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。
英文摘要:
Deep generative models, such as Variational Autoencoders (VAEs), have been employed widely in computational creativity research. However, such models discourage out-of-distribution generation to avoid spurious sample generation, limiting their creativity. Thus, incorporating research on human creativity into generative deep learning techniques presents an opportunity to make their outputs more compelling and human-like. As we see the emergence of generative models directed to creativity research, a need for machine learning-based surrogate metrics to characterize creative output from these models is imperative. We propose group-based subset scanning to quantify, detect, and characterize creative processes by detecting a subset of anomalous node-activations in the hidden layers of generative models. Our experiments on original, typically decoded, and "creatively decoded" (Das et al 2020) image datasets reveal that the proposed subset scores distribution is more useful for detecting creative processes in the activation space rather than the pixel space. Further, we found that creative samples generate larger subsets of anomalies than normal or non-creative samples across datasets. The node activations highlighted during the creative decoding process are different from those responsible for normal sample generation.
Evolutionary Computation: A Paradigm of Natural Generative AI and Creative Synthesis
Centre for Frontier AI Research (CFAR), Agency for Science, Technology and Research (A*STAR), Singapore(前沿人工智能研究中心(CFAR),科技研究局(A*STAR),新加坡);School of Mechanical Sciences, Indian Institute of Technology (IIT) Goa(机械科学学院,印度理工学院(IIT) Goa);School of Computer Science and Technology, Dalian University of Technology (DLUT)(计算机科学与技术学院,大连理工大学);College of Computing and Data Science (CCDS), Nanyang Technological University (NTU)(计算与数据科学学院(CCDS),南洋理工大学);Honda Research Institute Europe (HRI-EU), Offenbach am Main, Germany(本田欧洲研究院(HRI-EU),德国Offenbach am Main);Data Science and Artificial Intelligence Research Centre, School of Computer Science and Engineering, Nanyang Technological University (NTU)(数据科学与人工智能研究中心,计算机科学与工程学院,南洋理工大学)
Selective Imperfection as a Generative Framework for Analysis, Creativity and Discovery
Laboratory for Atomistic and Molecular Mechanics(原子分子力学实验室);Center for Computational Science and Engineering(计算科学与工程中心);Schwarzman College of Computing(施瓦茨曼计算学院);Massachusetts Institute of Technology(麻省理工学院)
Enhancing Creativity in 3D Generative Design via a TRIZ-Inspired Text-to-CAD Framework
Cho Chun Shik Graduate School of Mobility, KAIST, Daejeon, Republic of Korea(韩国科学技术院(KAIST)赵正植移动研究生院,大田,韩国);Samsung Electronics, Suwon, Republic of Korea(三星电子,水原,韩国);Department of Mechanical Engineering, Hanyang University, Ansan, Republic of Korea(汉阳大学机械工程系,安山,韩国);Narnia Labs, Daejeon, Republic of Korea(纳尼亚实验室,大田,韩国)
End-to-End Framework Integrating Generative AI and Deep Reinforcement Learning for Autonomous Ultrasound Scanning
organization= Concordia Institute for Information Systems Engineering, Concordia University , city= Montreal , country= Canada;organization= Department of Mechanical, Industrial;Aerospace Engineering, Concordia University , city= Montreal , country= Canada;organization= Department of Medicine, Laval University , city= Quebec , country= Canada;organization= Department of Electrical, Computer;Software Engineering, Ontario Tech University , city= Oshawa , country= Canada;organization= Department of Computer Science, Khalifa University , city= Abu Dhabi , country= UAE;organization= Artificial Intelligence \& Cyber Systems Research Center, Department of CSM, Lebanese American University , city= Beirut , country= Lebanon;organization= Department of Electrical Engineering;Computer Science, Khalifa University , city= Abu Dhabi , country= UAE
Automating Generative Deep Learning for Artistic Purposes: Challenges and Opportunities
Queen Mary University of London(伦敦玛丽女王大学);Goldsmiths, University of London(伦敦大学金史密斯学院);University of the Arts London(伦敦艺术大学);Aalto University(阿尔托大学);Finnish Center for Artificial Intelligence(芬兰人工智能中心);Monash University(莫纳什大学)