Meta tags:
author=Jiaojiao Fan;

Headings (most frequently used words):

image, generation, via, wasserstein, jiaojiao, fan, news, selected, publications, cosmos, world, foundation, model, platform, for, physical, ai, edify, high, quality, with, pixel, space, laplacian, diffusion, models, refdrop, controllable, consistency, in, or, video, reference, feature, guidance, variational, gradient, flow, scalable, computations, of, barycenter, input, convex, neural, networks,

Text of the page (most frequently used words):
image (12), and (10), for (10), nvidia (9), model (9), generation (8), edify (8), world (8), research (6), wasserstein (5), arxiv (5), #jiaojiao (5), fan (5), 2024 (5), models (5), cosmos (5), code (4), website (4), yongxin (4), chen (4), consistency (4), video (4), diffusion (4), pixel (4), space (4), text (4), the (4), physical (4), foundation (4), was (4), source (3), from (3), scalable (3), gradient (3), flow (3), refdrop (3), enables (3), photorealistic (3), with (3), paper (3), 2025 (3), open (3), amirhossein (2), taghvaei (2), using (2), input (2), convex (2), neural (2), networks (2), icml (2), via (2), qinsheng (2), zhang (2), 2022 (2), variational (2), training (2), reference (2), family (2), human (2), personalization (2), white (2), laplacian (2), platform (2), google (2), scholar (2), publications (2), scientist (2), working (2), georgia (2), release (2), during (2), customization (2), phd (2), where (2), design, leonid, keselman, 2020, novel, algorithm, computing, barycenters, long, talk, computations, barycenter, approach, formulations, haotian, xue, free, method, controlling, manipulating, attention, modules, consistent, character, multiple, subject, blending, diverse, content, creation, enhanced, temporal, videos, neurips, controllable, feature, guidance, cascaded, capable, generating, images, delivers, exceptional, results, remarkable, identity, also, supports, style, transfer, upsampling, controlnets, 360, hdr, panorama, high, quality, represents, foundational, breakthrough, modeling, developers, customize, deploy, that, can, understand, predict, interactions, sequences, jensen, huang, keynote, ces2025, see, full, list, selected, icml22, accepted, may, graduated, joined, genai, specifically, post, tech, jul, released, pyramid, nov, first, large, scale, predict1, jan, sota, autonomous, driving, robot, arm, manipulation, predict2, jun, news, linkedin, github, email, internship, developed, capabilities, which, successfully, shipped, part, deployed, enterprise, clients, advertising, applications, extremely, tasks, customized, output, photo, completed, machine, learning, advised, fortunate, internships, microsoft, snap, creative, vision, prof, institute, technology, santa, clara, california, work, currently, improving, sourced, groundbreaking, deep, imagination,

Text of the page (random words):
jiaojiao fan jiaojiao fan i am a research scientist at nvidia deep imagination research in santa clara california where i work on world foundation models currently i am working on improving cosmos nvidia s open sourced groundbreaking world foundation model for physical ai i completed my phd in machine learning at georgia institute of technology in 2024 where i was advised by prof yongxin chen during my phd i was fortunate to do internships at nvidia research snap creative vision research and microsoft research during my nvidia internship i developed customization capabilities for the edify text to image generation model which was successfully shipped as part of the edify model family and deployed by enterprise clients for advertising applications edify customization enables extremely photorealistic image generation for human personalization tasks e g from reference photo to customized output cv email github google scholar linkedin x news jun 2025 cosmos predict2 release sota world model for autonomous driving and robot arm manipulation jan 2025 cosmos predict1 release first open source large scale world model nov 2024 edify image paper was released photorealistic text to image generation with pixel space laplacian pyramid diffusion jul 2024 graduated from georgia tech and joined nvidia as a research scientist working on nvidia genai specifically edify text to image model post training may 2022 icml22 wasserstein gradient flow in pixel space is accepted selected publications see google scholar for the full list of publications cosmos world foundation model platform for physical ai nvidia white paper 2025 arxiv code website jensen huang keynote at ces2025 nvidia s open source video world model platform for physical ai represents a foundational breakthrough in world modeling cosmos enables developers to customize and deploy world foundation models that can understand and predict physical interactions in video sequences edify image high quality image generation with pixel space laplacian diffusion models nvidia white paper 2024 arxiv website edify image is a family of cascaded pixel space diffusion text to image models capable of generating photorealistic images edify image delivers exceptional human personalization results with remarkable identity consistency the model also supports style transfer 4k upsampling controlnets and 360 hdr panorama generation refdrop controllable consistency in image or video generation via reference feature guidance jiaojiao fan haotian xue qinsheng zhang yongxin chen neurips 2024 arxiv website refdrop is a training free method for controlling consistency in image and video generation by manipulating attention modules in diffusion models refdrop enables consistent character generation multiple subject blending diverse content creation and enhanced temporal consistency in videos variational wasserstein gradient flow jiaojiao fan qinsheng zhang amirhossein taghvaei yongxin chen icml 2022 arxiv code a scalable approach to wasserstein gradient flow using variational formulations scalable computations of wasserstein barycenter via input convex neural networks jiaojiao fan amirhossein taghvaei yongxin chen icml long talk 2020 arxiv code a novel scalable algorithm for computing wasserstein barycenters using input convex neural networks design and source code from leonid keselman s website