Zobrazeno 1 - 5
of 5
pro vyhledávání: '"Yeats, Eric"'
We leverage diffusion models to study the robustness-performance tradeoff of robust classifiers. Our approach introduces a simple, pretrained diffusion method to generate low-norm counterfactual examples (CEs): semantically altered data which results
Externí odkaz:
http://arxiv.org/abs/2404.10588
Autor:
Zhang, Jingyang, Sun, Jingwei, Yeats, Eric, Ouyang, Yang, Kuo, Martin, Zhang, Jianyi, Yang, Hao Frank, Li, Hai
The problem of pre-training data detection for large language models (LLMs) has received growing attention due to its implications in critical issues like copyright violation and test data contamination. Despite improved performance, existing methods
Externí odkaz:
http://arxiv.org/abs/2404.02936
Quantification of the number of variables needed to locally explain complex data is often the first step to better understanding it. Existing techniques from intrinsic dimension estimation leverage statistical models to glean this information from sa
Externí odkaz:
http://arxiv.org/abs/2312.06869
Disentangled learning representations have promising utility in many applications, but they currently suffer from serious reliability issues. We present Gaussian Channel Autoencoder (GCAE), a method which achieves reliable disentanglement via flexibl
Externí odkaz:
http://arxiv.org/abs/2302.04362
We present a self-supervised method to disentangle factors of variation in high-dimensional data that does not rely on prior knowledge of the underlying variation profile (e.g., no assumptions on the number or distribution of the individual latent va
Externí odkaz:
http://arxiv.org/abs/2209.10677