Search papers, labs, and topics across Lattice.
This paper revisits dataset distillation by introducing Influence Matching (Inf-Match), which focuses on aligning the final outcomes of model training rather than the intermediate processes. By employing a differentiable, sample-level influence estimator that operates in linear time, the authors effectively quantify the impact of data on model parameters without relying on complex inverse-Hessian computations. The results demonstrate that Inf-Match achieves superior accuracy on standard classification benchmarks and vision-language tasks, outperforming existing methods like NCFM by notable margins.
Aligning the final outcomes of training rather than just the training process leads to a significant boost in model performance, with Inf-Match achieving state-of-the-art results across multiple benchmarks.
We revisit dataset distillation from an outcome-centric perspective. Rather than aligning process surrogates (per-step gradients or training trajectories), Influence Matching (Inf-Match) aligns the final outcome of training: it learns a compact synthetic set whose effect on the converged parameters matches that of the full dataset. Concretely, we introduce a fully differentiable, sample-level influence estimator that quantifies parameter shifts from adding or removing data, without time-consuming inverse-Hessian products or convexity assumptions. The estimator runs in linear time by unrolling the optimization dynamics and applying a first-order Taylor approximation. We then learn the synthetic set by minimizing the mismatch between its influence and that of the real dataset, yielding outcome alignment rather than heuristic process imitation. Inf-Match delivers the best accuracy across standard classification benchmarks. For instance, on Tiny-ImageNet (IPC=10), Inf-Match attains 31.5\%, a +4.7\% improvement over NCFM. Beyond classification, Inf-Match scales to vision-language distillation on Flickr30K, outperforming strong process-matching baselines. For instance, with 200 to 1000 synthetic samples, our method achieved a leading impressive average on image/text retrieval tasks, higher than NCFM by 2.5\%. The code will be released via https://github.com/hrtan/infmatch.
Stanford HAI