Meta tags:
author=Ani Kembhavi;

Headings (most frequently used words):

recent, selected, works, for, 2025, 2026, awards, talks, service, press, best, paper, honorable, mention, award, cvpr, molmo, and, pixmo, jul, icvss, sicily, slides, program, chair, pc, iccv, here, are, some, from, my, teams, at, the, allen, institute, ai,

Text of the page (most frequently used words):
and (61), the (36), for (29), models (18), with (11), open (10), embodied (10), vision (10), 2023 (10), visual (9), our (9), award (9), using (8), language (8), multimodal (8), code (8), objaverse (8), computer (8), cvpr (8), 2024 (8), paper (8), learning (7), molmo (7), procthor (7), state (6), art (6), environments (6), manipulation (6), 2025 (6), best (6), simulation (5), real (5), robotics (5), training (5), data (5), objects (5), tasks (5), more (5), large (5), image (5), unified (5), university (5), this (4), was (4), robothor (4), research (4), interactive (4), that (4), model (4), physical (4), are (4), spoc (4), scale (4), works (4), navigation (4), ai2 (4), satlas (4), generated (4), from (4), natural (4), generation (4), institute (4), talk (4), jul (4), platform (3), framework (3), via (3), context (3), uses (3), attention (3), flow (3), bidaf (3), machine (3), comprehension (3), world (3), enables (3), recent (3), scaling (3), assets (3), trained (3), human (3), environment (3), thor (3), train (3), also (3), high (3), resolution (3), imagery (3), produce (3), images (3), over (3), diverse (3), datasets (3), programming (3), compositional (3), task (3), program (3), several (3), processing (3), results (3), across (3), benchmarks (3), procedural (3), text (3), their (3), selected (3), allen (3), neurips (3), chair (3), jun (3), aug (3), panel (3), talks (3), outstanding (3), awards (3), building (3), made (2), family (2), press (2), multi (2), different (2), without (2), before (2), bidirectional (2), modular (2), scenes (2), simulated (2), generate (2), available (2), imitating (2), shortest (2), reinforcement (2), agents (2), interact (2), perform (2), robot (2), global (2), which (2), input (2), geospatial (2), infrastructure (2), wind (2), turbines (2), view (2), synthesis (2), strong (2), resources (2), 10m (2), set (2), scans (2), than (2), neuro (2), symbolic (2), python (2), may (2), reasoning (2), virtual (2), tokens (2), video (2), applications (2), pixmo (2), teams (2), area (2), senior (2), college (2), park (2), workshop (2), washington (2), invited (2), workshops (2), labs (2), univ (2), 2026 (2), icra (2), mobile (2), corl (2), spent (2), years (2), director (2), ani (2), kembhavi (2), website, materialize, craft, scripts, compositions, videos, iconary, powered, drawing, guessing, game, creativity, allenact, source, lxmert, teaching, transformer, paint, generalizable, representations, gameplay, stage, hierarchical, process, represents, levels, granularity, directional, mechanism, obtain, query, aware, representation, early, summarization, extremely, popular, neural, emergence, bert, architetcure, built, movable, components, allowing, host, vastly, layouts, within, single, space, study, transfer, consists, paired, counterparts, key, factor, surprising, result, diversity, possible, procedurally, simulations, holodeck, massively, openly, resource, agent, path, experts, demonstrations, depth, sensors, makes, assumptions, about, target, paths, simuated, consisting, near, photo, realistic, indoor, where, can, navigate, extensively, used, community, policies, imitation, such, object, instruction, following, robots, contains, super, freely, low, sentinel, satellites, fidelity, entire, planet, visualizing, downloading, products, satellite, currently, includes, marine, offshore, platforms, renewable, energy, onshore, solar, farms, tree, cover, demonstrate, power, asset, librraies, zero123, novel, utilizing, 100, million, rendered, achieving, zero, shot, generalization, abilities, largest, public, quality, they, comprise, sources, including, manually, designed, photogrammetry, landmarks, everyday, items, professional, historic, antique, artifacts, less, year, these, have, become, facto, foundation, universes, annotated, visprog, approach, solving, complex, given, instructions, avoids, need, any, specific, instead, ability, like, programs, then, executed, get, both, solution, comprehensive, interpretable, rationale, each, line, invoke, one, off, shelf, routines, functions, intermediate, outputs, consumed, subsequent, parts, only, rgb, explicit, mapping, supervision, rearrangement, arm, sample, arbitrarily, customizable, performant, evaluate, scales, support, modalities, larger, parameter, scratch, pairs, 180m, clips, 130m, interleaved, trajectories, achieves, performance, grit, benchmark, first, spanning, classical, nlp, achieved, broad, unification, homogenizing, every, output, into, sequences, universal, compressors, autoregressive, audio, action, demo, while, current, interpret, express, full, potential, remains, untapped, goes, beyond, point, what, perceives, rich, interactions, worlds, empowering, next, capable, acting, interacting, most, powerful, closes, gap, between, proprietary, systems, wide, range, academic, well, evaluations, smaller, outperform, 10x, size, weights, here, some, past, iclr, emnlp, conferences, sac, iccv, service, amazon, adobe, maryland, architecture, indian, interior, design, hubli, india, summer, school, iiit, hyderabad, amld, generative, epfl, lausanne, switzerland, 50th, anniversary, celebration, umd, oct, colloquium, feb, gerald, masson, distinguished, lecture, johns, hopkins, mar, discussions, discussion, naver, europe, nov, oxford, bristol, dec, icvss, sicily, slides, iqa, question, answering, nvidia, pioneer, 2018, test, time, 2020, embodiment, robotic, harmonic, iros, poliformer, policy, transformers, masterful, navigators, 2022, honorable, mention, google, scholar, got, under, five, microsoft, working, search, prof, larry, davis, maryand, enjoy, deploying, users, work, two, decades, spans, affiliate, associate, professor, science, engineering, department, won, wayve, london, autonomous, driving, seattle, leading, focussed, researcher, meta, superintelligence, frontier, home, aniruddha,

Text of the page (random words):
aniruddha ani kembhavi ani kembhavi home awards selected works press i am an ai researcher at meta superintelligence labs building frontier multimodal models i was a director at wayve ai in london uk building world models for autonomous driving before that i spent 10 years at the allen institute for ai ai2 in seattle i was the senior director of computer vision leading research teams focussed on multimodal ai and embodied ai our research won several best and outstanding paper awards at cvpr neurips corl icra and more i am also an affiliate associate professor at the computer science engineering department at the university of washington i enjoy building and deploying ai applications for users to interact with my work over two decades spans computer vision robotics and natural language processing i got my ph d at the university of maryand college park under prof larry s davis and also spent five years at microsoft working in image and video search google scholar recent awards best paper honorable mention award for cvpr 2025 molmo and pixmo best paper award for cvpr 2023 visual programming compositional visual reasoning without training outstanding paper award for neurips 2022 procthor large scale embodied ai using procedural generation outstanding paper award for corl 2024 poliformer scaling on policy rl with transformers results in masterful navigators best paper award on mobile manipulation for iros 2024 harmonic mobile manipulation best paper award for icra 2024 open x embodiment robotic learning datasets and rt x models allen institute test of time award 2020 bidaf bidirectional attention flow for machine comprehension nvidia pioneer award 2018 iqa visual question answering in interactive environments recent talks jul 2026 icvss 2026 sicily slides dec 2025 univ of oxford and univ of bristol nov 2025 naver labs europe workshop on ai for robotics jun 2025 3 invited talks and 1 panel discussion at cvpr 2025 workshops jun 2024 5 invited talks and 2 panel discussions at cvpr 2024 workshops mar 2024 gerald m masson distinguished lecture at johns hopkins university feb 2024 university of washington robotics colloquium oct 2023 robotics and computer vision panel 50th anniversary celebration for cs umd aug 2023 amld generative ai workshop at epfl lausanne switzerland aug 2023 summer school on ai by iiit hyderabad aug 2023 ai architecture at indian institute for interior design hubli india jul 2023 talk at university of maryland college park jul 2023 talk at adobe research jul 2023 talk at amazon jun 2023 best paper award talk at cvpr 2023 service program chair pc iccv 2025 senior area chair sac cvpr 2024 area chair ac several past cvpr iclr neurips and emnlp conferences selected works here are some selected works from my teams at the allen institute for ai molmo and pixmo open weights and open data for state of the art multimodal models molmo is a family of open state of the art multimodal ai models our most powerful model closes the gap between open and proprietary systems across a wide range of academic benchmarks as well as human evaluations our smaller models outperform models 10x their size while current multimodal models interpret multimodal data and express it in natural language their full potential remains untapped molmo goes beyond by learning to point at what it perceives molmo enables rich interactions with physical and virtual worlds empowering the next generation of applications capable of acting and interacting with their environments molmo demo unified io scaling autoregressive multimodal models with vision language audio and action unified io is the first ai model to perform a large and diverse set of ai tasks spanning classical computer vision image synthesis vision and language and natural language processing nlp it achieved this broad unification by homogenizing every task s input and output into sequences of tokens using universal compressors unified io 2 scales this to support more modalities more tasks and larger models our 7b parameter model is trained from scratch on 1b image text pairs 1t text tokens 180m video clips 130m interleaved image text 3m 3d assets and 1m robot trajectories it achieves state of the art performance on the grit benchmark and strong results across more than 30 benchmarks in computer vision unified io unified io 2 code procthor large scale embodied ai using procedural generation procthor is a framework for procedural generation of embodied ai environments procthor enables us to sample arbitrarily large datasets of diverse interactive customizable and performant virtual environments to train and evaluate embodied agents models trained using only rgb images on procthor with no explicit mapping and no human task supervision produce state of the art results across 6 embodied ai benchmarks for navigation rearrangement and arm manipulation procthor code visual programming compositional neuro symbolic visual reasoning visual programming is a neuro symbolic approach to solving complex and compositional visual tasks given natural language instructions it avoids the need for any task specific training instead it uses the in context learning ability of large language models to generate python like modular programs which are then executed to get both the solution and a comprehensive and interpretable rationale each line of the generated program may invoke one of several off the shelf computer vision models image processing routines or python functions to produce intermediate outputs that may be consumed by subsequent parts of the program visprog code objaverse and objaverse xl universes with over 10m annotated 3d objects objaverse and objaverse xl are the largest public resources for 3d objects with 1m and 10m high quality assets they comprise objects from a diverse set of sources including manually designed objects photogrammetry scans of landmarks and everyday items and professional scans of historic and antique artifacts in less than a year these datasets have become the de facto resources for training foundation models for 3d computer vision we demonstrate the power of the objaverse asset librraies by training zero123 on novel view synthesis utilizing over 100 million multi view rendered images and achieving strong zero shot generalization abilities objaverse objaverse xl code satlas open geospatial data generated by ai satlas is a platform for visualizing and downloading global geospatial data products generated by our ai models using satellite images currently it includes marine infrastructure offshore wind turbines and platforms renewable energy infrastructure onshore wind turbines and solar farms and tree cover satlas also contains high resolution imagery on a global scale generated by our super resolution ai models which input freely available low resolution imagery from the sentinel 2 satellites and produce high fidelity imagery for the entire planet satlas code ai2 thor an interactive 3d simulated environment to train robots ai2 thor is a simuated environment consisting of near photo realistic 3d indoor scenes where ai agents can navigate and interact with objects to perform tasks it is extensively used in the community to train robot policies using reinforcement learning and imitation learning for tasks such as visual navigation object manipulation instruction following and more ai2 thor code spoc imitating shortest paths in simulation for real world navigation and manipulation spoc is an embodied navigation and manipulation agent trained by imitating shortest path experts in simulation spoc uses no human demonstrations no reinforcement learning no depth sensors and makes no assumptions about the target environment a key factor that enables this surprising result is the scale and diversity of our training data made possible by our recent works to procedurally generate simulations via procthor and holodeck and massively scaling up 3d assets via our openly available objaverse resource spoc robothor an open simulation to real embodied ai platform robothor is a framework to study simulation to real transfer for robotics it consists of simulated environments paired with physical counterparts in the real world the physical environments are built using modular and movable components allowing us to host scenes with vastly different layouts within a single physical space robothor code bidaf bidirectional attention flow for machine comprehension bidaf was an extremely popular and state of the art neural model for machine comprehension before the emergence of the bert architetcure it is a multi stage hierarchical process that represents the context at different levels of granularity and uses bi directional attention flow mechanism to obtain a query aware context representation without early summarization code press molmo a family of open state of the art multimodal ai models learning generalizable visual representations via interactive gameplay x lxmert teaching vision and language transformer models to paint allenact an open source framework for research in embodied ai ai creativity robothor an open simulation to real embodied ai platform iconary an ai powered drawing and guessing game craft scripts to compositions to videos this website was made using materialize