Haoyu Lu (卢浩宇)

Email: ruclhy1998[at]163.com      Google Scholar     Github     

Hi there! I am a Ph.D. Student at Renmin University of China, advised by Prof. Zhiwu Lu. I also work closely with Dr. Mingyu Ding at UC Berkeley and Prof. Bo Zhang at ZJU. Prior to the Ph.D. study, I received my B.E. degree in Computer Science from Renmin University of China in 2021. My research interests lie in multimodal foundation model and video understanding.

Introducing DeepSeek-VL, an open-source Vision-Language (VL) Model designed for real-world vision and language understanding applications.

Introducing WuDao-WenLan, a large-scale Chinese Vision-Language Pre-Training Model.

We introduce Video Diffusion Transformer (VDT), which pioneers the use of transformers in diffusion-based video generation.

We propose UniAdapter, which unifies unimodal and multimodal adapters for parameter-efficient cross-modal adaptation on pre-trained vision-language models.

We propose Language-Guided Denoising Network (key-frame selection) for video-language modeling.

We propose a cross-modal MoCo-based continual learning algorithm with bidirectional momentum update (BMU).

We develop a foundation model pre-trained with huge multimodal data, which can be quickly adapted for various downstream cognitive tasks.

We propose COllaborative Two-Stream vision-language pre-training model (COTS) for cross-modal retrieval by enhancing cross-modal interaction.

We explores how to design a single neural network capable of adapting to multiple heterogeneous vision tasks, such as image segmentation, 3D detection, and video recognition.

We propose Motion Vector based Cross Guidance Contrastive learning for video self-supervised learning.

We propose a pretext task for self-supervised video representation learning by exploiting spatiotemporal continuity in videos.

The website template was adapted from Jingxiang Sun.