arXiv Machine Learning September 25, 2026 By Yanran Wu, Sana Lakdawala, Renzo Tassara Miller, Chongyang Bai, Sharath Ciddu, Shivendra Pratap Singh, Kungang Li, Sandeep Pandey, Chunwei Liu

Let Training Guide Selection: Online Synthetic Data Filtering via Real-Anchored Utility

Read the original on arXiv Machine Learning →

The paper introduces FROST, an online framework that filters synthetic training data by estimating its utility through gradient feedback anchored in real data. FROST calibrates batch utility against recent history to decide when to filter, removing 20–30% of synthetic samples while improving performance on image classification and LLM fine-tuning tasks. The method is also applied to a large‑scale industrial ads re‑ranking system, yielding significant gains over an optimized production baseline.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

arXiv AI

2d ago

Training-Aware Target Coverage for Synthetic Data Selection

arXiv:2610.00814v1 Announce Type: cross Abstract: Synthetic data are increasingly used to scale LLM training, yet more synthetic data do not necessarily produce better models. Useful synthetic data m...

By Yang Ba, Michelle V. Mancenido, Rong Pan

llms fine-tuning

arXiv Machine Learning

Jun 16

Discriminative Span as a Predictor of Synthetic Data Utility via Classifier Reconstruction

arXiv:2605. 09697v3 Announce Type: replace-cross Abstract: In many real-world computer vision applications, including medical imaging and industrial inspection, binary classification tasks are characterized by a severe scarcity of positive samples.

By Radhika Amar Desai, Modigari Narendra

rag computer-vision data-engineering

arXiv AI

Jul 7

Post-Generation Curation of Synthetic Images via Homogeneous-Heterogeneous Splitting

arXiv:2607. 02637v1 Announce Type: cross Abstract: Recent generative models can produce high-quality synthetic images, offering scalable training training data for data-hungry models.

By Disheng Liu, Tuo Liang, Chaoda Song, Yu Yin

llms computer-vision fine-tuning benchmarks safety

arXiv Machine Learning

Sep 4

A Real-Calibrated Synthetic-First Data Engine

The paper introduces the Real‑Calibrated Synthetic‑First Data Engine, a modular pipeline that integrates controllable diffusion‑based synthetic image generation with multi‑stage curation, filtering, and optional uncertainty‑driven selection and human verification. Designed as a CLI‑based framework, it allows independent configuration of generation, filtering, selection, and validation modules to enhance reproducibility and flexibility in real‑world data workflows. Empirical tests on human pose estimation demonstrate that synthetic data can boost a real‑data baseline when used as low‑cost augmentation, though synthetic‑only training still lags behind real‑only performance, underscoring the importance of data‑centric orchestration in low‑data regimes.

By Yukang Shen, Zhiguo Liu, Yingshu Li, Yan Huang

agents diffusion computer-vision data-engineering

arXiv Machine Learning

Aug 11

From Fake to Real: Pretraining on Balanced Synthetic Images to Prevent Spurious Correlations in Image Recognition

arXiv:2308. 04553v4 Announce Type: replace-cross Abstract: Visual recognition models are prone to learning spurious correlations induced by a biased training set where certain conditions $B$ (\eg, Indoors) are over-represented in certain classes $Y$ (\eg, Big Dogs).

By Maan Qraitem, Kate Saenko, Bryan A. Plummer

computer-vision benchmarks safety

arXiv Computer Vision

3d ago

Efficient Dataset Distillation for Pre-Trained Self-Supervised Models via Statistical Flow Matching

arXiv:2602.05391v3 Announce Type: replace Abstract: Dataset distillation seeks to synthesize a compact surrogate dataset that enables performance comparable to training on the original dataset for do...

By Qianxin Xia, Jiawei Du, Yuhan Zhang, Xin Zhang, Xuewan He, Wenbo Jiang, Jielei Wang, Tao Luo, Guoming Lu

diffusion efficiency benchmarks

One email a morning, machine-written

Related stories

Training-Aware Target Coverage for Synthetic Data Selection

Discriminative Span as a Predictor of Synthetic Data Utility via Classifier Reconstruction

Post-Generation Curation of Synthetic Images via Homogeneous-Heterogeneous Splitting

A Real-Calibrated Synthetic-First Data Engine

From Fake to Real: Pretraining on Balanced Synthetic Images to Prevent Spurious Correlations in Image Recognition

Efficient Dataset Distillation for Pre-Trained Self-Supervised Models via Statistical Flow Matching