Samm Sun — accepted AI-conference papers

2026

Dynamic Important Example Mining for Reinforcement Finetuning

Haoru Tan, Sitong Wu, Yanfeng Chen, Shizhen Zhao, Yang-Tian Sun, Tianjia Liu +6

CVPR 2026

Reinforcement fine-tuning (RFT) is increasingly used to strengthen the reasoning abilities of large models, yet its effectiveness is bound by how training data are selected and used. Most data-centric RFT methods rely on static or heuristic sample selection, implicitly assuming a sample's value is f

Cited by 0 Source ↗code ↗Scholar ↗

2024

Surge Phenomenon in Optimal Learning Rate and Batch Size Scaling

Shuaipeng Li, Penghao Zhao, Hailin Zhang, Samm Sun, Hao Wu, Dian Jiao +7

NeurIPS 2024poster

In current deep learning tasks, Adam-style optimizers—such as Adam, Adagrad, RMSprop, Adafactor, and Lion—have been widely used as alternatives to SGD-style optimizers. These optimizers typically update model parameters using the sign of gradients, resulting in more stable convergence curves. The l…

Cited by 6 Source ↗PDF ↗Scholar ↗