latent representations
Latent representations refer to internal features or patterns extracted by a model that are not directly observable in the input data. These representations encapsulate the underlying structure of the data and are often used for tasks like dimensionality reduction and generative modeling.
- Adaptive Time Encoding for Irregular Multivariate Time-Series Classification
- CORAL: Disentangling Latent Representations in Long-Tailed Diffusion
- Cross-Modal Representational Knowledge Distillation for Enhanced Spike-informed LFP Modeling
- Dense SAE Latents Are Features, Not Bugs
- Discovering Latent Graphs with GFlowNets for Diverse Conditional Image Generation
- Incomplete Multi-view Clustering via Hierarchical Semantic Alignment and Cooperative Completion
- Incomplete Multi-view Deep Clustering with Data Imputation and Alignment
- Lookahead Routing for Large Language Models
- Multi-View Oriented GPLVM: Expressiveness and Efficiency
- Promptable 3-D Object Localization with Latent Diffusion Models
- SCOUT: Teaching Pre-trained Language Models to Enhance Reasoning via Flow Chain-of-Thought
- SGCD: Stain-Guided CycleDiffusion for Unsupervised Domain Adaptation of Histopathology Image Classification
- STaRFormer: Semi-Supervised Task-Informed Representation Learning via Dynamic Attention-Based Regional Masking for Sequential Data
- Shape-Informed Clustering of Multi-Dimensional Functional Data via Deep Functional Autoencoders
- Sparse Diffusion Autoencoder for Test-time Adapting Prediction of Complex Systems
- Sparse Image Synthesis via Joint Latent and RoI Flow
- Test-Time Spectrum-Aware Latent Steering for Zero-Shot Generalization in Vision-Language Models
- Towards a Golden Classifier-Free Guidance Path via Foresight Fixed Point Iterations
- Venus-MAXWELL: Efficient Learning of Protein-Mutation Stability Landscapes using Protein Language Models
- Videos are Sample-Efficient Supervisions: Behavior Cloning from Videos via Latent Representations