Meta tags:
description=Hsin-Ping Huang is a Research Engineer at Apple working on multimodal generative modeling for Apple Intelligence.;
author=Hsin-Ping Huang;

Headings (most frequently used words):

-

Text of the page (most frequently used words):
nbsp (30), and (21), huang (14), hsin (13), ping (12), for (11), paper (10), ming (10), yang (10), research (9), with (9), page (8), from (8), hsuan (8), generation (8), project (7), image (7), the (7), university (6), that (5), supervised (5), learning (5), sun (5), 2025 (5), human (4), domain (4), classification (4), media (4), generating (4), 2020 (4), jia (4), few (4), wang (4), liu (4), deqing (4), video (4), chuan (4), models (4), apple (4), cvpr (3), eccv (3), wacv (3), computer (3), fellowship (3), merced (3), 2024 (3), present (3), unsupervised (3), adaptive (3), discourse (3), semantic (3), scene (3), hung (3), tseng (3), shot (3), large (3), using (3), 2021 (3), vision (3), self (3), austin (3), intern (3), google (3), this (2), reviewer (2), iccv (2), neurips (2), finalist (2), 2022 (2), meta (2), phd (2), adversarial (2), relations (2), implicit (2), adaptation (2), view (2), synthesis (2), multiple (2), bin (2), semi (2), acoustic (2), event (2), representations (2), novel (2), views (2), transformation (2), singh (2), transformers (2), face (2), anti (2), spoofing (2), yaojie (2), autoflow (2), framework (2), charles (2), herrmann (2), long (2), take (2), videos (2), via (2), model (2), context (2), scale (2), xuhui (2), jiang (2), move (2), motion (2), conditioned (2), text (2), diffusion (2), zhan (2), zhou (2), kitten (2), visual (2), entities (2), hexiang (2), xinyi (2), national (2), taiwan (2), 2017 (2), texas (2), adobe (2), california (2), engineer (2), borrows, designs, website, jon, barron, conference, iclr, accv, aaai, ijcai, journal, tpami, ijcv, tip, graphics, forum, spl, area, chair, professional, activities, understanding, bobcat, honors, awards, network, reconstruction, component, leverages, explicit, classify, conll, 2019, junyi, jessy, relation, address, free, viewpoint, renderings, synthesized, label, map, synthesizing, plane, mpi, representation, ying, lee, study, audio, amount, unlabeled, data, these, icassp, chao, krishna, puvvada, tackle, stylization, stylized, images, constructing, point, cloud, aggregating, style, statistics, modulating, features, linear, maneesh, saurabh, saini, stylize, introducing, ensemble, adapters, feature, wise, layers, samples, hartwig, adam, jinwei, yuan, taihong, xiao, wen, sheng, chu, robust, cross, optical, flow, training, datasets, replaces, need, ground, truth, labels, loss, search, metric, 2023, stone, kyle, sargent, erika, junhwa, hur, propose, coherent, events, decoupling, into, keyframe, frame, interpolation, effective, keyframes, guidance, factor, allows, detailed, control, over, objects, appearances, location, optimizing, inserted, attention, layer, annotations, yukun, zhu, fine, grained, controllable, object, appearance, generates, sequences, prompt, trained, dataset, annotated, motions, feng, difan, jui, hsien, benchmark, evaluating, ability, generate, real, world, highlighting, even, advanced, struggle, entity, fidelity, tmlr, 2026, kelvin, chan, wei, chang, gaurav, tomar, hagai, taitelbaum, yonatan, bitton, knowledge, intensive, evaluation, publications, 2013, amazon, applied, scientist, deepmind, student, researcher, eecs, github, linkedin, scholar, email, honored, interests, lie, machine, focus, manipulation, fortunate, receive, advice, prior, received, under, supervision, prof, completed, science, electrical, engineering, lab, where, develop, multimodal, generative, intelligence, work, focuses, power, editing, experiences, including, genmoji, playground,

Text of the page (random words):
hsin ping huang hsin ping huang i am a research engineer at apple where i develop multimodal generative models for apple intelligence my work focuses on apple s diffusion models that power image generation and editing experiences including image playground and genmoji prior to this i received my ph d from university of california merced under the supervision of prof ming hsuan yang in the vision and learning lab i completed my m s in computer science from the university of texas at austin and my b s in electrical engineering from national taiwan university my research interests lie in computer vision and machine learning with a focus on image video and 3d generation and manipulation i am fortunate to intern at google with deqing sun yu chuan su hexiang hu lu jiang charles herrmann yaojie liu and xinyi wang at adobe research with zhan xu and yang zhou and to receive advice from hung yu tseng and jia bin huang i am honored to be a finalist for the meta phd research fellowship email nbsp nbsp cv nbsp nbsp google scholar nbsp nbsp linkedin nbsp nbsp github apple research engineer 2025 present university of california merced ph d in eecs 2020 2025 adobe research research intern 2024 google research deepmind student researcher 2021 2024 amazon applied scientist intern 2020 university of texas at austin m s in cs 2017 2020 national taiwan university b s in ee 2013 2017 publications kitten a knowledge intensive evaluation of image generation on visual entities hsin ping huang xinyi wang yonatan bitton hagai taitelbaum gaurav singh tomar ming wei chang xuhui jia kelvin c k chan hexiang hu yu chuan su ming hsuan yang tmlr 2026 kitten is a benchmark for evaluating text to image models ability to generate real world visual entities highlighting that even advanced models struggle with entity fidelity project page nbsp nbsp paper move in 2d 2d conditioned human motion generation hsin ping huang yang zhou jui hsien wang difan liu feng liu ming hsuan yang zhan xu cvpr 2025 move in 2d generates human motion sequences conditioned on a scene image and text prompt using a diffusion model trained on a large scale dataset of annotated human motions project page nbsp nbsp paper fine grained controllable video generation via object appearance and context hsin ping huang yu chuan su deqing sun lu jiang xuhui jia yukun zhu ming hsuan yang wacv 2025 factor is a video generation model that allows detailed control over objects appearances context and location by optimizing the inserted attention layer with large scale annotations project page nbsp nbsp paper nbsp nbsp media ak generating long take videos via effective keyframes and guidance hsin ping huang yu chuan su ming hsuan yang wacv 2025 we propose a framework for generating long take videos with multiple coherent events by decoupling video generation into keyframe generation and frame interpolation paper nbsp nbsp media ak self supervised autoflow hsin ping huang charles herrmann junhwa hur erika lu kyle sargent austin stone ming hsuan yang deqing sun cvpr 2023 self supervised autoflow is a framework for learning optical flow training datasets that replaces the need for ground truth labels by using a self supervised loss as the search metric project page nbsp nbsp paper adaptive transformers for robust few shot cross domain face anti spoofing hsin ping huang deqing sun yaojie liu wen sheng chu taihong xiao jinwei yuan hartwig adam ming hsuan yang eccv 2022 we present adaptive vision transformers for face anti spoofing introducing ensemble adapters and feature wise transformation layers for domain adaptation with few samples project page nbsp nbsp paper learning to stylize novel views hsin ping huang hung yu tseng saurabh saini maneesh singh ming hsuan yang iccv 2021 we tackle 3d scene stylization generating stylized images from novel views by constructing a point cloud aggregating style statistics and modulating features with a linear transformation project page nbsp nbsp paper nbsp nbsp media ak unsupervised and semi supervised few shot acoustic event classification hsin ping huang krishna c puvvada ming sun chao wang icassp 2021 we study semi supervised few shot acoustic event classification learning audio representations from a large amount of unlabeled data and using these representations for classification paper semantic view synthesis hsin ping huang hung yu tseng hsin ying lee jia bin huang eccv 2020 we address semantic view synthesis generating free viewpoint renderings of a synthesized scene from a semantic label map by synthesizing a multiple plane image mpi representation project page nbsp nbsp paper nbsp nbsp media ak unsupervised adversarial domain adaptation for implicit discourse relation classification hsin ping huang junyi jessy li conll 2019 we present an unsupervised adversarial domain adaptive network with a reconstruction component that leverages explicit discourse relations to classify implicit discourse relations paper honors and awards uc merced bobcat fellowship 2024 meta phd research fellowship finalist ar vr human understanding 2022 professional activities area chair neurips 26 journal reviewer tpami ijcv tip computer graphics forum spl conference reviewer cvpr 22 23 25 iccv 21 23 25 eccv 22 24 neurips 24 25 iclr 25 26 wacv 25 accv 24 aaai 23 24 25 26 ijcai 23 24 this page borrows designs from jon barron s website