| Proj No. | A3223-251 |
| Title | diffusion and flow-based generative models |
| Summary | Generative modeling has witnessed significant advancements with the emergence of diffusion models and flow-based models. This research explores the theoretical foundations, training dynamics, and comparative performance of these two classes of generative models. Diffusion models, inspired by non-equilibrium thermodynamics, progressively transform noise into structured data, while flow-based models leverage invertible transformations to enable exact likelihood estimation. The project aims to analyze their efficiency, scalability, and applicability across domains such as image synthesis, text generation, and scientific data modeling. Additionally, we plan to investigate hybrid approaches and optimization techniques to enhance sample quality and training stability. |
| Supervisor | A/P Wen Bihan (Loc:S2 > S2 B2B > S2 B2B 54, Ext: +65 67904708) |
| Co-Supervisor | - |
| RI Co-Supervisor | - |
| Lab | Centre for Information Sciences & System (CISS) (Loc: S2-B4b-05) |
| Single/Group: | Single |
| Area: | Digital Media Processing and Computer Engineering |
| ISP/RI/SMP/SCP?: |