fidelity
In AI, fidelity refers to the degree to which a generated output matches the intended or desired outcomes. High-fidelity models produce results that closely resemble the expected reality, critical in tasks such as image synthesis or simulation.
- Chain-of-Zoom: Extreme Super-Resolution via Scale Autoregression and Preference Alignment
- DGSolver: Diffusion Generalist Solver with Universal Posterior Sampling for Image Restoration
- Generating Multi-Table Time Series EHR from Latent Space with Minimal Preprocessing
- Geometry-Aware Edge Pooling for Graph Neural Networks
- GnnXemplar: Exemplars to Explanations - Natural Language Rules for Global GNN Interpretability
- GnnXemplar: Exemplars to Explanations - Natural Language Rules for Global GNN Interpretability
- Noise Hypernetworks: Amortizing Test-Time Compute in Diffusion Models
- Projection-Manifold Regularized Latent Diffusion for Robust General Image Fusion
- Robust Explanations of Graph Neural Networks via Graph Curvatures
- Salient Concept-Aware Generative Data Augmentation
- Self-Supervised Selective-Guided Diffusion Model for Old-Photo Face Restoration
- SparseDiT: Token Sparsification for Efficient Diffusion Transformer
- WMCopier: Forging Invisible Watermarks on Arbitrary Images
- Walking the Schrödinger Bridge: A Direct Trajectory for Text-to-3D Generation