Goto

Collaborating Authors

 editing


Designer Baby Companies Are in Turmoil

WIRED

Bootstrap Bio and Manhattan Genomics, which were pursuing gene editing in human embryos to prevent serious disease, have shut down. Two companies that launched last year with plans to create gene-edited babies have already shut down, citing money issues and internal conflict. One of them, Manhattan Genomics of New York, closed abruptly shortly after announcing a team of scientific advisers in October that included a prominent fertility doctor, a data scientist who worked for de-extinction company Colossal Biosciences, and a scientist who pioneered a "three-parent" IVF technique. The other, California-based Bootstrap Bio, said it ceased operations in late 2025, as first reported by Mother Jones. Manhattan Genomics and Bootstrap Bio had ambitions to edit DNA in human embryos with the goal of preventing serious disease in babies.


HairDiffusion: VividMulti-Colored HairEditingviaLatentDiffusion

Neural Information Processing Systems

Hair editing is a critical image synthesis task that aims to edit hair color and hairstyle using textdescriptions orreference images, while preserving irrelevant attributes(e.g.,identity,background,cloth).






f04351c9fa1e22797c7d32c1f6d23948-Paper-Datasets_and_Benchmarks_Track.pdf

Neural Information Processing Systems

Generative AI has revolutionised visual content editing, empowering users to effortlessly modify images and videos. However, not all edits are equal. To perform realistic edits in domains such as natural image or medical imaging, modifications must respect causal relationships inherent to the data generation process.



DreamSteerer: EnhancingSourceImageConditioned EditabilityusingPersonalizedDiffusionModels

Neural Information Processing Systems

However, such a solution often shows unsatisfactory editability on the source image. To address this, we propose DreamSteerer, a plug-in method for augmenting existing T2I personalization methods. Specifically, we enhance the source image conditioned editability of a personalized diffusion model via a novel Editability Driven Score Distillation (EDSD) objective. Moreover, we identify amodetrapping issuewithEDSD, andpropose amodeshifting regularization with spatial feature guided sampling to avoid such an issue.