generative model
A generative model is a type of model that learns to generate data points similar to a training dataset, capturing underlying distributions and complexities to create new samples, commonly utilized in unsupervised learning.
- A solvable model of learning generative diffusion: theory and insights
- Alchemist: Turning Public Text-to-Image Data into Generative Gold
- Amortized Active Generation of Pareto Sets
- Balanced Conic Rectified Flow
- BioCG: Constrained Generative Modeling for Biochemical Interaction Prediction
- Contextual Thompson Sampling via Generation of Missing Data
- Domain-RAG: Retrieval-Guided Compositional Image Generation for Cross-Domain Few-Shot Object Detection
- Flattening Hierarchies with Policy Bootstrapping
- Inference-Time Personalized Alignment with a Few User Preference Queries
- Kuramoto Orientation Diffusion Models
- MOF-BFN: Metal-Organic Frameworks Structure Prediction via Bayesian Flow Networks
- Model-Guided Dual-Role Alignment for High-Fidelity Open-Domain Video-to-Audio Generation
- Multiverse: Your Language Models Secretly Decide How to Parallelize and Merge Generation
- Near-Optimal Sample Complexity for Online Constrained MDPs
- Non-Asymptotic Analysis Of Data Augmentation For Precision Matrix Estimation
- OSCAR: One-Step Diffusion Codec Across Multiple Bit-rates
- On the Emergence of Linear Analogies in Word Embeddings
- ProSpero: Active Learning for Robust Protein Design Beyond Wild-Type Neighborhoods
- ProtInvTree: Deliberate Protein Inverse Folding with Reward-guided Tree Search
- Quantization-Free Autoregressive Action Transformer
- RUAGO: Effective and Practical Retain-Free Unlearning via Adversarial Attack and OOD Generator
- Self-Calibrating BCIs: Ranking and Recovery of Mental Targets Without Labels
- Synthetic-powered predictive inference
- Track, Inpaint, Resplat: Subject-driven 3D and 4D Generation with Progressive Texture Infilling
- When Worse is Better: Navigating the Compression Generation Trade-off In Visual Tokenization