tuned for downstream text processing tasks. Recently, researchers have introduced several parameter-efficient fine-tuning methods that optimize input prompts or adjust a small number…

Computation and Language · Computer Science 2024-06-07 Saeed Najafi , Alona Fyshe

Pretraining data of large language models composes multiple domains (e.g., web texts, academic papers, codes), whose mixture proportions crucially impact the competence of outcome models. While existing endeavors rely on heuristics or…

Computation and Language · Computer Science 2025-03-21 Jiasheng Ye , Peiju Liu , Tianxiang Sun , Jun Zhan , Yunhua Zhou , Xipeng Qiu

Continual pre-training on small-scale task-specific data is an effective method for improving large language models in new target fields, yet it risks catastrophic forgetting of their original capabilities. A common solution is to re-weight…

Machine Learning · Computer Science 2026-04-14 Kailai Yang , Xiao Liu , Lei Ji , Hao Li , Xiao Liang , Zhiwei Liu , Yeyun Gong , Peng Cheng , Mao Yang

Domain reweighting can improve sample efficiency and downstream generalization, but data-mixture optimization for multimodal midtraining remains largely unexplored. Current multimodal training recipes tune mixtures along a single dimension,…

Machine Learning · Computer Science 2026-04-17 Bingbing Wen , Sirajul Salekin , Feiyang Kang , Bill Howe , Lucy Lu Wang , Javier Movellan , Manjot Bilkhu

Training data mixtures greatly impact the generalization performance of large language models. Existing domain reweighting methods often rely on costly weight computations and require retraining when new data is introduced. To this end, we…

Machine Learning · Computer Science 2025-06-02 Wanyun Xie , Francesco Tonin , Volkan Cevher

Pretrained language models (PLMs) are trained on massive corpora, but often need to specialize to specific domains. A parameter-efficient adaptation method suggests training an adapter for each domain on the task of language modeling. This…

Computation and Language · Computer Science 2023-03-29 Alexandra Chronopoulou , Matthew E. Peters , Alexander Fraser , Jesse Dodge

The ever-growing diversity of pre-training text corpora has equipped language models with generalization capabilities across various downstream tasks. However, such diverse datasets are often too large for academic budgets; hence, most…

Computation and Language · Computer Science 2023-04-18 Jean Kaddour

While Large Language Models (LLMs) have revolutionized artificial intelligence, fine-tuning LLMs is extraordinarily computationally expensive, preventing smaller businesses and research teams with limited GPU resources from engaging with…

Machine Learning · Computer Science 2025-08-26 Daniel Frees , Aditri Bhagirath , Moritz Bolling

Large language models (LLMs) increasingly serve as the backbone for classifying text associated with distinct domains and simultaneously several labels (classes). When encountering domain shifts, e.g., classifier of movie reviews from IMDb…

Computation and Language · Computer Science 2024-05-06 Miruna Beţianu , Abele Mălan , Marco Aldinucci , Robert Birke , Lydia Chen

Pretraining large language models (LLMs) on vast and heterogeneous datasets is crucial for achieving state-of-the-art performance across diverse downstream tasks. However, current training paradigms treat all samples equally, overlooking…

Machine Learning · Computer Science 2025-02-11 Daouda Sow , Herbert Woisetschläger , Saikiran Bulusu , Shiqiang Wang , Hans-Arno Jacobsen , Yingbin Liang

Machine learning algorithms minimizing the average training loss usually suffer from poor generalization performance due to the greedy exploitation of correlations among the training data, which are not stable under distributional shifts.…

Machine Learning · Computer Science 2022-07-29 Xi Leng , Xiaoying Tang , Yatao Bian

Modern machine learning pipelines are increasingly combining and mixing data from diverse and disparate sources, e.g., pre-training large language models. Yet, finding the optimal data mixture is a challenging and open problem. We formalize…

Machine Learning · Computer Science 2026-01-16 Anvith Thudi , Evianne Rovers , Yangjun Ruan , Tristan Thrush , Chris J. Maddison

For most natural language processing tasks, the dominant practice is to finetune large pretrained transformer models (e.g., BERT) using smaller downstream datasets. Despite the success of this approach, it remains unclear to what extent…

Computation and Language · Computer Science 2023-05-29 Kundan Krishna , Saurabh Garg , Jeffrey P. Bigham , Zachary C. Lipton

Existing pretraining data mixing methods for large language models (LLMs) typically follow a domain-wise methodology, a top-down process that first determines domain weights and then performs uniform data sampling across each domain.…

Computation and Language · Computer Science 2025-03-04 Xiangyu Xi , Deyang Kong , Jian Yang , Jiawei Yang , Zhengyu Chen , Wei Wang , Jingang Wang , Xunliang Cai , Shikun Zhang , Wei Ye