{"ID":2880595,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2508.13653","arxiv_id":"2508.13653","title":"GRAFT: Gradient-Aware Fast MaxVol Technique for Dynamic Data Sampling","abstract":"Training modern neural networks on large datasets is computationally and environmentally costly. We introduce GRAFT, a scalable in-training subset selection method that (i) extracts a low-rank feature representation for each batch, (ii) applies a Fast MaxVol sampler to select a small, diverse subset that spans the batch's dominant subspace, and (iii) dynamically adjusts the subset size using a gradient-approximation criterion. By operating in low-rank subspaces and training on carefully chosen examples instead of full batches, GRAFT preserves the training trajectory while reducing wall-clock time, energy consumption, and $\\mathrm{CO}_2$ emissions. Across multiple benchmarks, GRAFT matches or exceeds recent selection baselines in both accuracy and efficiency, providing a favorable trade-off between accuracy, efficiency, and emissions.","short_abstract":"Training modern neural networks on large datasets is computationally and environmentally costly. We introduce GRAFT, a scalable in-training subset selection method that (i) extracts a low-rank feature representation for each batch, (ii) applies a Fast MaxVol sampler to select a small, diverse subset that spans the batc...","url_abs":"https://arxiv.org/abs/2508.13653","url_pdf":"https://arxiv.org/pdf/2508.13653v2","authors":"[\"Ashish Jha\",\"Anh huy Phan\",\"Razan Dibo\",\"Valentin Leplat\"]","published":"2025-08-19T09:03:39Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"cs.AI\",\"math.NA\"]","methods":"[]","has_code":false}