Traditional wisdom in generative modeling literature is that spurious samples that a model can generate are errors and they should be avoided. Recent research, however, has shown interest in studying or even exploiting such samples instead of eliminating them. In this paper, we ask the question whether such samples can be eliminated all together without sacrificing coverage of the generating distribution. For the class of models we consider, we experimentally demonstrate that this is not possible without losing the ability to model some of the test samples. While our results need to be confirmed on a broader set of model families, these initial findings provide partial evidence that spurious samples share structural properties with the learned dataset, which, in turn, suggests they are not simply errors but a feature of deep generative nets.
Sparse Causal Discovery with Generative Intervention for Unsupervised Graph Domain Adaptation
State Key Laboratory for Multimedia Information Processing, School of Computer Science, PKU-Anker LLM Lab(多媒体信息处理国家重点实验室,计算机科学学院,PKU-Anker LLM实验室);Peking University(北京大学);University of California, Los Angeles(加州大学洛杉矶分校);Paul G. Allen School of Computer Science(保罗·G·艾伦计算机科学学院);University of Washington(华盛顿大学);Hong Kong University of Science and Technology(香港科技大学)
Breaking the Spurious Causality of Conditional Generation via Fairness Intervention with Corrective Sampling
Samsung Advanced Institute of Technology (SAIT)(三星先进技术研究院);Korea Advanced Institute of Science & Technology (KAIST)(韩国科学技术院);Pohang University of Science and Technology (POSTECH)(浦项科技大学)
The Clever Hans Mirage: A Comprehensive Survey on Spurious Correlations in Machine Learning
University of Virginia(弗吉尼亚大学);Purdue University(普渡大学);University of Illinois Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校);Pennsylvania State University(宾夕法尼亚州立大学);Amazon(亚马逊);New York University(纽约大学)
Model Collapse Is Not a Bug but a Feature in Machine Unlearning for LLMs
Dept. of Computer Science & Munich Data Science Institute, Technical University of Munich(计算机科学系及慕尼黑数据科学研究所,技术大学慕尼黑);Mila, Université de Montréal(蒙特利尔大学Mila)
Towards creativity characterization of generative models via group-based subset scanning
IBM Research Africa(IBM非洲研究院);IBM Research(IBM研究院)
深度生成模型,如变分自编码器(VAE),已被广泛用于计算创造力的研究。然而,这样的模型不鼓励非分布生成以避免虚假样本生成,限制了它们的创造力。因此,将对人类创造力的研究纳入生成式深度学习技术中,可以使其输出更有吸引力,更像人类。当我们看到面向创造力研究的生成模型的出现时,迫切需要基于机器学习的替代指标来表征这些模型的创造性输出。我们提出了基于组的子集扫描,通过检测生成模型的隐藏层中的异常节点激活的子集来量化,检测和表征创造性过程。我们对原始、典型解码和“创造性解码”(Das et al 2020)图像数据集的实验表明,所提出的子集分数分布对于检测激活空间而不是像素空间中的创造性过程更有用。此外,我们发现,在数据集上,创意样本比正常或非创意样本生成更大的异常子集。在创造性解码过程中突出显示的节点激活与负责正常样本生成的节点激活不同。
Quantifying Language Models' Sensitivity to Spurious Features in Prompt Design or: How I learned to start worrying about prompt formatting
Paul G. Allen School of Computer Science & Engineering, University of Washington(华盛顿大学保罗·G·艾伦计算机科学与工程学院);Allen Institute for Artificial Intelligence(艾伦人工智能研究所);University of California, Berkeley(加州大学伯克利分校)
ZeroDiff: Solidified Visual-Semantic Correlation in Zero-Shot Learning
Xi’an Jiaotong-Liverpool University(西交利物浦大学);University of Liverpool(利物浦大学);University of Nottingham(诺丁汉大学);Mohamed bin Zayed University of Artificial Intelligence(穆罕默德·本·扎耶德人工智能大学);Westlake University(西湖大学);Duke Kunshan University(昆山杜克大学)