Meta tags:
author= Yichen Li;
Headings (most frequently used words):
-
Text of the page (most frequently used words):
and (24), for (19), multimodal (14), yichen (14), 2026 (12), research (12), learning (11), video (11), paper (11), model (9), architecture (9), project (9), page (9), code (8), physics (7), 2025 (6), 2024 (6), robotics (6), generation (6), post (6), training (6), with (6), attention (6), new (6), icml (5), summer (5), built (5), evolutionary (5), wojciech (5), matusik (5), open (5), efficient (5), 2020 (4), cvpr (4), eccv (4), the (4), based (4), models (4), assembly (4), antonio (4), torralba (4), world (4), projective (4), seeds (4), hardware (4), 2023 (3), course (3), assistant (3), geometric (3), methods (3), random (3), iclr (3), iccv (3), nvidia (3), diffusion (3), including (3), zeroth (3), order (3), optimization (3), other (3), embedding (3), leonidas (3), guibas (3), from (3), language (3), through (3), aware (3), variants (3), effort (3), junior (3), blog (3), generic (3), june (3), mit (3), prof (3), rank (2), fall (2), 2021 (2), visual (2), large (2), reviewer (2), acm (2), siggraph (2), tpami (2), sense (2), space (2), academic (2), point (2), cloud (2), system (2), data (2), decomposition (2), using (2), adobe (2), dense (2), reward (2), dynamics (2), over (2), kate (2), saenko (2), domain (2), lin (2), shao (2), kaichun (2), part (2), image (2), karl (2), willis (2), hui (2), shinjiro (2), sueda (2), jieliang (2), luo (2), yunsheng (2), tian (2), all (2), planning (2), dataset (2), joshua (2), physical (2), multi (2), chao (2), liu (2), jointly (2), towards (2), action (2), conditioned (2), advantage (2), weighted (2), matching (2), pretraining (2), autoregressive (2), eses (2), stable (2), lens (2), controlled (2), experiments (2), key (2), normalized (2), shuffled (2), controls (2), help (2), memory (2), pareto (2), front (2), compiler (2), awm (2), any (2), foundation (2), end (2), may (2), students (2), college, prize, excellence, computer, science, gpa, 1st, robert, shillman, fellowship, awards, s898, deep, cs468, algorithms, non, euclidean, spring, cs231n, convolutional, neural, networks, recognition, teaching, thoughts, workshop, organizer, rss, geometry, era, conference, neurips, journal, tog, ieee, multisensory, services, completion, utilizing, raycasting, patent, filed, layer, source, separation, gaussian, kernel, fast, processing, pointnet, replacement, systems, per, frame, unified, simulation, framework, soft, articulated, rigid, body, designed, anisotropic, young, modulus, across, regimes, error, reduction, team, baseline, professional, experience, xingchao, peng, domain2vec, unsupervised, adaptation, minhyuk, sung, single, 2022, asia, jie, assemble, them, generalizable, disassembly, icra, sachin, chitta, jacob, edward, farhad, javid, pingchuan, bassel, omari, asap, automated, sequence, complex, feasibility, 2019, stan, sclaroff, svetlana, lazebnik, kevin, shih, bryan, plummer, revisiting, ended, phrase, detection, tao, peter, chen, preconditioners, conjugate, gradient, pde, solver, jiequan, zhuang, wang, yueqi, duan, category, level, joint, shape, mike, foshey, francis, williams, tenenbaum, yilun, understand, tactile, signals, zhiyi, brandon, feng, preprint, generalized, shuchen, xue, chongjian, shilong, zhang, zhi, ming, aligning, publication, studies, rollout, reinforcement, densifying, feedback, shorter, generated, blocks, varl, reinforcing, explores, sample, strategies, low, parameter, updates, quantization, llm, lesson, technical, perspective, blogs, task, driven, sensor, differentiable, manifold, videos, dynamic, mode, self, supervised, kuramoto, coupling, offer, some, delibrately, unconventional, ideas, researchers, start, resource, scarce, academia, ware, calculus, architectures, deltas, quality, latency, tradeoff, cli, lora, perturbations, quantized, base, weights, population, established, mathematical, equivalence, between, ddpo, dsm, objectives, aligns, fine, tuning, autodecoder, way, one, given, modalities, input, that, can, serve, extended, embeddings, sole, authored, conditions, bounding, box, trajectories, reference, frames, fvd, improvement, trained, h100, gpus, series, accepted, organizing, workshops, starting, releasing, photo, credit, jiayuan, mao, github, twitter, google, scholar, email, yichenl, dot, edu, recognizing, scaling, dilemma, started, before, coming, worked, stanford, gordon, wetzstein, effective, mechanisms, inspired, design, alg, faster, convergence, backprop, approaches, interactions, multimodality, enjoy, principled, development, final, year, candidate, eecs, csail, advised, focuses, generative, perception, interactive, understanding,
Text of the page (random words):
yichen li yichen li i am a final year ph d candidate in eecs at mit csail advised by prof antonio torralba my ph d research focuses on generative foundation models for multimodal perception and interactive world understanding i enjoy principled development towards multimodal and world model approaches from video models interactions and multimodality post training and rl generic rl alg for faster convergence with zeroth order and dense backprop model architecture effective mechanisms of learning and hardware inspired architecture design recognizing the scaling dilemma of academic ml research i started an open research seeds effort to help junior students before coming to mit i worked with prof leonidas guibas and prof gordon wetzstein at stanford email yichenl at mit dot edu google scholar twitter github photo credit jiayuan mao new new june 2026 releasing ac a hardware aware pareto front architecture compiler new june 2026 starting an open research seeds effort for junior students new june 2026 organizing multimodal and sense space workshops at cvpr 2026 new may 2026 awm accepted to icml 2026 new may 2026 projective attention blog series multimodal world model rl post training model architecture iccv 2025 multimodal action conditioned video generation sole authored end to end jointly conditions on language bounding box trajectories and reference frames 40 fvd improvement trained on 8 h100 gpus iclr 2024 learning generic multimodal embeddings an autodecoder way of learning one generic embedding of any given modalities of input that can serve as a extended embedding for any large foundation model icml 2026 awm advantage weighted matching established mathematical equivalence between ddpo and dsm objectives aligns rl fine tuning with diffusion pretraining blog 2025 eses efficient and stable evolutionary rl zeroth order evolutionary rl using lora perturbations and quantized base weights for memory efficient population based post training blog 2026 projective attention a geometric lens on attention and controlled architecture experiments including key normalized variants and random shuffled controls cli 2026 hardware aware architecture compiler hardware ware model architecture calculus for architectures and deltas on the quality latency memory tradeoff pareto front open research seeds the open research seeds effort is to offer some delibrately unconventional research ideas to help junior researchers to start in resource scarce academia self supervised multimodal learning with kuramoto coupling learning physics from videos through dynamic mode decomposition task driven to sensor optimization through differentiable manifold optimization technical and perspective blogs projective attention and projective lesson of attention a geometric lens on attention and controlled architecture experiments including key normalized variants and random shuffled controls eses efficient and stable evolutionary rl for llm post training explores sample efficient evolutionary strategies for language model post training through low rank parameter updates and quantization aware variants varl reinforcing video autoregressive generation studies rollout efficient reinforcement learning for autoregressive video generation by densifying reward feedback over shorter generated blocks publication all multimodal video rl physics robotics other advantage weighted matching aligning rl with pretraining in diffusion models rl shuchen xue chongjian ge shilong zhang yichen li zhi ming ma icml 2026 paper code multimodal action conditioned video generation video multimodal yichen li antonio torralba iccv 2025 paper project page code generalized dynamics generation towards physical world model physics video robotics yichen li zhiyi li brandon feng antonio torralba preprint 2025 paper project page learning to jointly understand visual and tactile signals multimodal yichen li yilun du chao liu chao liu mike foshey francis williams joshua b tenenbaum wojciech matusik antonio torralba iclr 2024 paper project page dataset category level multi part multi joint 3d shape assembly robotics yichen li kaichun mo yueqi duan he wang jiequan zhuang lin shao wojciech matusik leonidas guibas cvpr 2024 paper project page data code learning preconditioners for conjugate gradient pde solver physics yichen li peter yichen chen tao du wojciech matusik icml 2023 paper video project page code revisiting image language for open ended phrase detection multimodal bryan plummer kevin shih yichen li ke xu svetlana lazebnik stan sclaroff kate saenko tpami 2019 paper asap automated sequence planning for complex assembly with physical feasibility robotics yunsheng tian karl d d willis bassel al omari jieliang luo pingchuan ma yichen li farhad javid edward gu joshua jacob shinjiro sueda hui li sachin chitta wojciech matusik icra 2024 paper project page dataset code assemble them all physics based planning for generalizable assembly by disassembly robotics yunsheng tian jie xu yichen li jieliang luo shinjiro sueda hui li karl d d willis wojciech matusik siggraph asia 2022 paper project page code 3d part assembly from a single image other yichen li kaichun mo lin shao minhyuk sung leonidas guibas eccv 2020 paper project page code domain2vec domain embedding for unsupervised domain adaptation other xingchao peng yichen li kate saenko eccv 2020 paper project page code professional experience nvidia research summer 2024 built a unified physics simulation framework for soft articulated and rigid body dynamics designed anisotropic young s modulus learning across physics regimes 51 error reduction over team baseline adobe research summer 2025 built rl post training systems for video diffusion models including zeroth order evolutionary optimization and dense per frame reward methods nvidia research summer 2023 built a gaussian kernel based architecture for fast point cloud processing as a pointnet replacement adobe research summer 2021 built a video layer decomposition system using source separation methods nvidia research summer 2020 built a point cloud completion system utilizing raycasting based data generation us patent filed academic services workshop organizer cvpr 2026 multimodal learning sense of space rss 2025 multimodal multisensory robotics eccv 2024 geometry in the large model era conference reviewer cvpr iccv eccv icml iclr neurips acm siggraph journal reviewer acm tog ieee tpami random thoughts 弱水 teaching cs231n convolutional neural networks for visual recognition spring 2021 course assistant ca cs468 geometric algorithms non euclidean methods fall 2020 course assistant ca 6 s898 deep learning fall 2023 course assistant ca awards robert j shillman fellowship college prize for excellence in computer science gpa rank 1st
|
