V
主页
20240417【视频生成与理解】罗平:Efficient Diffusion Transformer for Image and Video Generat…
发布人
报告嘉宾:罗平 (香港大学) 报告时间:2024年4月17日 (星期三)晚上20:00 (北京时间) 报告题目:Efficient Diffusion Transformer for Image and Video Generation 报告人简介: Ping Luo is an Associate Professor in the Department of Computer Science at the University of Hong Kong, an Associate Director of the HKU Musketeers Foundation Institute of Data Science (HKU IDS), and a Deputy Director of the Joint Research Lab of HKU and Shanghai AI Lab. He obtained his Ph.D. in Information Engineering from the Chinese University of Hong Kong in 2014, under the supervision of Professor Xiaoou Tang and Xiaogang Wang. Before joining HKU in 2019, he was a Research Director in SenseTime. He has published 100+ papers in international conferences and journals such as TPAMI, ICML, ICLR, NeurIPS, and CVPR, with over 48,000 citations on Google Scholar. He was awarded the 2015 AAAI Easily Accessible Paper, nominated for the 2022 Computational Visual Media Journal's Best Paper of the Year, won the 2022 ACL Outstanding Paper, the 2023 World Artificial Intelligence Conference (WAIC) Outstanding Papers, and was a candidate for the Best Paper at ICCV’23. He was recognized as one of the innovators under 35 in the Asia-Pacific region by the MIT Technology Review (MIT TR35) in 2020. He has mentored 30 Ph.D. students, many of whom have received significant awards such as the Nvidia Fellowship, Baidu Fellowship, WAIC Yunfan Award, etc. 个人主页: https://www.cs.hku.hk/index.php/people/academic-staff/pluo https://scholar.google.com.hk/citations?user=aXdjxb4AAAAJ 报告摘要: This talk will introduce a series of our recent work on the model, data, and computing efficiency of image and video generation developed from 2022 to 2024, such as RAPHAEL (NeurIPS’23), Video DiT (ICLR’24), PixArt-alpha (ICLR’24), PixArt-delta (arXiv:2401.05252), PixArt-sigma (arXiv:2403.04692), GenTron (CVPR’24), DiffAgent (CVPR’24), and implicit prompting (arXiv:2403.02118), with emphasis on the technical details, philosophy, and experience in conducting these researches.
打开封面
下载高清视频
观看高清视频
视频下载器
Runway发布Gen-3 Alpha Image to Video,将任何图像用作视频生成的第一帧
20240424【空天智能感知】邹征夏:基于虚实混合增强的复杂场景感知与推演
【YOLOv5教程】花9888买的YOLOv5全套教程2024完整版现分享给大家!免费白嫖!值得收藏!——附赠课程与资料!——(人工智能、深度学习、机器学习)
20240403【以人为中心的生成、感知与决策】晏轶超:三维数字人的重建、编辑与驱动研究
【AI短片基础5】视频生成:Runway、即梦、可灵 | 文生、图生、视频生视频技巧
【VALSE2024】0506《Workshop :三维重建与生成》
20240529【可控的视觉内容生成】刘希慧:Towards Controllable and Compositional Visual Content ……
【VALSE2024】0505 马月昕《特邀报告:三维场景理解的前世、今生与未来》
20240529【可控的视觉内容生成】陈铠:Geometric-Controllable Visual Generation: A Systematic ……
【VALSE2024】0506《Workshop :具身智能的视觉与学习》
20240117【学术新人“修炼手册”】Panel
20231025【面向视觉的零样本学习】李晶晶:基于生成模型的零样本视觉识别
20200610 图像复原-去伪存真
【VALSE2024】0505 王兴刚《APR:面向大模型的新型高效率网络架构》
【VALSE论文速览-168期】Learning A Sparse Transformer Network for Effective Image Derain
【VALSE2024】0505 赵恒爽《APR:视觉基础大模型》
20230322【大规模预训练模型的可信性】庞天宇:Trustworthy Diffusion Models & Diffusion Models for……
20240612【可信基础模型】韩波:Exploring Trustworthy Foundation Models under Imperfect Data
【VALSE2024】0505 俞扬《APR:世界模型与具身决策》
20210816【VALSE短教程】《视觉语言导航》特邀讲师:吴琦副教授(澳大利亚阿德莱德大学)
20230906【多模态行人重识别的研究进展与未来】叶茫:多模态行人重识别进展与挑战
20200722 迁移学习:他山之石,可以攻玉
20221228【时光倒转万物生:扩散概率模型专题研讨】宋佳铭:用扩散模型高效求解逆问题
太厉害了 已跪!终于有人能把OpenCV图像处理讲的这么通俗易懂了,现在计算机视觉opencv全套分享给大家。
20240717【面向事件相机的物体检测与跟踪】王逍:Visual Object Tracking using an Event Camera
20211215【学无止境:深度连续学习】李玺:基于深度神经网络的持续性学习理论与方法
20220309【让机器看懂视频:视频分割与目标追踪】杨宗鑫:视频理解中的多目标联合分割
20200415 机器学习 vs 压缩感知:核磁共振成像与重建
【VALSE2023】0611《Workshop :机器人具身智能》
20210414【元学习研究的进展与未来】Panel
图像分割、目标检测、特征提取、边缘检测、图像滤波、人脸识别,这绝对是B站最适合入门学习的OpenCV计算机视觉课程!从入门到图像处理实战!人工智能丨深度学习丨
20201014 自监督学习
20240522【计算机辅助诊疗:过去,现在和未来】骆路阳:Understanding and Learning from Imperfect Medical
20220309【让机器看懂视频:视频分割与目标追踪】Panel
Transformer如此火爆,是不是可以舍弃卷积了呢?迪哥带你分析在不同应用场景下Transformer与卷积的实用性,零基础必看!
20210303【底层视觉与图像处理】Panel
20240103【大模型的可信之道:挑战、发展与机遇】Panel
20220105【标签噪声学习专题论坛】Panel
20230419【知识嵌入的跨模态学习】王文冠:Knowledge-Driven Perception
20220113 VALSE Student Webinar【论文写作那些事儿】庞树茂:科技论文写作经验分享