Always Learning, Always Mixing: Efficient and Simple Data Mixing All The Time

• URL: http://arxiv.org/abs/2605.15220v1
• Date: Wed, 13 May 2026 02:29:19 GMT
• Title: Always Learning, Always Mixing: Efficient and Simple Data Mixing All The Time
• Abstract summary: We introduce OP-Mix, a data mixing algorithm that operates across the entire language model training lifecycle.<n>In pretraining, OP-Mix improves upon training without mixing by 6.3% in average perplexity.<n>For continual learning, OP-Mix matches the performance of both retraining and on-policy distillation while using 66% and 95% less overall compute.
• Score: 51.671620992989375
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Data mixing decides how to combine different sources or types of data and is a consequential problem throughout language model training. In pretraining, data composition is a key determinant of model quality; in continual learning and adaptation, it governs what is retained and acquired. Yet existing data mixing methods address only one phase of this lifecycle at a time: some require smaller proxy models tied to a single training phase, others assume a fixed domain set, and continual learning lacks principled guidance altogether. We argue that data mixing is fundamentally an online decision making problem -- one that recurs throughout training and demands a single, unified solution. We introduce OP-Mix (On-Policy Mix), a data mixing algorithm that operates across the entire language model training lifecycle. Our main insight is that candidate data mixtures can be cheaply simulated by interpolating between low-rank adapters trained directly on the current model, eliminating separate proxy models and ensuring the search is always grounded in the model's actual learning dynamics. Across pretraining, continual midtraining, and continual instruction tuning, OP-Mix consistently finds near-optimal mixtures while using a fraction of the compute of the baselines. In pretraining, OP-Mix improves upon training without mixing by 6.3% in average perplexity. For continual learning, OP-Mix matches the performance of both retraining and on-policy distillation while using 66% and 95% less overall compute, respectively. OP-Mix suggests a different view of language model training: not a sequence of distinct phases, but a single continuous process of learning from data.

Related papers

• Decouple Searching from Training: Scaling Data Mixing via Model Merging for Large Language Model Pre-training [16.022416196267937]
  We propose Decouple Searching from Training Mix (DeMix), a novel framework that leverages model merging to predict optimal data ratios.<n>We show that DeMix breaks the trade-off between sufficiency, accuracy and efficiency, obtaining the optimal mixture with higher benchmark performance at lower search cost.
  arXiv  Detail & Related papers  (2026-01-31T14:27:46Z)
• Aioli: A Unified Optimization Framework for Language Model Data Mixing [74.50480703834508]
  We show that no existing method consistently outperforms a simple stratified sampling baseline in terms of average test perplexity.<n>We derive a new online method named Aioli, which directly estimates the mixing law parameters throughout training and uses them to dynamically adjust proportions.
  arXiv  Detail & Related papers  (2024-11-08T17:50:24Z)
• RC-Mixup: A Data Augmentation Strategy against Noisy Data for Regression Tasks [27.247270530020664]
  We study the problem of robust data augmentation for regression tasks in the presence of noisy data. C-Mixup is more selective in which samples to mix based on their label distances for better regression performance. We propose RC-Mixup, which tightly integrates C-Mixup with multi-round robust training methods for a synergistic effect.
  arXiv  Detail & Related papers  (2024-05-28T08:02:42Z)
• Data Mixing Laws: Optimizing Data Mixtures by Predicting Language Modeling Performance [55.872926690722714]
  We study the predictability of model performance regarding the mixture proportions in function forms.<n>We propose nested use of the scaling laws of training steps, model sizes, and our data mixing law.<n>Our method effectively optimize the training mixture of a 1B model trained for 100B tokens in RedPajama.
  arXiv  Detail & Related papers  (2024-03-25T17:14:00Z)
• Efficient Online Data Mixing For Language Model Pre-Training [101.45242332613944]
  Existing data selection methods suffer from slow and computationally expensive processes. Data mixing, on the other hand, reduces the complexity of data selection by grouping data points together. We develop an efficient algorithm for Online Data Mixing (ODM) that combines elements from both data selection and data mixing.
  arXiv  Detail & Related papers  (2023-12-05T00:42:35Z)
• Integrating Local Real Data with Global Gradient Prototypes for Classifier Re-Balancing in Federated Long-Tailed Learning [60.41501515192088]
  Federated Learning (FL) has become a popular distributed learning paradigm that involves multiple clients training a global model collaboratively. The data samples usually follow a long-tailed distribution in the real world, and FL on the decentralized and long-tailed data yields a poorly-behaved global model. In this work, we integrate the local real data with the global gradient prototypes to form the local balanced datasets.
  arXiv  Detail & Related papers  (2023-01-25T03:18:10Z)
• Learning from aggregated data with a maximum entropy model [73.63512438583375]
  We show how a new model, similar to a logistic regression, may be learned from aggregated data only by approximating the unobserved feature distribution with a maximum entropy hypothesis. We present empirical evidence on several public datasets that the model learned this way can achieve performances comparable to those of a logistic model trained with the full unaggregated data.
  arXiv  Detail & Related papers  (2022-10-05T09:17:27Z)
• Towards Understanding the Data Dependency of Mixup-style Training [14.803285140800542]
  In the Mixup training paradigm, a model is trained using convex combinations of data points and their associated labels. Despite seeing very few true data points during training, models trained using Mixup seem to still minimize the original empirical risk. For a large class of linear models and linearly separable datasets, Mixup training leads to learning the same classifier as standard training.
  arXiv  Detail & Related papers  (2021-10-14T18:13:57Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.