A cast of foundation models that can generate HD videos and synchronized audio. Enables additional capabilities like precise instruction-based video editing and generation of personalized videos.

Leveraging point trajectories and self-supervised representations for few-shot action recogntion.

Repurpose generative models as discriminative models to evaluate generative performance.

State-of-the-art Diffusion-based Text-to-Video generation method.

A novel and effective method to enable precise instance-level control for text-to-image generation.

Unsupervised Video Instance Segmentation without any video annotations or densely labeled pre-training.

Localize multiple objects in a real world images in an unsupervised fashion without any training using Self-supervised Transformers.

Improve performance of detectors trained using missing annotations by posing the problem as region-based semi-supervised learning.

Equip machines with capabilities to automatically discover and learn models for new categories from a large unlabeled dataset.

Automatically discover and attribute open-world GAN images.

Novel architecture based on denoising auto-encoders to improve few shot learning.

Empirical evidence that a single large neural network is usually more accurate than a boosted ensemble of neural networks with the same number of total parameters

Proposed a real-time system for image-based vehicle re-identification and natural language-based vehicle retrieval.

Proposed a Real-Time system for Vehicle Re-identification, Multi-Camera Tracking, and Anomaly Detection in a network of traffic cameras.

Humans interact with functionally similar objects in a similar manner.

A method for exploiting the spatial layout information of a human and an object for detecting HOIs in images.

Proposed a novel dual-path adaptive attention model for Vehicle re-identification.

Training deep networks with the ability to align features achieves state of the art performance on Person Re-identification.