Variance Alignment Score: A Simple But Tough-to-Beat Data Selection Method for Multimodal Contrastive Learning

• URL: http://arxiv.org/abs/2402.02055v1
• Date: Sat, 3 Feb 2024 06:29:04 GMT
• Title: Variance Alignment Score: A Simple But Tough-to-Beat Data Selection Method for Multimodal Contrastive Learning
• Authors: Yiping Wang, Yifang Chen, Wendan Yan, Kevin Jamieson, Simon Shaolei Du
• Abstract summary: We propose a principled metric named Variance Alignment Score (VAS), which has the form $langle Sigma_texttest, Sigma_irangle$. We show that applying VAS and CLIP scores together can outperform baselines by a margin of $1.3%$ on 38 evaluation sets for noisy dataset DataComp and $2.5%$ on VTAB for high-quality dataset CC12M.
• Score: 17.40655778450583
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: In recent years, data selection has emerged as a core issue for large-scale visual-language model pretraining, especially on noisy web-curated datasets. One widely adopted strategy assigns quality scores such as CLIP similarity for each sample and retains the data pairs with the highest scores. However, these approaches are agnostic of data distribution and always fail to select the most informative samples. To solve this problem, we propose a simple yet theoretically principled metric named Variance Alignment Score (VAS), which has the form $\langle \Sigma_{\text{test}}, \Sigma_i\rangle$. Here, $\Sigma_{\text{test}}$ represents the target (cross-)covariance matrix we aim to align, potentially based on prior knowledge, while $\Sigma_i$ denotes the tensor product of single or multi-modal representations for the $i$-th sample. We further design a new data selection method that maximizes the total VAS. We provide theoretical analysis in a simplified setting to demonstrate the theoretical advantage of VAS over random or other existing data selection. Experimentally, applying VAS and CLIP scores together can outperform baselines by a margin of $1.3\%$ average on 38 evaluation sets for noisy dataset DataComp and $2.5\%$ on VTAB for high-quality dataset CC12M. Additionally, our ablation study also shows visual features are better than text for calculating VAS, and the related classical experimental design methods may fail under this context.

Related papers

• Adapt-$\infty$: Scalable Lifelong Multimodal Instruction Tuning via Dynamic Data Selection [89.42023974249122]
  Adapt-$infty$ is a new multi-way and adaptive data selection approach for Lifelong Instruction Tuning. We construct pseudo-skill clusters by grouping gradient-based sample vectors. We select the best-performing data selector for each skill cluster from a pool of selector experts.
  arXiv  Detail & Related papers  (2024-10-14T15:48:09Z)
• BWS: Best Window Selection Based on Sample Scores for Data Pruning across Broad Ranges [12.248397169100784]
  Data subset selection aims to find a smaller yet informative subset of a large dataset that can approximate the full-dataset training. We introduce a universal and efficient data subset selection method, Best Window Selection (BWS), by proposing a method to choose the best window subset from samples ordered based on their difficulty scores.
  arXiv  Detail & Related papers  (2024-06-05T08:33:09Z)
• Data-Efficient Learning via Clustering-Based Sensitivity Sampling: Foundation Models and Beyond [28.651041302245538]
  We present a new data selection approach based on $k$-means clustering and sampling sensitivity. We show how it can be applied on linear regression, leading to a new sampling strategy that surprisingly matches the performances of leverage score sampling.
  arXiv  Detail & Related papers  (2024-02-27T09:03:43Z)
• Towards a statistical theory of data selection under weak supervision [7.540077751816086]
  Given a sample of size $N$, it is often useful to select a subsample of smaller size $nN$ to be used for statistical estimation or learning. We assume to be given $N$ unlabeled samples $bold x_i_ile N$, and to be given access to a surrogate model' that can predict labels $y_i$ better than random guessing.
  arXiv  Detail & Related papers  (2023-09-25T22:23:27Z)
• Sample-Efficient Linear Representation Learning from Non-IID Non-Isotropic Data [4.971690889257356]
  We introduce an adaptation of the alternating minimization-descent scheme proposed by Collins and Nayer and Vaswani. We show that vanilla alternating-minimization descent fails catastrophically even for iid, but mildly non-isotropic data. Our analysis unifies and generalizes prior work, and provides a flexible framework for a wider range of applications.
  arXiv  Detail & Related papers  (2023-08-08T17:56:20Z)
• Project and Probe: Sample-Efficient Domain Adaptation by Interpolating Orthogonal Features [119.22672589020394]
  We propose a lightweight, sample-efficient approach that learns a diverse set of features and adapts to a target distribution by interpolating these features. Our experiments on four datasets, with multiple distribution shift settings for each, show that Pro$2$ improves performance by 5-15% when given limited target data.
  arXiv  Detail & Related papers  (2023-02-10T18:58:03Z)
• Bias Mimicking: A Simple Sampling Approach for Bias Mitigation [57.17709477668213]
  We introduce a new class-conditioned sampling method: Bias Mimicking. Bias Mimicking improves underrepresented groups' accuracy of sampling methods by 3% over four benchmarks.
  arXiv  Detail & Related papers  (2022-09-30T17:33:00Z)
• Pareto Optimization for Active Learning under Out-of-Distribution Data Scenarios [79.02009938011447]
  We propose a sampling scheme, which selects optimal subsets of unlabeled samples with fixed batch size from the unlabeled data pool. Experimental results show its effectiveness on both classical Machine Learning (ML) and Deep Learning (DL) tasks.
  arXiv  Detail & Related papers  (2022-07-04T04:11:44Z)
• Improving Contrastive Learning on Imbalanced Seed Data via Open-World Sampling [96.8742582581744]
  We present an open-world unlabeled data sampling framework called Model-Aware K-center (MAK) MAK follows three simple principles: tailness, proximity, and diversity. We demonstrate that MAK can consistently improve both the overall representation quality and the class balancedness of the learned features.
  arXiv  Detail & Related papers  (2021-11-01T15:09:41Z)
• 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)

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.