Goto

Collaborating Authors

AI-generative Art Predicted To Be Next Trend For NFT Sector - AI Summary

#artificialintelligence

Art NFTs, in particular, made a big impact last year with Christie's reporting over $93 million in nonfungible token sales during its fourth annual Art Tech Summit that took place this past August. While notable, much of the crypto art scene appears to be dominated by cartoons and memes, as projects like CryptoPunks and Bored Ape Yacht Club have taken center stage. Known as "AI-generative NFTs," these nonfungible tokens are becoming increasingly popular within the art community, along with those interested in emerging technologies like artificial intelligence, blockchain and the Metaverse. Being able to work with an AI to bring your ideas to life is an experience like no other, it augments creativity in a way that feels like freedom, a type of play you haven't experienced since you were a child." While Eponym lets users create their own art NFTs, Metascapes is another project that was developed by three photographers looking to combine human expression with computer algorithms. While the potential for AI-generative NFTs is apparent, the question of whether or not artificial intelligence can be trusted to generate quality images based on text or photographs remains a concern. For instance, Fisher mentioned that Eponym has two versions of its generator available to the public, one on the company's Discord channel operating as a chatbot and the other as a private link that contains more complex algorithms capable of creating more advanced images. For example, Fisher remarked that Eponym's next project will feature interactive virtual identities where users can take their own portraits to create 3D avatars and animate them using artificial intelligence. Art NFTs, in particular, made a big impact last year with Christie's reporting over $93 million in nonfungible token sales during its fourth annual Art Tech Summit that took place this past August. While notable, much of the crypto art scene appears to be dominated by cartoons and memes, as projects like CryptoPunks and Bored Ape Yacht Club have taken center stage. Known as "AI-generative NFTs," these nonfungible tokens are becoming increasingly popular within the art community, along with those interested in emerging technologies like artificial intelligence, blockchain and the Metaverse. Being able to work with an AI to bring your ideas to life is an experience like no other, it augments creativity in a way that feels like freedom, a type of play you haven't experienced since you were a child."


Copulas as High-Dimensional Generative Models: Vine Copula Autoencoders

Neural Information Processing Systems

We introduce the vine copula autoencoder (VCAE), a flexible generative model for high-dimensional distributions built in a straightforward three-step procedure. Second, the multivariate distribution of the encoded data is estimated with vine copulas. Third, a generative model is obtained by combining the estimated distribution with the decoder part of the AE. As such, the proposed approach can transform any already trained AE into a flexible generative model at a low computational cost. This is an advantage over existing generative models such as adversarial networks and variational AEs which can be difficult to train and can impose strong assumptions on the latent space.


Advanced Generative Design for Manufacturing

#artificialintelligence

Restricting the design space available for a generative outcome can have its benefits. It eliminates the chance of unwanted contact with other components and allows extra room for stress response and deflection. The upfront setup effort required for obstacle offsets is faster and more elegant than post-processing an outcome after the generative study has been completed. Synthesis resolution controls the mesh size or level of detail of the generative solution. A low resolution will result in a faster solution at the expense of design space detail.


Nonparametric Density Estimation under Adversarial Losses

Neural Information Processing Systems

We study minimax convergence rates of nonparametric density estimation under a large class of loss functions called ``adversarial losses'', which, besides classical L^p losses, includes maximum mean discrepancy (MMD), Wasserstein distance, and total variation distance. These losses are closely related to the losses encoded by discriminator networks in generative adversarial networks (GANs). In a general framework, we study how the choice of loss and the assumed smoothness of the underlying density together determine the minimax rate. We also discuss implications for training GANs based on deep ReLU networks, and more general connections to learning implicit generative models in a minimax statistical sense.


Lifted Generative Parameter Learning

AAAI Conferences

Statistical relational learning (SRL) augments probabilistic models with relational representations and facilitates reasoning over sets of objects. When learning the probabilistic parameters for SRL models, however, one often resorts to reasoning over individual objects. To address this challenge, we compile a Markov logic network into a compact and efficient first-order data structure and use weighted first-order model counting to exactly optimize the likelihood of the parameters in a lifted manner. By exploiting the relational structure in the model, it is possible to learn more accurate parameters and dramatically improve the run time of the likelihood calculation. This allows us to calculate the exact likelihood for models where previously only approximate inference was feasible. Results on real-world data sets show that this approach learns more accurate models.