Critical Synthesis
Critical synthesis in various fields focuses on generating realistic and diverse data, often using machine learning models to overcome limitations of real-world datasets. Current research emphasizes the development and application of diffusion models, generative adversarial networks (GANs), and transformer-based architectures for tasks ranging from image and speech synthesis to molecular design and controller synthesis. This work is significant for expanding datasets in data-scarce domains, improving the performance and robustness of AI systems, and enabling new applications in medicine, materials science, and beyond.
Papers
ResNCT: A Deep Learning Model for the Synthesis of Nephrographic Phase Images in CT Urography
Syed Jamal Safdar Gardezi, Lucas Aronson, Peter Wawrzyn, Hongkun Yu, E. Jason Abel, Daniel D. Shapiro, Meghan G. Lubner, Joshua Warner, Giuseppe Toia, Lu Mao, Pallavi Tiwari, Andrew L. Wentland
Certified Policy Verification and Synthesis for MDPs under Distributional Reach-avoidance Properties
S. Akshay, Krishnendu Chatterjee, Tobias Meggendorfer, Đorđe Žikelić
Synthesis of pulses from particle detectors with a Generative Adversarial Network (GAN)
Alberto Regadío, Luis Esteban, Sebastián Sánchez-Prieto
Noise-robust zero-shot text-to-speech synthesis conditioned on self-supervised speech-representation model with adapters
Kenichi Fujita, Hiroshi Sato, Takanori Ashihara, Hiroki Kanagawa, Marc Delcroix, Takafumi Moriya, Yusuke Ijima