Convex Dataset Valuation for Post-Training

• URL: http://arxiv.org/abs/2605.16704v1
• Date: Fri, 15 May 2026 23:35:07 GMT
• Title: Convex Dataset Valuation for Post-Training
• Abstract summary: We propose a scalable convex dataset-level valuation method based on mean matching (KMM) in gradient space.<n>We show that our approach consistently outperforms existing valuation baselines, achieving stronger performance with low overhead.
• Score: 21.22512177964122
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Improving LLM performance on downstream tasks sometimes requires leveraging auxiliary datasets during post-training. In practice, however, developers face constraints on compute, labeling, and licensing costs that preclude using all available data, necessitating principled dataset-level selection. These constraints are increasingly shaped by dataset marketplaces, where data acquisition is governed by budgets and negotiation. We study dataset valuation as a subset selection problem during LLM post-training. Our goal is to identify and weight auxiliary datasets so as to maximize target task performance given constrained budgets. We first show that commonly used gradient alignment scores provide a reasonable yet incomplete valuation signal, as they ignore redundancy among datasets. To address this, we propose a scalable convex dataset-level valuation method based on kernel mean matching (KMM) in gradient space, which jointly accounts for alignment with the target task and redundancy across auxiliary datasets. Through extensive experiments across diverse post-training settings and tasks, we show that our approach consistently outperforms existing valuation baselines, achieving stronger performance with low computational overhead. Our results position dataset valuation as a practical decision tool for post-training data selection in market-constrained large language model settings. The code is available at https://github.com/uiuctml/convex_data_valuation.

Related papers

• Swift Cross-Dataset Pruning: Enhancing Fine-Tuning Efficiency in Natural Language Understanding [2.379669478864599]
  Current cross-dataset pruning techniques for fine-tuning often rely on computationally expensive sample ranking processes.<n>We propose Swift Cross-Dataset Pruning (SCDP), which uses TF-IDF embeddings with geometric median to rapidly evaluate sample importance.<n> Experimental results on six diverse datasets demonstrate the effectiveness of our method, spanning various tasks and scales.
  arXiv  Detail & Related papers  (2025-01-05T03:52:04Z)
• Domain Adaptation for Offline Reinforcement Learning with Limited Samples [2.3674123304219816]
  offline reinforcement learning learns effective policies from a static target dataset.<n>It relies on the quality and size of the target dataset and it degrades if limited samples in the target dataset are available.<n>This paper proposes the first framework that theoretically explores the impact of the weights assigned to each dataset on the performance of offline RL.
  arXiv  Detail & Related papers  (2024-08-22T05:38:48Z)
• Neural Dynamic Data Valuation: A Stochastic Optimal Control Approach [15.538631565455448]
  This paper introduces a new framework that formulates data valuation as an optimal control problem to capture the dynamic evolution of data utility over time.<n>Unlike static approaches, NDDV models data interactions through continuous trajectories that reflect both individual and collective learning dynamics.
  arXiv  Detail & Related papers  (2024-04-30T13:39:26Z)
• Pessimistic Value Iteration for Multi-Task Data Sharing in Offline Reinforcement Learning [116.87367592920171]
  Offline Reinforcement Learning (RL) has shown promising results in learning a task-specific policy from a fixed dataset. In scenarios where the dataset for a specific task is limited, a natural approach is to improve offline RL with datasets from other tasks. We propose an uncertainty-based Multi-Task Data Sharing (MTDS) approach that shares the entire dataset without data selection.
  arXiv  Detail & Related papers  (2024-04-30T08:16:52Z)
• How Much Data are Enough? Investigating Dataset Requirements for Patch-Based Brain MRI Segmentation Tasks [74.21484375019334]
  Training deep neural networks reliably requires access to large-scale datasets. To mitigate both the time and financial costs associated with model development, a clear understanding of the amount of data required to train a satisfactory model is crucial. This paper proposes a strategic framework for estimating the amount of annotated data required to train patch-based segmentation networks.
  arXiv  Detail & Related papers  (2024-04-04T13:55:06Z)
• How to Train Data-Efficient LLMs [56.41105687693619]
  We study data-efficient approaches for pre-training language models (LLMs) We find that Ask-LLM and Density sampling are the best methods in their respective categories. In our comparison of 19 samplers, involving hundreds of evaluation tasks and pre-training runs, we find that Ask-LLM and Density are the best methods in their respective categories.
  arXiv  Detail & Related papers  (2024-02-15T02:27:57Z)
• Building Manufacturing Deep Learning Models with Minimal and Imbalanced Training Data Using Domain Adaptation and Data Augmentation [15.333573151694576]
  We propose a novel domain adaptation (DA) approach to address the problem of labeled training data scarcity for a target learning task. Our approach works for scenarios where the source dataset and the dataset available for the target learning task have same or different feature spaces. We evaluate our combined approach using image data for wafer defect prediction.
  arXiv  Detail & Related papers  (2023-05-31T21:45:34Z)
• DataPerf: Benchmarks for Data-Centric AI Development [81.03754002516862]
  DataPerf is a community-led benchmark suite for evaluating ML datasets and data-centric algorithms. We provide an open, online platform with multiple rounds of challenges to support this iterative development. The benchmarks, online evaluation platform, and baseline implementations are open source.
  arXiv  Detail & Related papers  (2022-07-20T17:47:54Z)
• Open-Set Semi-Supervised Learning for 3D Point Cloud Understanding [62.17020485045456]
  It is commonly assumed in semi-supervised learning (SSL) that the unlabeled data are drawn from the same distribution as that of the labeled ones. We propose to selectively utilize unlabeled data through sample weighting, so that only conducive unlabeled data would be prioritized.
  arXiv  Detail & Related papers  (2022-05-02T16:09:17Z)
• How to distribute data across tasks for meta-learning? [59.608652082495624]
  We show that the optimal number of data points per task depends on the budget, but it converges to a unique constant value for large budgets. Our results suggest a simple and efficient procedure for data collection.
  arXiv  Detail & Related papers  (2021-03-15T15:38:47Z)
• Efficient Conditional Pre-training for Transfer Learning [71.01129334495553]
  We propose efficient filtering methods to select relevant subsets from the pre-training dataset. We validate our techniques by pre-training on ImageNet in both the unsupervised and supervised settings. We improve standard ImageNet pre-training by 1-3% by tuning available models on our subsets and pre-training on a dataset filtered from a larger scale dataset.
  arXiv  Detail & Related papers  (2020-11-20T06:16:15Z)

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.