TANDEM: Bi-Level Data Mixture Optimization with Twin Networks

• URL: http://arxiv.org/abs/2606.04401v1
• Date: Wed, 03 Jun 2026 03:28:46 GMT
• Title: TANDEM: Bi-Level Data Mixture Optimization with Twin Networks
• Abstract summary: We propose a bi-level optimization problem for large language models (LLMs)<n>We simplify the problem into a single-level penalized form and solve with twin networks.<n>Our proposed method measures the data efficacy through the difference between the twin models and up-weights domains.
• Score: 37.94726322431112
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: The capabilities of large language models (LLMs) significantly depend on training data drawn from various domains. Optimizing domain-specific mixture ratios can be modeled as a bi-level optimization problem, which we simplify into a single-level penalized form and solve with twin networks: a proxy model trained on primary data and a dynamically updated reference model trained with additional data. Our proposed method, Twin Networks for bi-level DatA mixturE optiMization (TANDEM), measures the data efficacy through the difference between the twin models and up-weights domains that benefit more from the additional data. TANDEM provides theoretical guarantees and wider applicability, compared to prior approaches. Furthermore, our bi-level perspective suggests new settings to study domain reweighting such as data-restricted scenarios and supervised fine-tuning, where optimized mixture ratios significantly improve the performance. Extensive experiments validate TANDEM's effectiveness in all scenarios.

Related papers

• Bagging-Based Model Merging for Robust General Text Embeddings [73.51674133699196]
  General-purpose text embedding models underpin a wide range of NLP and information retrieval applications.<n>We present a systematic study of multi-task training for text embeddings from two perspectives: data scheduling and model merging.<n>We propose Bagging-based rObust mOdel Merging (BOOM), which trains multiple embedding models on sampled subsets and merges them into a single model.
  arXiv  Detail & Related papers  (2026-02-05T15:45:08Z)
• FoundIR-v2: Optimizing Pre-Training Data Mixtures for Image Restoration Foundation Model [89.23522479092537]
  We propose a high-capacity diffusion-based image restoration foundation model, FoundIR-v2.<n>By leveraging the data mixing law, our method ensures a balanced dataset composition.<n>Our method achieves favorable performance against state-of-the-art approaches.
  arXiv  Detail & Related papers  (2025-12-10T03:10:52Z)
• Heterogeneous Self-Supervised Acoustic Pre-Training with Local Constraints [64.15709757611369]
  We propose a new self-supervised pre-training approach to dealing with heterogeneous data.<n>The proposed approach can significantly improve the adaptivity of the self-supervised pre-trained model for the downstream supervised fine-tuning tasks.
  arXiv  Detail & Related papers  (2025-08-27T15:48:50Z)
• Data Mixing Optimization for Supervised Fine-Tuning of Large Language Models [5.626695215498112]
  We frame data mixing as an optimization problem and introduce a novel method designed to minimize validation loss.<n>Our approach parametrizes the loss by modeling effective data transferred and leveraging scaling laws for fine-tuning.<n>We show that our algorithm achieves excellent overall and individual performance across all domains.
  arXiv  Detail & Related papers  (2025-08-16T07:28:39Z)
• IDEAL: Data Equilibrium Adaptation for Multi-Capability Language Model Alignment [29.703775936837012]
  Large Language Models (LLMs) have achieved impressive performance through Supervised Fine-tuning (SFT) on diverse instructional datasets.<n>When training on multiple capabilities simultaneously, the mixture training dataset, governed by volumes of data from different domains, is a critical factor that directly impacts the final model's performance.<n>We introduce an innovative data equilibrium framework designed to effectively optimize volumes of data from different domains within mixture SFT datasets.
  arXiv  Detail & Related papers  (2025-05-19T06:42:44Z)
• Adaptive Data Optimization: Dynamic Sample Selection with Scaling Laws [59.03420759554073]
  We introduce Adaptive Data Optimization (ADO), an algorithm that optimize data distributions in an online fashion, concurrent with model training. ADO does not require external knowledge, proxy models, or modifications to the model update. ADO uses per-domain scaling laws to estimate the learning potential of each domain during training and adjusts the data mixture accordingly.
  arXiv  Detail & Related papers  (2024-10-15T17:47:44Z)
• BiMix: A Bivariate Data Mixing Law for Language Model Pretraining [47.77701041534746]
  The impact of pretraining data composition on model performance remains poorly understood.<n>$textbfBiMix$ provides a systematic framework for understanding and optimizing data mixtures.<n>Our work contributes both theoretical insights into data mixing dynamics and practical tools for enhancing LLM training efficiency.
  arXiv  Detail & Related papers  (2024-05-23T09:44:02Z)
• Comparative Analysis of Different Efficient Fine Tuning Methods of Large Language Models (LLMs) in Low-Resource Setting [0.0]
  We try to push the understanding of different fine-tuning strategies for large language models (LLMs) We compare state-of-the-art methods like vanilla fine-tuning and Pattern-Based Fine-Tuning (PBFT) on pre-trained models across two datasets, COLA and MNLI. Our findings suggest that these alternative strategies can exhibit out-of-domain generalization comparable to that of vanilla FT and PBFT.
  arXiv  Detail & Related papers  (2024-05-21T20:08:52Z)
• Pseudo-Trilateral Adversarial Training for Domain Adaptive Traversability Prediction [8.145900996884993]
  Traversability prediction is a fundamental perception capability for autonomous navigation. We propose a novel perception model that adopts a coarse-to-fine alignment (CALI) to perform unsupervised domain adaptation (UDA) We show the superiorities of our proposed models over multiple baselines in several challenging domain adaptation setups.
  arXiv  Detail & Related papers  (2023-06-26T00:39:32Z)
• Few-Shot Named Entity Recognition: A Comprehensive Study [92.40991050806544]
  We investigate three schemes to improve the model generalization ability for few-shot settings. We perform empirical comparisons on 10 public NER datasets with various proportions of labeled data. We create new state-of-the-art results on both few-shot and training-free settings.
  arXiv  Detail & Related papers  (2020-12-29T23:43:16Z)

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.