Using Scaling Laws for Data Source Utility Estimation in Domain-Specific Pre-Training

• URL: http://arxiv.org/abs/2507.22250v1
• Date: Tue, 29 Jul 2025 21:56:45 GMT
• Title: Using Scaling Laws for Data Source Utility Estimation in Domain-Specific Pre-Training
• Authors: Oleksiy Ostapenko, Charles Guille-Escuret, Luke Kumar, Max Tian, Denis Kocetkov, Gopeshh Subbaraj, Raymond Li, Joel Lamy-Poirier, Sebastien Paquet, Torsten Scholak,
• Abstract summary: We introduce a framework for optimizing domain-specific dataset construction in foundation model training.<n>Our approach extends the usual point estimate approaches, aka micro-annealing, to estimating scaling laws.<n>We validate our approach through experiments on a pre-trained model with 7 billion parameters.
• Score: 4.90288999217624
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We introduce a framework for optimizing domain-specific dataset construction in foundation model training. Specifically, we seek a cost-efficient way to estimate the quality of data sources (e.g. synthetically generated or filtered web data, etc.) in order to make optimal decisions about resource allocation for data sourcing from these sources for the stage two pre-training phase, aka annealing, with the goal of specializing a generalist pre-trained model to specific domains. Our approach extends the usual point estimate approaches, aka micro-annealing, to estimating scaling laws by performing multiple annealing runs of varying compute spent on data curation and training. This addresses a key limitation in prior work, where reliance on point estimates for data scaling decisions can be misleading due to the lack of rank invariance across compute scales -- a phenomenon we confirm in our experiments. By systematically analyzing performance gains relative to acquisition costs, we find that scaling curves can be estimated for different data sources. Such scaling laws can inform cost effective resource allocation across different data acquisition methods (e.g. synthetic data), data sources (e.g. user or web data) and available compute resources. We validate our approach through experiments on a pre-trained model with 7 billion parameters. We adapt it to: a domain well-represented in the pre-training data -- the medical domain, and a domain underrepresented in the pretraining corpora -- the math domain. We show that one can efficiently estimate the scaling behaviors of a data source by running multiple annealing runs, which can lead to different conclusions, had one used point estimates using the usual micro-annealing technique instead. This enables data-driven decision-making for selecting and optimizing data sources.

Related papers

• Formal Bayesian Transfer Learning via the Total Risk Prior [1.8570591025615457]
  We show how a particular instantiation of our prior leads to a Bayesian Lasso in a transformed coordinate system.<n>We also demonstrate that recently proposed minimax-frequentist transfer learning techniques may be viewed as an approximate Maximum a Posteriori approach to our model.
  arXiv  Detail & Related papers  (2025-07-31T17:55:16Z)
• Data Mixing Agent: Learning to Re-weight Domains for Continual Pre-training [30.915768238214653]
  Data Mixing Agent is an end-to-end framework that learns to re-weight domains.<n>It generalizes well across unseen source fields, target models, and domain spaces without retraining.
  arXiv  Detail & Related papers  (2025-07-21T14:01:54Z)
• Capturing the Temporal Dependence of Training Data Influence [100.91355498124527]
  We formalize the concept of trajectory-specific leave-one-out influence, which quantifies the impact of removing a data point during training.<n>We propose data value embedding, a novel technique enabling efficient approximation of trajectory-specific LOO.<n>As data value embedding captures training data ordering, it offers valuable insights into model training dynamics.
  arXiv  Detail & Related papers  (2024-12-12T18:28:55Z)
• Source-Free Domain-Invariant Performance Prediction [68.39031800809553]
  We propose a source-free approach centred on uncertainty-based estimation, using a generative model for calibration in the absence of source data. Our experiments on benchmark object recognition datasets reveal that existing source-based methods fall short with limited source sample availability. Our approach significantly outperforms the current state-of-the-art source-free and source-based methods, affirming its effectiveness in domain-invariant performance estimation.
  arXiv  Detail & Related papers  (2024-08-05T03:18:58Z)
• LAVA: Data Valuation without Pre-Specified Learning Algorithms [20.578106028270607]
  We introduce a new framework that can value training data in a way that is oblivious to the downstream learning algorithm. We develop a proxy for the validation performance associated with a training set based on a non-conventional class-wise Wasserstein distance between training and validation sets. We show that the distance characterizes the upper bound of the validation performance for any given model under certain Lipschitz conditions.
  arXiv  Detail & Related papers  (2023-04-28T19:05:16Z)
• Uncertainty-guided Source-free Domain Adaptation [77.3844160723014]
  Source-free domain adaptation (SFDA) aims to adapt a classifier to an unlabelled target data set by only using a pre-trained source model. We propose quantifying the uncertainty in the source model predictions and utilizing it to guide the target adaptation.
  arXiv  Detail & Related papers  (2022-08-16T08:03:30Z)
• Source-Free Domain Adaptation via Distribution Estimation [106.48277721860036]
  Domain Adaptation aims to transfer the knowledge learned from a labeled source domain to an unlabeled target domain whose data distributions are different. Recently, Source-Free Domain Adaptation (SFDA) has drawn much attention, which tries to tackle domain adaptation problem without using source data. In this work, we propose a novel framework called SFDA-DE to address SFDA task via source Distribution Estimation.
  arXiv  Detail & Related papers  (2022-04-24T12:22:19Z)
• Invariance Learning in Deep Neural Networks with Differentiable Laplace Approximations [76.82124752950148]
  We develop a convenient gradient-based method for selecting the data augmentation. We use a differentiable Kronecker-factored Laplace approximation to the marginal likelihood as our objective.
  arXiv  Detail & Related papers  (2022-02-22T02:51:11Z)
• Leveraging Unlabeled Data to Predict Out-of-Distribution Performance [63.740181251997306]
  Real-world machine learning deployments are characterized by mismatches between the source (training) and target (test) distributions. In this work, we investigate methods for predicting the target domain accuracy using only labeled source data and unlabeled target data. We propose Average Thresholded Confidence (ATC), a practical method that learns a threshold on the model's confidence, predicting accuracy as the fraction of unlabeled examples.
  arXiv  Detail & Related papers  (2022-01-11T23:01:12Z)
• Deep Learning with Multiple Data Set: A Weighted Goal Programming Approach [2.7393821783237184]
  Large-scale data analysis is growing at an exponential rate as data proliferates in our societies. Deep Learning models require plenty of resources, and distributed training is needed. This paper presents a Multicriteria approach for distributed learning.
  arXiv  Detail & Related papers  (2021-11-27T07:10:25Z)

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.