4 papers
Variance-Aware Adaptive Weighting for Diffusion Model Training
Nanlong Sun, Lei Shi
Diffusion models have recently achieved remarkable success in generative modeling, yet their training dynamics across different noise levels remain highly imbalanced, which can lea…
LLM-as-RNN: A Recurrent Language Model for Memory Updates and Sequence Prediction
Yuxing Lu, J. Ben Tamo, Weichen Zhao +5
Large language models are strong sequence predictors, yet standard inference relies on immutable context histories. After making an error at generation step t, the model lacks an u…
Language-Guided Tuning: Enhancing Numeric Optimization with Textual Feedback
Yuxing Lu, Yucheng Hu, Nan Sun +1
Configuration optimization remains a critical bottleneck in machine learning, requiring coordinated tuning across model architecture, training strategy, feature engineering, and hy…
END: Robust Dual-Decoder Watermarking Framework Against Non-Differentiable Distortions
Nan Sun, Han Fang, Yuxing Lu +2
DNN-based watermarking methods have rapidly advanced, with the ``Encoder-Noise Layer-Decoder'' (END) framework being the most widely used. To ensure end-to-end training, the noise…