![[2506.01085] Learning What Matters: Prioritized Concept Learning via Relative Error-driven Sample Selection](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2506.01085] Learning What Matters: Prioritized Concept Learning via Relative Error-driven Sample Selection
Summary
The paper presents PROGRESS, a framework for prioritized concept learning in vision-language models, enabling efficient sample selection based on relative error during training.
Why It Matters
This research addresses the challenges of instruction tuning in vision-language models by proposing a more efficient method for sample selection, which can significantly reduce data and compute requirements. This has implications for advancing AI capabilities while minimizing resource consumption, making it relevant for researchers and practitioners in machine learning and AI.
Key Takeaways
- PROGRESS dynamically selects learning samples based on model needs.
- The framework requires no upfront annotations and minimizes supervision.
- It outperforms existing methods with less data and computational resources.
- Demonstrates strong generalization across different model architectures.
- Offers a scalable solution for efficient learning in AI.
Computer Science > Computer Vision and Pattern Recognition arXiv:2506.01085 (cs) [Submitted on 1 Jun 2025 (v1), last revised 25 Feb 2026 (this version, v2)] Title:Learning What Matters: Prioritized Concept Learning via Relative Error-driven Sample Selection Authors:Shivam Chandhok, Qian Yang, Oscar Manas, Kanishk Jain, Leonid Sigal, Aishwarya Agrawal View a PDF of the paper titled Learning What Matters: Prioritized Concept Learning via Relative Error-driven Sample Selection, by Shivam Chandhok and 5 other authors View PDF HTML (experimental) Abstract:Instruction tuning has been central to the success of recent vision-language models (VLMs), but it remains expensive-requiring large-scale datasets, high-quality annotations, and large compute budgets. We propose PRioritized cOncept learninG via Relative Error-driven Sample Selection (PROGRESS), a data- and compute-efficient framework that enables VLMs to dynamically select what to learn next based on their evolving needs during training. At each stage, the model tracks its learning progress across skills and selects the most informative samples-those it has not already mastered and that are not too difficult to learn at the current stage of training. This strategy effectively controls skill acquisition and the order in which skills are learned. Specifically, we sample from skills showing the highest learning progress, prioritizing those with the most rapid improvement. Unlike prior methods, PROGRESS requires no upfront answer...
