Search Results - changhoon+kim

4 Results Sort By:
Attributing Image Generative Models Using Latent Fingerprints
Background Artificial intelligence (AI) image generation has opened up endless possibilities for creativity, allowing architects, designers, and artists to produce visual models rapidly. However, this technology is responsible for creating deepfakes, which are highly realistic but fake images and videos. Deepfakes can be used to spread misinformation,...
Published: 10/4/2024   |   Inventor(s): Guangyu Nie, Changhoon Kim, Yezhou Yang, Yi Ren
Keywords(s):  
Category(s): Physical Science, Artificial Intellegence/Machine Learning, Wireless & Networking, Intelligence & Security
METHOD FOR PRODUCING A POPULATION OF SYMMETRICALLY BARCODED TRANSPOSOMES
METHOD FOR PRODUCING A POPULATION OF SYMMETRICALLY BARCODED TRANSPOSOMES Researchers at UCSF have developed a method for producing a population of symmetrically barcoded transposomes that allow the reconstruction of a tagmented nucleic acid sequence without relying on alignment to a reference sequence. Many next generation sequencing workflows rely...
Published: 7/6/2023   |   Inventor(s): Derek Bogdanoff, Chang Kim, Tomasz Nowakowski
Keywords(s): Adaptor, Alignment, Barcoding, Next-Generation Sequencing, Sequence Reconstruction, Tagmentation, Transposase
Category(s): Technology Classifications > Biology
METHODS FOR FRAGMENTING COMPLIMENTARY DNA
METHODS FOR FRAGMENTING COMPLIMENTARY DNA Researchers at UCSF have discovered a new method for fragmenting complimentary DNA for sequencing applications. Sequencing of nucleic acids is one of the most important and transformational methods in biomedical sciences. In recent years, high throughput sequencing has become commonplace in both academic and...
Published: 7/6/2023   |   Inventor(s): Derek Bogdanoff, Chang Kim, Tomasz Nowakowski
Keywords(s): Barcode, cDNA, High-Throughput, Long-read Sequencing, Short Read Sequencing, Transposon, UMI
Category(s): Technology Classifications > Biology
Decentralized Attribution of Generative Models
­Background Growing applications of generative models have led to new threats such as malicious personation and digital copyright infringement. One solution to these threats is model attribution, i.e., the identification of user-end models where the contents under question are generated from. Existing studies showed empirical feasibility of attribution...
Published: 2/23/2023   |   Inventor(s): Yezhou Yang, Changhoon Kim, Yi Ren
Keywords(s): Security
Category(s): Physical Science, Computing & Information Technology, Intelligence & Security