AI ML DL

AI ML DL

Preventing Stable Diffusion ‘Copyright Infringement’ by Poisoning The Source Data

A new paper proposes a method to protect the work of artists from being incorporated into Stable Diffusion, by using adversarial data perturbations to adversely affect generated results. Though the method even works with the image-to-image method of Stable Diffusion, it may require concerted public and organizational will to implement at scale – in common with many other proposed ‘data poisoning’ methods of this kind.

Read More »
Image Synthesis labeling
AI ML DL

Image Synthesis Has an SEO Problem

Images uploaded to the internet are being scraped at scale for ingestion into AI datasets. But the captions associated with the images were written for SEO purposes, and not for the benefit of machine learning systems. Badly-captioned images can, therefore, negatively influence the accuracy and usability of multimodal systems that are trained on them. Here we take a look at the problems, and some of the possible solutions.

Read More »
AI ML DL

Creating Better Avatars with a Dual-Domain Approach

A new academic collaboration, including contributors from Microsoft, has developed an improved technique capable of fitting user-submitted images into a ‘deepfake puppetry’ workflow in only thirty seconds, with notably improved fidelity to the original identity..

Read More »
AI ML DL

The ‘Cheap’ Decisions That Can Affect Image Synthesis

A new academic collaboration, including contributors from Microsoft, has developed an improved technique capable of fitting user-submitted images into a ‘deepfake puppetry’ workflow in only thirty seconds, with notably improved fidelity to the original identity..

Read More »
AI ML DL

Temporally Coherent Human Video Deepfakes Via Diffusion

A new research collaboration between Poland and the UK may offer the first effective method to obtain a much-cherished ‘holy grail’ of deepfake image synthesis research – the ability to generate temporally coherent human video from latent diffusion systems such as Stable Diffusion.

Read More »
AI ML DL

Entanglement in Image Synthesis

In the field of image synthesis, entanglement is the enmeshing of data properties with data other properties, which can make it difficult or impossible to isolate a particular aspect. If you change that data, you end up changing ‘nearby’ data. It’s not a trivial challenge to solve.

Read More »