Showing 1–50 of 100 results for author: Vondrick, C

Robot Critics that Sweat the Small Stuff

Authors: Sruthi Sudhakar, Junbang Liang, Sreehari Rammohan, Pavel Tokmakov, Richard Zemel, Carl Vondrick

Abstract: Large vision-language models contain several priors about the world and object interactions, making them useful critics during inference to steer robot policies towards success. However, closed-loop robot manipulation requires judging small visual differences between success and failure, which remains a challenge for current VLMs. We introduce a method to fine-tune critics by constructing pairwise… ▽ More Large vision-language models contain several priors about the world and object interactions, making them useful critics during inference to steer robot policies towards success. However, closed-loop robot manipulation requires judging small visual differences between success and failure, which remains a challenge for current VLMs. We introduce a method to fine-tune critics by constructing pairwise progress supervision using success and failure rollouts obtained from a policy. Our fine-tuned critic excels at fine-grained progress reasoning and subtle failure detection, outperforming prior progress reasoning baselines. Additionally, we use an action-conditioned video model to predict the visual effect of several candidate actions sampled from a policy, and show that our critic can correctly identify successful candidates to execute, improving the average policy success rate by 11% across real-world tasks and 5.9% across simulation tasks. △ Less

Submitted 19 June, 2026; originally announced June 2026.

$A^2$: Smaller Self-Supervised ViTs Localize Better than Larger Ones

Authors: Sreehari Rammohan, Huy Ha, Carl Vondrick

Abstract: Robust visual classification often depends on localizing the main foreground objects in an image while ignoring contextual distractors. Surprisingly, we find that the attention maps of smaller self-supervised ViTs localize foreground objects better than those of larger ViTs. However, we still need large ViTs, because they extract richer representations from each patch. To get the best of both worl… ▽ More Robust visual classification often depends on localizing the main foreground objects in an image while ignoring contextual distractors. Surprisingly, we find that the attention maps of smaller self-supervised ViTs localize foreground objects better than those of larger ViTs. However, we still need large ViTs, because they extract richer representations from each patch. To get the best of both worlds, good localization and rich representations, we propose $A^2$, a simple method that leverages this inverse scaling finding by decoupling where to look (a small attention model) from what to extract (a large embedding model): we crop around the attention peaks of a small model and embed the crops with a larger model. $A^2$ uses entirely pretrained features, requires no group labels, and does not require per-dataset attention or backbone training. Across 5 benchmarks, $A^2$ is competitive with backbone-matched loss-level methods like DFR, and outperforms end-to-end attention training under stronger distribution shifts. △ Less

Submitted 2 June, 2026; originally announced June 2026.

Do multimodal models imagine electric sheep?

Authors: Santhosh Kumar Ramakrishnan, Carl Vondrick, Raja Giryes, Philipp Krähenbühl, Vladlen Koltun

Abstract: Yes. We find that large multimodal models develop mental imagery when solving spatial puzzles, and they do imagine sheep when solving sheep puzzles. We fine-tune a Qwen3.5 VLM to solve twelve diverse visual reasoning tasks -- including tangram, jigsaw, sokoban, 3D mental rotation, and rush hour -- that require understanding geometry, spatial relationships, and the consequences of actions. By super… ▽ More Yes. We find that large multimodal models develop mental imagery when solving spatial puzzles, and they do imagine sheep when solving sheep puzzles. We fine-tune a Qwen3.5 VLM to solve twelve diverse visual reasoning tasks -- including tangram, jigsaw, sokoban, 3D mental rotation, and rush hour -- that require understanding geometry, spatial relationships, and the consequences of actions. By supervising the model to predict the open-loop sequence of actions to solve a puzzle from an initial state, we show that the model's activations after each action encode meaningful visual information about the intermediate state. This finding suggests that an imperfect visual world model begins to form as a byproduct of learning to select correct actions, in the absence of any explicit visual supervision. Building on this observation, we propose two ways to sharpen and use the mental images formed by the model. We find that integrating as few as sixteen visual tokens per step into the chain of thought improves the average solve rate from 83% to 89%, with particularly strong gains on reasoning-heavy tasks such as jigsaw and 3D mental rotation. △ Less

Submitted 10 May, 2026; originally announced May 2026.

Few-Shot Design Optimization by Exploiting Auxiliary Information

Authors: Arjun Mani, Carl Vondrick, Richard Zemel

Abstract: Many real-world design problems involve optimizing an expensive black-box function $f(x)$, such as hardware design or drug discovery. Bayesian Optimization has emerged as a sample-efficient framework for this problem. However, the basic setting considered by these methods is simplified compared to real-world experimental setups, where experiments often generate a wealth of useful information. We i… ▽ More Many real-world design problems involve optimizing an expensive black-box function $f(x)$, such as hardware design or drug discovery. Bayesian Optimization has emerged as a sample-efficient framework for this problem. However, the basic setting considered by these methods is simplified compared to real-world experimental setups, where experiments often generate a wealth of useful information. We introduce a new setting where an experiment generates high-dimensional auxiliary information $h(x)$ along with the performance measure $f(x)$; moreover, a history of previously solved tasks from the same task family is available for accelerating optimization. A key challenge of our setting is learning how to represent and utilize $h(x)$ for efficiently solving new optimization tasks beyond the task history. We develop a novel approach for this setting based on a neural model which predicts $f(x)$ for unseen designs given a few-shot context containing observations of $h(x)$. We evaluate our method on two challenging domains, robotic hardware design and neural network hyperparameter tuning, and introduce a novel design problem and large-scale benchmark for the former. On both domains, our method utilizes auxiliary feedback effectively to achieve more accurate few-shot prediction and faster optimization of design tasks, significantly outperforming several methods for multi-task optimization. △ Less

Submitted 12 February, 2026; originally announced February 2026.

4D Gaussian Splatting as a Learned Dynamical System

Authors: Arnold Caleb Asiimwe, Carl Vondrick

Abstract: We reinterpret 4D Gaussian Splatting as a continuous-time dynamical system, where scene motion arises from integrating a learned neural dynamical field rather than applying per-frame deformations. This formulation, which we call EvoGS, treats the Gaussian representation as an evolving physical system whose state evolves continuously under a learned motion law. This unlocks capabilities absent in d… ▽ More We reinterpret 4D Gaussian Splatting as a continuous-time dynamical system, where scene motion arises from integrating a learned neural dynamical field rather than applying per-frame deformations. This formulation, which we call EvoGS, treats the Gaussian representation as an evolving physical system whose state evolves continuously under a learned motion law. This unlocks capabilities absent in deformation-based approaches:(1) sample-efficient learning from sparse temporal supervision by modeling the underlying motion law; (2) temporal extrapolation enabling forward and backward prediction beyond observed time ranges; and (3) compositional dynamics that allow localized dynamics injection for controllable scene synthesis. Experiments on dynamic scene benchmarks show that EvoGS achieves better motion coherence and temporal consistency compared to deformation-field baselines while maintaining real-time rendering △ Less

Submitted 22 December, 2025; originally announced December 2025.

New York Smells: A Large Multimodal Dataset for Olfaction

Authors: Ege Ozguroglu, Junbang Liang, Ruoshi Liu, Mia Chiquier, Michael DeTienne, Wesley Wei Qian, Alexandra Horowitz, Andrew Owens, Carl Vondrick

Abstract: While olfaction is central to how animals perceive the world, this rich chemical sensory modality remains largely inaccessible to machines. One key bottleneck is the lack of diverse, multimodal olfactory training data collected in natural settings. We present New York Smells, a large dataset of paired image and olfactory signals captured ``in the wild.'' Our dataset contains 7,000 smell-image pair… ▽ More While olfaction is central to how animals perceive the world, this rich chemical sensory modality remains largely inaccessible to machines. One key bottleneck is the lack of diverse, multimodal olfactory training data collected in natural settings. We present New York Smells, a large dataset of paired image and olfactory signals captured ``in the wild.'' Our dataset contains 7,000 smell-image pairs from 3,500 distinct objects across indoor and outdoor environments, with approximately 70$\times$ more objects than existing olfactory datasets. Our benchmark has three tasks: cross-modal smell-to-image retrieval, recognizing scenes, objects, and materials from smell alone, and fine-grained discrimination between grass species. Through experiments on our dataset, we find that visual data enables cross-modal olfactory representation learning, and that our learned olfactory representations outperform widely-used hand-crafted features. △ Less

Submitted 2 August, 2026; v1 submitted 25 November, 2025; originally announced November 2025.

Comments: Project website at https://smell.cs.columbia.edu

CAViAR: Critic-Augmented Video Agentic Reasoning

Authors: Sachit Menon, Ahmet Iscen, Arsha Nagrani, Tobias Weyand, Carl Vondrick, Cordelia Schmid

Abstract: Video understanding has seen significant progress in recent years, with models' performance on perception from short clips continuing to rise. Yet, multiple recent benchmarks, such as LVBench, Neptune, and ActivityNet-RTL, show performance wanes for tasks requiring complex reasoning on videos as queries grow more complex and videos grow longer. In this work, we ask: can existing perception capabil… ▽ More Video understanding has seen significant progress in recent years, with models' performance on perception from short clips continuing to rise. Yet, multiple recent benchmarks, such as LVBench, Neptune, and ActivityNet-RTL, show performance wanes for tasks requiring complex reasoning on videos as queries grow more complex and videos grow longer. In this work, we ask: can existing perception capabilities be leveraged to successfully perform more complex video reasoning? In particular, we develop a large language model agent given access to video modules as subagents or tools. Rather than following a fixed procedure to solve queries as in previous work such as Visual Programming, ViperGPT, and MoReVQA, the agent uses the results of each call to a module to determine subsequent steps. Inspired by work in the textual reasoning domain, we introduce a critic to distinguish between instances of successful and unsuccessful sequences from the agent. We show that the combination of our agent and critic achieve strong performance on the previously-mentioned datasets. △ Less

Submitted 9 September, 2025; originally announced September 2025.

Video Generators are Robot Policies

Authors: Junbang Liang, Pavel Tokmakov, Ruoshi Liu, Sruthi Sudhakar, Paarth Shah, Rares Ambrus, Carl Vondrick

Abstract: Despite tremendous progress in dexterous manipulation, current visuomotor policies remain fundamentally limited by two challenges: they struggle to generalize under perceptual or behavioral distribution shifts, and their performance is constrained by the size of human demonstration data. In this paper, we use video generation as a proxy for robot policy learning to address both limitations simulta… ▽ More Despite tremendous progress in dexterous manipulation, current visuomotor policies remain fundamentally limited by two challenges: they struggle to generalize under perceptual or behavioral distribution shifts, and their performance is constrained by the size of human demonstration data. In this paper, we use video generation as a proxy for robot policy learning to address both limitations simultaneously. We propose Video Policy, a modular framework that combines video and action generation that can be trained end-to-end. Our results demonstrate that learning to generate videos of robot behavior allows for the extraction of policies with minimal demonstration data, significantly improving robustness and sample efficiency. Our method shows strong generalization to unseen objects, backgrounds, and tasks, both in simulation and the real world. We further highlight that task success is closely tied to the generated video, with action-free video data providing critical benefits for generalizing to novel tasks. By leveraging large-scale video generative models, we achieve superior performance compared to traditional behavior cloning, paving the way for more scalable and data-efficient robot policy learning. △ Less

Submitted 1 August, 2025; originally announced August 2025.

MINERVA: Evaluating Complex Video Reasoning

Authors: Arsha Nagrani, Sachit Menon, Ahmet Iscen, Shyamal Buch, Ramin Mehran, Nilpa Jha, Anja Hauth, Yukun Zhu, Carl Vondrick, Mikhail Sirotenko, Cordelia Schmid, Tobias Weyand

Abstract: Multimodal LLMs are turning their focus to video benchmarks, however most video benchmarks only provide outcome supervision, with no intermediate or interpretable reasoning steps. This makes it challenging to assess if models are truly able to combine perceptual and temporal information to reason about videos, or simply get the correct answer by chance or by exploiting linguistic biases. To remedy… ▽ More Multimodal LLMs are turning their focus to video benchmarks, however most video benchmarks only provide outcome supervision, with no intermediate or interpretable reasoning steps. This makes it challenging to assess if models are truly able to combine perceptual and temporal information to reason about videos, or simply get the correct answer by chance or by exploiting linguistic biases. To remedy this, we provide a new video reasoning dataset called MINERVA for modern multimodal models. Each question in the dataset comes with 5 answer choices, as well as detailed, hand-crafted reasoning traces. Our dataset is multimodal, diverse in terms of video domain and length, and consists of complex multi-step questions. Extensive benchmarking shows that our dataset provides a challenge for frontier open-source and proprietary models. We perform fine-grained error analysis to identify common failure modes across various models, and create a taxonomy of reasoning errors. We use this to explore both human and LLM-as-a-judge methods for scoring video reasoning traces, and find that failure modes are primarily related to temporal localization, followed by visual perception errors, as opposed to logical or completeness errors. The dataset, along with questions, answer candidates and reasoning traces will be publicly available under https://github.com/google-deepmind/neptune?tab=readme-ov-file\#minerva. △ Less

Submitted 1 May, 2025; originally announced May 2025.

Towards LLM Agents for Earth Observation

Authors: Chia Hsiang Kao, Wenting Zhao, Shreelekha Revankar, Samuel Speas, Snehal Bhagat, Rajeev Datta, Cheng Perng Phoo, Utkarsh Mall, Carl Vondrick, Kavita Bala, Bharath Hariharan

Abstract: Earth Observation (EO) provides critical planetary data for environmental monitoring, disaster management, climate science, and other scientific domains. Here we ask: Are AI systems ready for reliable Earth Observation? We introduce \datasetnamenospace, a benchmark of 140 yes/no questions from NASA Earth Observatory articles across 13 topics and 17 satellite sensors. Using Google Earth Engine API… ▽ More Earth Observation (EO) provides critical planetary data for environmental monitoring, disaster management, climate science, and other scientific domains. Here we ask: Are AI systems ready for reliable Earth Observation? We introduce \datasetnamenospace, a benchmark of 140 yes/no questions from NASA Earth Observatory articles across 13 topics and 17 satellite sensors. Using Google Earth Engine API as a tool, LLM agents can only achieve an accuracy of 33% because the code fails to run over 58% of the time. We improve the failure rate for open models by fine-tuning synthetic data, allowing much smaller models (Llama-3.1-8B) to achieve comparable accuracy to much larger ones (e.g., DeepSeek-R1). Taken together, our findings identify significant challenges to be solved before AI agents can automate earth observation, and suggest paths forward. The project page is available at https://iandrover.github.io/UnivEarth. △ Less

Submitted 12 September, 2025; v1 submitted 16 April, 2025; originally announced April 2025.

Comments: Accepted at ICML 2025 Workshop TerraBytes

Can LLM feedback enhance review quality? A randomized study of 20K reviews at ICLR 2025

Authors: Nitya Thakkar, Mert Yuksekgonul, Jake Silberg, Animesh Garg, Nanyun Peng, Fei Sha, Rose Yu, Carl Vondrick, James Zou

Abstract: Peer review at AI conferences is stressed by rapidly rising submission volumes, leading to deteriorating review quality and increased author dissatisfaction. To address these issues, we developed Review Feedback Agent, a system leveraging multiple large language models (LLMs) to improve review clarity and actionability by providing automated feedback on vague comments, content misunderstandings, a… ▽ More Peer review at AI conferences is stressed by rapidly rising submission volumes, leading to deteriorating review quality and increased author dissatisfaction. To address these issues, we developed Review Feedback Agent, a system leveraging multiple large language models (LLMs) to improve review clarity and actionability by providing automated feedback on vague comments, content misunderstandings, and unprofessional remarks to reviewers. Implemented at ICLR 2025 as a large randomized control study, our system provided optional feedback to more than 20,000 randomly selected reviews. To ensure high-quality feedback for reviewers at this scale, we also developed a suite of automated reliability tests powered by LLMs that acted as guardrails to ensure feedback quality, with feedback only being sent to reviewers if it passed all the tests. The results show that 27% of reviewers who received feedback updated their reviews, and over 12,000 feedback suggestions from the agent were incorporated by those reviewers. This suggests that many reviewers found the AI-generated feedback sufficiently helpful to merit updating their reviews. Incorporating AI feedback led to significantly longer reviews (an average increase of 80 words among those who updated after receiving feedback) and more informative reviews, as evaluated by blinded researchers. Moreover, reviewers who were selected to receive AI feedback were also more engaged during paper rebuttals, as seen in longer author-reviewer discussions. This work demonstrates that carefully designed LLM-generated review feedback can enhance peer review quality by making reviews more specific and actionable while increasing engagement between reviewers and authors. The Review Feedback Agent is publicly available at https://github.com/zou-group/review_feedback_agent. △ Less

Submitted 13 April, 2025; originally announced April 2025.

Comments: 30 pages, 7 figures

Teaching Humans Subtle Differences with DIFFusion

Authors: Mia Chiquier, Orr Avrech, Yossi Gandelsman, Berthy Feng, Katherine Bouman, Carl Vondrick

Abstract: Scientific expertise often requires recognizing subtle visual differences that remain challenging to articulate even for domain experts. We present a system that leverages generative models to automatically discover and visualize minimal discriminative features between categories while preserving instance identity. Our method generates counterfactual visualizations with subtle, targeted transforma… ▽ More Scientific expertise often requires recognizing subtle visual differences that remain challenging to articulate even for domain experts. We present a system that leverages generative models to automatically discover and visualize minimal discriminative features between categories while preserving instance identity. Our method generates counterfactual visualizations with subtle, targeted transformations between classes, performing well even in domains where data is sparse, examples are unpaired, and category boundaries resist verbal description. Experiments across six domains, including black hole simulations, butterfly taxonomy, and medical imaging, demonstrate accurate transitions with limited training data, highlighting both established discriminative features and novel subtle distinctions that measurably improved category differentiation. User studies confirm our generated counterfactuals significantly outperform traditional approaches in teaching humans to correctly differentiate between fine-grained classes, showing the potential of generative models to advance visual learning and scientific research. △ Less

Submitted 15 May, 2025; v1 submitted 10 April, 2025; originally announced April 2025.

DiSciPLE: Learning Interpretable Programs for Scientific Visual Discovery

Authors: Utkarsh Mall, Cheng Perng Phoo, Mia Chiquier, Bharath Hariharan, Kavita Bala, Carl Vondrick

Abstract: Visual data is used in numerous different scientific workflows ranging from remote sensing to ecology. As the amount of observation data increases, the challenge is not just to make accurate predictions but also to understand the underlying mechanisms for those predictions. Good interpretation is important in scientific workflows, as it allows for better decision-making by providing insights into… ▽ More Visual data is used in numerous different scientific workflows ranging from remote sensing to ecology. As the amount of observation data increases, the challenge is not just to make accurate predictions but also to understand the underlying mechanisms for those predictions. Good interpretation is important in scientific workflows, as it allows for better decision-making by providing insights into the data. This paper introduces an automatic way of obtaining such interpretable-by-design models, by learning programs that interleave neural networks. We propose DiSciPLE (Discovering Scientific Programs using LLMs and Evolution) an evolutionary algorithm that leverages common sense and prior knowledge of large language models (LLMs) to create Python programs explaining visual data. Additionally, we propose two improvements: a program critic and a program simplifier to improve our method further to synthesize good programs. On three different real-world problems, DiSciPLE learns state-of-the-art programs on novel tasks with no prior literature. For example, we can learn programs with 35% lower error than the closest non-interpretable baseline for population density estimation. △ Less

Submitted 14 February, 2025; originally announced February 2025.

Generative Data Mining with Longtail-Guided Diffusion

Authors: David S. Hayden, Mao Ye, Timur Garipov, Gregory P. Meyer, Carl Vondrick, Zhao Chen, Yuning Chai, Eric Wolff, Siddhartha S. Srinivasa

Abstract: It is difficult to anticipate the myriad challenges that a predictive model will encounter once deployed. Common practice entails a reactive, cyclical approach: model deployment, data mining, and retraining. We instead develop a proactive longtail discovery process by imagining additional data during training. In particular, we develop general model-based longtail signals, including a differentiab… ▽ More It is difficult to anticipate the myriad challenges that a predictive model will encounter once deployed. Common practice entails a reactive, cyclical approach: model deployment, data mining, and retraining. We instead develop a proactive longtail discovery process by imagining additional data during training. In particular, we develop general model-based longtail signals, including a differentiable, single forward pass formulation of epistemic uncertainty that does not impact model parameters or predictive performance but can flag rare or hard inputs. We leverage these signals as guidance to generate additional training data from a latent diffusion model in a process we call Longtail Guidance (LTG). Crucially, we can perform LTG without retraining the diffusion model or the predictive model, and we do not need to expose the predictive model to intermediate diffusion states. Data generated by LTG exhibit semantically meaningful variation, yield significant generalization improvements on numerous image classification benchmarks, and can be analyzed by a VLM to proactively discover, textually explain, and address conceptual gaps in a deployed predictive model. △ Less

Submitted 26 June, 2025; v1 submitted 3 February, 2025; originally announced February 2025.

Comments: 20 pages

Journal ref: Proceedings of the 42nd International Conference on Machine Learning (ICML), 2025

MedAutoCorrect: Image-Conditioned Autocorrection in Medical Reporting

Authors: Arnold Caleb Asiimwe, Dídac Surís, Pranav Rajpurkar, Carl Vondrick

Abstract: In medical reporting, the accuracy of radiological reports, whether generated by humans or machine learning algorithms, is critical. We tackle a new task in this paper: image-conditioned autocorrection of inaccuracies within these reports. Using the MIMIC-CXR dataset, we first intentionally introduce a diverse range of errors into reports. Subsequently, we propose a two-stage framework capable of… ▽ More In medical reporting, the accuracy of radiological reports, whether generated by humans or machine learning algorithms, is critical. We tackle a new task in this paper: image-conditioned autocorrection of inaccuracies within these reports. Using the MIMIC-CXR dataset, we first intentionally introduce a diverse range of errors into reports. Subsequently, we propose a two-stage framework capable of pinpointing these errors and then making corrections, simulating an \textit{autocorrection} process. This method aims to address the shortcomings of existing automated medical reporting systems, like factual errors and incorrect conclusions, enhancing report reliability in vital healthcare applications. Importantly, our approach could serve as a guardrail, ensuring the accuracy and trustworthiness of automated report generation. Experiments on established datasets and state of the art report generation models validate this method's potential in correcting medical reporting errors. △ Less

Submitted 3 December, 2024; originally announced December 2024.

Self-Improving Autonomous Underwater Manipulation

Authors: Ruoshi Liu, Huy Ha, Mengxue Hou, Shuran Song, Carl Vondrick

Abstract: Underwater robotic manipulation faces significant challenges due to complex fluid dynamics and unstructured environments, causing most manipulation systems to rely heavily on human teleoperation. In this paper, we introduce AquaBot, a fully autonomous manipulation system that combines behavior cloning from human demonstrations with self-learning optimization to improve beyond human teleoperation p… ▽ More Underwater robotic manipulation faces significant challenges due to complex fluid dynamics and unstructured environments, causing most manipulation systems to rely heavily on human teleoperation. In this paper, we introduce AquaBot, a fully autonomous manipulation system that combines behavior cloning from human demonstrations with self-learning optimization to improve beyond human teleoperation performance. With extensive real-world experiments, we demonstrate AquaBot's versatility across diverse manipulation tasks, including object grasping, trash sorting, and rescue retrieval. Our real-world experiments show that AquaBot's self-optimized policy outperforms a human operator by 41% in speed. AquaBot represents a promising step towards autonomous and self-improving underwater manipulation systems. We open-source both hardware and software implementation details. △ Less

Submitted 24 October, 2024; originally announced October 2024.

Comments: Project Page: https://aquabot.cs.columbia.edu/

Differentiable Robot Rendering

Authors: Ruoshi Liu, Alper Canberk, Shuran Song, Carl Vondrick

Abstract: Vision foundation models trained on massive amounts of visual data have shown unprecedented reasoning and planning skills in open-world settings. A key challenge in applying them to robotic tasks is the modality gap between visual data and action data. We introduce differentiable robot rendering, a method allowing the visual appearance of a robot body to be directly differentiable with respect to… ▽ More Vision foundation models trained on massive amounts of visual data have shown unprecedented reasoning and planning skills in open-world settings. A key challenge in applying them to robotic tasks is the modality gap between visual data and action data. We introduce differentiable robot rendering, a method allowing the visual appearance of a robot body to be directly differentiable with respect to its control parameters. Our model integrates a kinematics-aware deformable model and Gaussians Splatting and is compatible with any robot form factors and degrees of freedom. We demonstrate its capability and usage in applications including reconstruction of robot poses from images and controlling robots through vision language models. Quantitative and qualitative results show that our differentiable rendering model provides effective gradients for robotic control directly from pixels, setting the foundation for the future applications of vision foundation models in robotics. △ Less

Submitted 17 October, 2024; originally announced October 2024.

Comments: Project Page: https://drrobot.cs.columbia.edu/

EraseDraw: Learning to Draw Step-by-Step via Erasing Objects from Images

Authors: Alper Canberk, Maksym Bondarenko, Ege Ozguroglu, Ruoshi Liu, Carl Vondrick

Abstract: Creative processes such as painting often involve creating different components of an image one by one. Can we build a computational model to perform this task? Prior works often fail by making global changes to the image, inserting objects in unrealistic spatial locations, and generating inaccurate lighting details. We observe that while state-of-the-art models perform poorly on object insertion,… ▽ More Creative processes such as painting often involve creating different components of an image one by one. Can we build a computational model to perform this task? Prior works often fail by making global changes to the image, inserting objects in unrealistic spatial locations, and generating inaccurate lighting details. We observe that while state-of-the-art models perform poorly on object insertion, they can remove objects and erase the background in natural images very well. Inverting the direction of object removal, we obtain high-quality data for learning to insert objects that are spatially, physically, and optically consistent with the surroundings. With this scalable automatic data generation pipeline, we can create a dataset for learning object insertion, which is used to train our proposed text conditioned diffusion model. Qualitative and quantitative experiments have shown that our model achieves state-of-the-art results in object insertion, particularly for in-the-wild images. We show compelling results on diverse insertion prompts and images across various domains.In addition, we automate iterative insertion by combining our insertion model with beam search guided by CLIP. △ Less

Submitted 23 December, 2024; v1 submitted 31 August, 2024; originally announced September 2024.

Controlling the World by Sleight of Hand

Authors: Sruthi Sudhakar, Ruoshi Liu, Basile Van Hoorick, Carl Vondrick, Richard Zemel

Abstract: Humans naturally build mental models of object interactions and dynamics, allowing them to imagine how their surroundings will change if they take a certain action. While generative models today have shown impressive results on generating/editing images unconditionally or conditioned on text, current methods do not provide the ability to perform object manipulation conditioned on actions, an impor… ▽ More Humans naturally build mental models of object interactions and dynamics, allowing them to imagine how their surroundings will change if they take a certain action. While generative models today have shown impressive results on generating/editing images unconditionally or conditioned on text, current methods do not provide the ability to perform object manipulation conditioned on actions, an important tool for world modeling and action planning. Therefore, we propose to learn an action-conditional generative models by learning from unlabeled videos of human hands interacting with objects. The vast quantity of such data on the internet allows for efficient scaling which can enable high-performing action-conditional models. Given an image, and the shape/location of a desired hand interaction, CosHand, synthesizes an image of a future after the interaction has occurred. Experiments show that the resulting model can predict the effects of hand-object interactions well, with strong generalization particularly to translation, stretching, and squeezing interactions of unseen objects in unseen environments. Further, CosHand can be sampled many times to predict multiple possible effects, modeling the uncertainty of forces in the interaction/environment. Finally, method generalizes to different embodiments, including non-human hands, i.e. robot hands, suggesting that generative video models can be powerful models for robotics. △ Less

Submitted 13 August, 2024; originally announced August 2024.

Dreamitate: Real-World Visuomotor Policy Learning via Video Generation

Authors: Junbang Liang, Ruoshi Liu, Ege Ozguroglu, Sruthi Sudhakar, Achal Dave, Pavel Tokmakov, Shuran Song, Carl Vondrick

Abstract: A key challenge in manipulation is learning a policy that can robustly generalize to diverse visual environments. A promising mechanism for learning robust policies is to leverage video generative models, which are pretrained on large-scale datasets of internet videos. In this paper, we propose a visuomotor policy learning framework that fine-tunes a video diffusion model on human demonstrations o… ▽ More A key challenge in manipulation is learning a policy that can robustly generalize to diverse visual environments. A promising mechanism for learning robust policies is to leverage video generative models, which are pretrained on large-scale datasets of internet videos. In this paper, we propose a visuomotor policy learning framework that fine-tunes a video diffusion model on human demonstrations of a given task. At test time, we generate an example of an execution of the task conditioned on images of a novel scene, and use this synthesized execution directly to control the robot. Our key insight is that using common tools allows us to effortlessly bridge the embodiment gap between the human hand and the robot manipulator. We evaluate our approach on four tasks of increasing complexity and demonstrate that harnessing internet-scale generative models allows the learned policy to achieve a significantly higher degree of generalization than existing behavior cloning approaches. △ Less

Submitted 24 June, 2024; originally announced June 2024.

Comments: Project page: https://dreamitate.cs.columbia.edu/

Whiteboard-of-Thought: Thinking Step-by-Step Across Modalities

Authors: Sachit Menon, Richard Zemel, Carl Vondrick

Abstract: When presented with questions involving visual thinking, humans naturally switch reasoning modalities, often forming mental images or drawing visual aids. Large language models have shown promising results in arithmetic and symbolic reasoning by expressing intermediate reasoning in text as a chain of thought, yet struggle to extend this capability to answer text queries that are easily solved by v… ▽ More When presented with questions involving visual thinking, humans naturally switch reasoning modalities, often forming mental images or drawing visual aids. Large language models have shown promising results in arithmetic and symbolic reasoning by expressing intermediate reasoning in text as a chain of thought, yet struggle to extend this capability to answer text queries that are easily solved by visual reasoning, even with extensive multimodal pretraining. We introduce a simple method, whiteboard-of-thought prompting, to unlock the visual reasoning capabilities of multimodal large language models across modalities. Whiteboard-of-thought prompting provides multimodal large language models with a metaphorical `whiteboard' to draw out reasoning steps as images, then returns these images back to the model for further processing. We find this can be accomplished with no demonstrations or specialized modules, instead leveraging models' existing ability to write code with libraries such as Matplotlib and Turtle. This simple approach shows state-of-the-art results on four difficult natural language tasks that involve visual and spatial reasoning. We identify multiple settings where GPT-4o using chain-of-thought fails dramatically, including more than one where it achieves $0\%$ accuracy, while whiteboard-of-thought enables up to $92\%$ accuracy in these same settings. We present a detailed exploration of where the technique succeeds as well as its sources of error. △ Less

Submitted 20 June, 2024; originally announced June 2024.

Comments: Project website: whiteboard.cs.columbia.edu/

How Video Meetings Change Your Expression

Authors: Sumit Sarin, Utkarsh Mall, Purva Tendulkar, Carl Vondrick

Abstract: Do our facial expressions change when we speak over video calls? Given two unpaired sets of videos of people, we seek to automatically find spatio-temporal patterns that are distinctive of each set. Existing methods use discriminative approaches and perform post-hoc explainability analysis. Such methods are insufficient as they are unable to provide insights beyond obvious dataset biases, and the… ▽ More Do our facial expressions change when we speak over video calls? Given two unpaired sets of videos of people, we seek to automatically find spatio-temporal patterns that are distinctive of each set. Existing methods use discriminative approaches and perform post-hoc explainability analysis. Such methods are insufficient as they are unable to provide insights beyond obvious dataset biases, and the explanations are useful only if humans themselves are good at the task. Instead, we tackle the problem through the lens of generative domain translation: our method generates a detailed report of learned, input-dependent spatio-temporal features and the extent to which they vary between the domains. We demonstrate that our method can discover behavioral differences between conversing face-to-face (F2F) and on video-calls (VCs). We also show the applicability of our method on discovering differences in presidential communication styles. Additionally, we are able to predict temporal change-points in videos that decouple expressions in an unsupervised way, and increase the interpretability and usefulness of our model. Finally, our method, being generative, can be used to transform a video call to appear as if it were recorded in a F2F setting. Experiments and visualizations show our approach is able to discover a range of behaviors, taking a step towards deeper understanding of human behaviors. △ Less

Submitted 2 June, 2024; originally announced June 2024.

Comments: Project webpage is available at: https://facet.cs.columbia.edu

Generative Camera Dolly: Extreme Monocular Dynamic Novel View Synthesis

Authors: Basile Van Hoorick, Rundi Wu, Ege Ozguroglu, Kyle Sargent, Ruoshi Liu, Pavel Tokmakov, Achal Dave, Changxi Zheng, Carl Vondrick

Abstract: Accurate reconstruction of complex dynamic scenes from just a single viewpoint continues to be a challenging task in computer vision. Current dynamic novel view synthesis methods typically require videos from many different camera viewpoints, necessitating careful recording setups, and significantly restricting their utility in the wild as well as in terms of embodied AI applications. In this pape… ▽ More Accurate reconstruction of complex dynamic scenes from just a single viewpoint continues to be a challenging task in computer vision. Current dynamic novel view synthesis methods typically require videos from many different camera viewpoints, necessitating careful recording setups, and significantly restricting their utility in the wild as well as in terms of embodied AI applications. In this paper, we propose $\textbf{GCD}$, a controllable monocular dynamic view synthesis pipeline that leverages large-scale diffusion priors to, given a video of any scene, generate a synchronous video from any other chosen perspective, conditioned on a set of relative camera pose parameters. Our model does not require depth as input, and does not explicitly model 3D scene geometry, instead performing end-to-end video-to-video translation in order to achieve its goal efficiently. Despite being trained on synthetic multi-view video data only, zero-shot real-world generalization experiments show promising results in multiple domains, including robotics, object permanence, and driving environments. We believe our framework can potentially unlock powerful applications in rich dynamic scene understanding, perception for robotics, and interactive 3D video viewing experiences for virtual reality. △ Less

Submitted 5 July, 2024; v1 submitted 23 May, 2024; originally announced May 2024.

Comments: Accepted to ECCV 2024. Project webpage is available at: https://gcd.cs.columbia.edu/

Evolving Interpretable Visual Classifiers with Large Language Models

Authors: Mia Chiquier, Utkarsh Mall, Carl Vondrick

Abstract: Multimodal pre-trained models, such as CLIP, are popular for zero-shot classification due to their open-vocabulary flexibility and high performance. However, vision-language models, which compute similarity scores between images and class labels, are largely black-box, with limited interpretability, risk for bias, and inability to discover new visual concepts not written down. Moreover, in practic… ▽ More Multimodal pre-trained models, such as CLIP, are popular for zero-shot classification due to their open-vocabulary flexibility and high performance. However, vision-language models, which compute similarity scores between images and class labels, are largely black-box, with limited interpretability, risk for bias, and inability to discover new visual concepts not written down. Moreover, in practical settings, the vocabulary for class names and attributes of specialized concepts will not be known, preventing these methods from performing well on images uncommon in large-scale vision-language datasets. To address these limitations, we present a novel method that discovers interpretable yet discriminative sets of attributes for visual recognition. We introduce an evolutionary search algorithm that uses a large language model and its in-context learning abilities to iteratively mutate a concept bottleneck of attributes for classification. Our method produces state-of-the-art, interpretable fine-grained classifiers. We outperform the latest baselines by 18.4% on five fine-grained iNaturalist datasets and by 22.2% on two KikiBouba datasets, despite the baselines having access to privileged information about class names. △ Less

Submitted 15 April, 2024; originally announced April 2024.

SelfIE: Self-Interpretation of Large Language Model Embeddings

Authors: Haozhe Chen, Carl Vondrick, Chengzhi Mao

Abstract: How do large language models (LLMs) obtain their answers? The ability to explain and control an LLM's reasoning process is key for reliability, transparency, and future model developments. We propose SelfIE (Self-Interpretation of Embeddings), a framework that enables LLMs to interpret their own embeddings in natural language by leveraging their ability to respond to inquiries about a given passag… ▽ More How do large language models (LLMs) obtain their answers? The ability to explain and control an LLM's reasoning process is key for reliability, transparency, and future model developments. We propose SelfIE (Self-Interpretation of Embeddings), a framework that enables LLMs to interpret their own embeddings in natural language by leveraging their ability to respond to inquiries about a given passage. Capable of interpreting open-world concepts in the hidden embeddings, SelfIE reveals LLM internal reasoning in cases such as making ethical decisions, internalizing prompt injection, and recalling harmful knowledge. SelfIE's text descriptions on hidden embeddings also open up new avenues to control LLM reasoning. We propose Supervised Control, which allows editing open-ended concepts while only requiring gradient computation of individual layer. We extend RLHF to hidden embeddings and propose Reinforcement Control that erases harmful knowledge in LLM without supervision targets. △ Less

Submitted 25 March, 2024; v1 submitted 16 March, 2024; originally announced March 2024.

PaperBot: Learning to Design Real-World Tools Using Paper

Authors: Ruoshi Liu, Junbang Liang, Sruthi Sudhakar, Huy Ha, Cheng Chi, Shuran Song, Carl Vondrick

Abstract: Paper is a cheap, recyclable, and clean material that is often used to make practical tools. Traditional tool design either relies on simulation or physical analysis, which is often inaccurate and time-consuming. In this paper, we propose PaperBot, an approach that directly learns to design and use a tool in the real world using paper without human intervention. We demonstrated the effectiveness a… ▽ More Paper is a cheap, recyclable, and clean material that is often used to make practical tools. Traditional tool design either relies on simulation or physical analysis, which is often inaccurate and time-consuming. In this paper, we propose PaperBot, an approach that directly learns to design and use a tool in the real world using paper without human intervention. We demonstrated the effectiveness and efficiency of PaperBot on two tool design tasks: 1. learning to fold and throw paper airplanes for maximum travel distance 2. learning to cut paper into grippers that exert maximum gripping force. We present a self-supervised learning framework that learns to perform a sequence of folding, cutting, and dynamic manipulation actions in order to optimize the design and use of a tool. We deploy our system to a real-world two-arm robotic system to solve challenging design tasks that involve aerodynamics (paper airplane) and friction (paper gripper) that are impossible to simulate accurately. △ Less

Submitted 14 March, 2024; originally announced March 2024.

Comments: Project Website: https://paperbot.cs.columbia.edu/

GES: Generalized Exponential Splatting for Efficient Radiance Field Rendering

Authors: Abdullah Hamdi, Luke Melas-Kyriazi, Jinjie Mai, Guocheng Qian, Ruoshi Liu, Carl Vondrick, Bernard Ghanem, Andrea Vedaldi

Abstract: Advancements in 3D Gaussian Splatting have significantly accelerated 3D reconstruction and generation. However, it may require a large number of Gaussians, which creates a substantial memory footprint. This paper introduces GES (Generalized Exponential Splatting), a novel representation that employs Generalized Exponential Function (GEF) to model 3D scenes, requiring far fewer particles to represe… ▽ More Advancements in 3D Gaussian Splatting have significantly accelerated 3D reconstruction and generation. However, it may require a large number of Gaussians, which creates a substantial memory footprint. This paper introduces GES (Generalized Exponential Splatting), a novel representation that employs Generalized Exponential Function (GEF) to model 3D scenes, requiring far fewer particles to represent a scene and thus significantly outperforming Gaussian Splatting methods in efficiency with a plug-and-play replacement ability for Gaussian-based utilities. GES is validated theoretically and empirically in both principled 1D setup and realistic 3D scenes. It is shown to represent signals with sharp edges more accurately, which are typically challenging for Gaussians due to their inherent low-pass characteristics. Our empirical analysis demonstrates that GEF outperforms Gaussians in fitting natural-occurring signals (e.g. squares, triangles, and parabolic signals), thereby reducing the need for extensive splitting operations that increase the memory footprint of Gaussian Splatting. With the aid of a frequency-modulated loss, GES achieves competitive performance in novel-view synthesis benchmarks while requiring less than half the memory storage of Gaussian Splatting and increasing the rendering speed by up to 39%. The code is available on the project website https://abdullahamdi.com/ges . △ Less

Submitted 24 May, 2024; v1 submitted 15 February, 2024; originally announced February 2024.

Comments: CVPR 2024 paper. project website https://abdullahamdi.com/ges

pix2gestalt: Amodal Segmentation by Synthesizing Wholes

Authors: Ege Ozguroglu, Ruoshi Liu, Dídac Surís, Dian Chen, Achal Dave, Pavel Tokmakov, Carl Vondrick

Abstract: We introduce pix2gestalt, a framework for zero-shot amodal segmentation, which learns to estimate the shape and appearance of whole objects that are only partially visible behind occlusions. By capitalizing on large-scale diffusion models and transferring their representations to this task, we learn a conditional diffusion model for reconstructing whole objects in challenging zero-shot cases, incl… ▽ More We introduce pix2gestalt, a framework for zero-shot amodal segmentation, which learns to estimate the shape and appearance of whole objects that are only partially visible behind occlusions. By capitalizing on large-scale diffusion models and transferring their representations to this task, we learn a conditional diffusion model for reconstructing whole objects in challenging zero-shot cases, including examples that break natural and physical priors, such as art. As training data, we use a synthetically curated dataset containing occluded objects paired with their whole counterparts. Experiments show that our approach outperforms supervised baselines on established benchmarks. Our model can furthermore be used to significantly improve the performance of existing object recognition and 3D reconstruction methods in the presence of occlusions. △ Less

Submitted 25 January, 2024; originally announced January 2024.

Comments: Website: https://gestalt.cs.columbia.edu/

Raidar: geneRative AI Detection viA Rewriting

Authors: Chengzhi Mao, Carl Vondrick, Hao Wang, Junfeng Yang

Abstract: We find that large language models (LLMs) are more likely to modify human-written text than AI-generated text when tasked with rewriting. This tendency arises because LLMs often perceive AI-generated text as high-quality, leading to fewer modifications. We introduce a method to detect AI-generated content by prompting LLMs to rewrite text and calculating the editing distance of the output. We dubb… ▽ More We find that large language models (LLMs) are more likely to modify human-written text than AI-generated text when tasked with rewriting. This tendency arises because LLMs often perceive AI-generated text as high-quality, leading to fewer modifications. We introduce a method to detect AI-generated content by prompting LLMs to rewrite text and calculating the editing distance of the output. We dubbed our geneRative AI Detection viA Rewriting method Raidar. Raidar significantly improves the F1 detection scores of existing AI content detection models -- both academic and commercial -- across various domains, including News, creative writing, student essays, code, Yelp reviews, and arXiv papers, with gains of up to 29 points. Operating solely on word symbols without high-dimensional features, our method is compatible with black box LLMs, and is inherently robust on new content. Our results illustrate the unique imprint of machine-generated text through the lens of the machines themselves. △ Less

Submitted 14 April, 2024; v1 submitted 23 January, 2024; originally announced January 2024.

Comments: Accepted by ICLR 2024, Large Language Models, Detection

Remote Sensing Vision-Language Foundation Models without Annotations via Ground Remote Alignment

Authors: Utkarsh Mall, Cheng Perng Phoo, Meilin Kelsey Liu, Carl Vondrick, Bharath Hariharan, Kavita Bala

Abstract: We introduce a method to train vision-language models for remote-sensing images without using any textual annotations. Our key insight is to use co-located internet imagery taken on the ground as an intermediary for connecting remote-sensing images and language. Specifically, we train an image encoder for remote sensing images to align with the image encoder of CLIP using a large amount of paired… ▽ More We introduce a method to train vision-language models for remote-sensing images without using any textual annotations. Our key insight is to use co-located internet imagery taken on the ground as an intermediary for connecting remote-sensing images and language. Specifically, we train an image encoder for remote sensing images to align with the image encoder of CLIP using a large amount of paired internet and satellite images. Our unsupervised approach enables the training of a first-of-its-kind large-scale vision language model (VLM) for remote sensing images at two different resolutions. We show that these VLMs enable zero-shot, open-vocabulary image classification, retrieval, segmentation and visual question answering for satellite images. On each of these tasks, our VLM trained without textual annotations outperforms existing VLMs trained with supervision, with gains of up to 20% for classification and 80% for segmentation. △ Less

Submitted 11 December, 2023; originally announced December 2023.

Interpreting and Controlling Vision Foundation Models via Text Explanations

Authors: Haozhe Chen, Junfeng Yang, Carl Vondrick, Chengzhi Mao

Abstract: Large-scale pre-trained vision foundation models, such as CLIP, have become de facto backbones for various vision tasks. However, due to their black-box nature, understanding the underlying rules behind these models' predictions and controlling model behaviors have remained open challenges. We present a framework for interpreting vision transformer's latent tokens with natural language. Given a la… ▽ More Large-scale pre-trained vision foundation models, such as CLIP, have become de facto backbones for various vision tasks. However, due to their black-box nature, understanding the underlying rules behind these models' predictions and controlling model behaviors have remained open challenges. We present a framework for interpreting vision transformer's latent tokens with natural language. Given a latent token, our framework retains its semantic information to the final layer using transformer's local operations and retrieves the closest text for explanation. Our approach enables understanding of model visual reasoning procedure without needing additional model training or data collection. Based on the obtained interpretations, our framework allows for model editing that controls model reasoning behaviors and improves model robustness against biases and spurious correlations. △ Less

Submitted 16 October, 2023; originally announced October 2023.

SHIFT3D: Synthesizing Hard Inputs For Tricking 3D Detectors

Authors: Hongge Chen, Zhao Chen, Gregory P. Meyer, Dennis Park, Carl Vondrick, Ashish Shrivastava, Yuning Chai

Abstract: We present SHIFT3D, a differentiable pipeline for generating 3D shapes that are structurally plausible yet challenging to 3D object detectors. In safety-critical applications like autonomous driving, discovering such novel challenging objects can offer insight into unknown vulnerabilities of 3D detectors. By representing objects with a signed distanced function (SDF), we show that gradient error s… ▽ More We present SHIFT3D, a differentiable pipeline for generating 3D shapes that are structurally plausible yet challenging to 3D object detectors. In safety-critical applications like autonomous driving, discovering such novel challenging objects can offer insight into unknown vulnerabilities of 3D detectors. By representing objects with a signed distanced function (SDF), we show that gradient error signals allow us to smoothly deform the shape or pose of a 3D object in order to confuse a downstream 3D detector. Importantly, the objects generated by SHIFT3D physically differ from the baseline object yet retain a semantically recognizable shape. Our approach provides interpretable failure modes for modern 3D object detectors, and can aid in preemptive discovery of potential safety risks within 3D perception systems before these risks become critical failures. △ Less

Submitted 11 September, 2023; originally announced September 2023.

Comments: Accepted by ICCV 2023

Objaverse-XL: A Universe of 10M+ 3D Objects

Authors: Matt Deitke, Ruoshi Liu, Matthew Wallingford, Huong Ngo, Oscar Michel, Aditya Kusupati, Alan Fan, Christian Laforte, Vikram Voleti, Samir Yitzhak Gadre, Eli VanderBilt, Aniruddha Kembhavi, Carl Vondrick, Georgia Gkioxari, Kiana Ehsani, Ludwig Schmidt, Ali Farhadi

Abstract: Natural language processing and 2D vision models have attained remarkable proficiency on many tasks primarily by escalating the scale of training data. However, 3D vision tasks have not seen the same progress, in part due to the challenges of acquiring high-quality 3D data. In this work, we present Objaverse-XL, a dataset of over 10 million 3D objects. Our dataset comprises deduplicated 3D objects… ▽ More Natural language processing and 2D vision models have attained remarkable proficiency on many tasks primarily by escalating the scale of training data. However, 3D vision tasks have not seen the same progress, in part due to the challenges of acquiring high-quality 3D data. In this work, we present Objaverse-XL, a dataset of over 10 million 3D objects. Our dataset comprises deduplicated 3D 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. Representing the largest scale and diversity in the realm of 3D datasets, Objaverse-XL enables significant new possibilities for 3D vision. Our experiments demonstrate the improvements enabled with the scale provided by Objaverse-XL. We show that by training Zero123 on novel view synthesis, utilizing over 100 million multi-view rendered images, we achieve strong zero-shot generalization abilities. We hope that releasing Objaverse-XL will enable further innovations in the field of 3D vision at scale. △ Less

Submitted 11 July, 2023; originally announced July 2023.

Sin3DM: Learning a Diffusion Model from a Single 3D Textured Shape

Authors: Rundi Wu, Ruoshi Liu, Carl Vondrick, Changxi Zheng

Abstract: Synthesizing novel 3D models that resemble the input example has long been pursued by graphics artists and machine learning researchers. In this paper, we present Sin3DM, a diffusion model that learns the internal patch distribution from a single 3D textured shape and generates high-quality variations with fine geometry and texture details. Training a diffusion model directly in 3D would induce la… ▽ More Synthesizing novel 3D models that resemble the input example has long been pursued by graphics artists and machine learning researchers. In this paper, we present Sin3DM, a diffusion model that learns the internal patch distribution from a single 3D textured shape and generates high-quality variations with fine geometry and texture details. Training a diffusion model directly in 3D would induce large memory and computational cost. Therefore, we first compress the input into a lower-dimensional latent space and then train a diffusion model on it. Specifically, we encode the input 3D textured shape into triplane feature maps that represent the signed distance and texture fields of the input. The denoising network of our diffusion model has a limited receptive field to avoid overfitting, and uses triplane-aware 2D convolution blocks to improve the result quality. Aside from randomly generating new samples, our model also facilitates applications such as retargeting, outpainting and local editing. Through extensive qualitative and quantitative evaluation, we show that our method outperforms prior methods in generation quality of 3D shapes. △ Less

Submitted 20 February, 2024; v1 submitted 24 May, 2023; originally announced May 2023.

Comments: Accepted to ICLR 2024. Project page: https://Sin3DM.github.io, Code: https://github.com/Sin3DM/Sin3DM

Tracking through Containers and Occluders in the Wild

Authors: Basile Van Hoorick, Pavel Tokmakov, Simon Stent, Jie Li, Carl Vondrick

Abstract: Tracking objects with persistence in cluttered and dynamic environments remains a difficult challenge for computer vision systems. In this paper, we introduce $\textbf{TCOW}$, a new benchmark and model for visual tracking through heavy occlusion and containment. We set up a task where the goal is to, given a video sequence, segment both the projected extent of the target object, as well as the sur… ▽ More Tracking objects with persistence in cluttered and dynamic environments remains a difficult challenge for computer vision systems. In this paper, we introduce $\textbf{TCOW}$, a new benchmark and model for visual tracking through heavy occlusion and containment. We set up a task where the goal is to, given a video sequence, segment both the projected extent of the target object, as well as the surrounding container or occluder whenever one exists. To study this task, we create a mixture of synthetic and annotated real datasets to support both supervised learning and structured evaluation of model performance under various forms of task variation, such as moving or nested containment. We evaluate two recent transformer-based video models and find that while they can be surprisingly capable of tracking targets under certain settings of task variation, there remains a considerable performance gap before we can claim a tracking model to have acquired a true notion of object permanence. △ Less

Submitted 4 May, 2023; originally announced May 2023.

Comments: Accepted at CVPR 2023. Project webpage is available at: https://tcow.cs.columbia.edu/

Humans as Light Bulbs: 3D Human Reconstruction from Thermal Reflection

Authors: Ruoshi Liu, Carl Vondrick

Abstract: The relatively hot temperature of the human body causes people to turn into long-wave infrared light sources. Since this emitted light has a larger wavelength than visible light, many surfaces in typical scenes act as infrared mirrors with strong specular reflections. We exploit the thermal reflections of a person onto objects in order to locate their position and reconstruct their pose, even if t… ▽ More The relatively hot temperature of the human body causes people to turn into long-wave infrared light sources. Since this emitted light has a larger wavelength than visible light, many surfaces in typical scenes act as infrared mirrors with strong specular reflections. We exploit the thermal reflections of a person onto objects in order to locate their position and reconstruct their pose, even if they are not visible to a normal camera. We propose an analysis-by-synthesis framework that jointly models the objects, people, and their thermal reflections, which allows us to combine generative models with differentiable rendering of reflections. Quantitative and qualitative experiments show our approach works in highly challenging cases, such as with curved mirrors or when the person is completely unseen by a normal camera. △ Less

Submitted 2 May, 2023; originally announced May 2023.

Comments: Website: https://thermal.cs.columbia.edu/

SURFSUP: Learning Fluid Simulation for Novel Surfaces

Authors: Arjun Mani, Ishaan Preetam Chandratreya, Elliot Creager, Carl Vondrick, Richard Zemel

Abstract: Modeling the mechanics of fluid in complex scenes is vital to applications in design, graphics, and robotics. Learning-based methods provide fast and differentiable fluid simulators, however most prior work is unable to accurately model how fluids interact with genuinely novel surfaces not seen during training. We introduce SURFSUP, a framework that represents objects implicitly using signed dista… ▽ More Modeling the mechanics of fluid in complex scenes is vital to applications in design, graphics, and robotics. Learning-based methods provide fast and differentiable fluid simulators, however most prior work is unable to accurately model how fluids interact with genuinely novel surfaces not seen during training. We introduce SURFSUP, a framework that represents objects implicitly using signed distance functions (SDFs), rather than an explicit representation of meshes or particles. This continuous representation of geometry enables more accurate simulation of fluid-object interactions over long time periods while simultaneously making computation more efficient. Moreover, SURFSUP trained on simple shape primitives generalizes considerably out-of-distribution, even to complex real-world scenes and objects. Finally, we show we can invert our model to design simple objects to manipulate fluid flow. △ Less

Submitted 8 September, 2023; v1 submitted 12 April, 2023; originally announced April 2023.

Comments: Website: https://surfsup.cs.columbia.edu/

Zero-1-to-3: Zero-shot One Image to 3D Object

Authors: Ruoshi Liu, Rundi Wu, Basile Van Hoorick, Pavel Tokmakov, Sergey Zakharov, Carl Vondrick

Abstract: We introduce Zero-1-to-3, a framework for changing the camera viewpoint of an object given just a single RGB image. To perform novel view synthesis in this under-constrained setting, we capitalize on the geometric priors that large-scale diffusion models learn about natural images. Our conditional diffusion model uses a synthetic dataset to learn controls of the relative camera viewpoint, which al… ▽ More We introduce Zero-1-to-3, a framework for changing the camera viewpoint of an object given just a single RGB image. To perform novel view synthesis in this under-constrained setting, we capitalize on the geometric priors that large-scale diffusion models learn about natural images. Our conditional diffusion model uses a synthetic dataset to learn controls of the relative camera viewpoint, which allow new images to be generated of the same object under a specified camera transformation. Even though it is trained on a synthetic dataset, our model retains a strong zero-shot generalization ability to out-of-distribution datasets as well as in-the-wild images, including impressionist paintings. Our viewpoint-conditioned diffusion approach can further be used for the task of 3D reconstruction from a single image. Qualitative and quantitative experiments show that our method significantly outperforms state-of-the-art single-view 3D reconstruction and novel view synthesis models by leveraging Internet-scale pre-training. △ Less

Submitted 20 March, 2023; originally announced March 2023.

Comments: Website: https://zero123.cs.columbia.edu/

ViperGPT: Visual Inference via Python Execution for Reasoning

Authors: Dídac Surís, Sachit Menon, Carl Vondrick

Abstract: Answering visual queries is a complex task that requires both visual processing and reasoning. End-to-end models, the dominant approach for this task, do not explicitly differentiate between the two, limiting interpretability and generalization. Learning modular programs presents a promising alternative, but has proven challenging due to the difficulty of learning both the programs and modules sim… ▽ More Answering visual queries is a complex task that requires both visual processing and reasoning. End-to-end models, the dominant approach for this task, do not explicitly differentiate between the two, limiting interpretability and generalization. Learning modular programs presents a promising alternative, but has proven challenging due to the difficulty of learning both the programs and modules simultaneously. We introduce ViperGPT, a framework that leverages code-generation models to compose vision-and-language models into subroutines to produce a result for any query. ViperGPT utilizes a provided API to access the available modules, and composes them by generating Python code that is later executed. This simple approach requires no further training, and achieves state-of-the-art results across various complex visual tasks. △ Less

Submitted 14 March, 2023; originally announced March 2023.

Comments: Website: https://viper.cs.columbia.edu/

Affective Faces for Goal-Driven Dyadic Communication

Authors: Scott Geng, Revant Teotia, Purva Tendulkar, Sachit Menon, Carl Vondrick

Abstract: We introduce a video framework for modeling the association between verbal and non-verbal communication during dyadic conversation. Given the input speech of a speaker, our approach retrieves a video of a listener, who has facial expressions that would be socially appropriate given the context. Our approach further allows the listener to be conditioned on their own goals, personalities, or backgro… ▽ More We introduce a video framework for modeling the association between verbal and non-verbal communication during dyadic conversation. Given the input speech of a speaker, our approach retrieves a video of a listener, who has facial expressions that would be socially appropriate given the context. Our approach further allows the listener to be conditioned on their own goals, personalities, or backgrounds. Our approach models conversations through a composition of large language models and vision-language models, creating internal representations that are interpretable and controllable. To study multimodal communication, we propose a new video dataset of unscripted conversations covering diverse topics and demographics. Experiments and visualizations show our approach is able to output listeners that are significantly more socially appropriate than baselines. However, many challenges remain, and we release our dataset publicly to spur further progress. See our website for video results, data, and code: https://realtalk.cs.columbia.edu. △ Less

Submitted 26 January, 2023; originally announced January 2023.

Adversarially Robust Video Perception by Seeing Motion

Authors: Lingyu Zhang, Chengzhi Mao, Junfeng Yang, Carl Vondrick

Abstract: Despite their excellent performance, state-of-the-art computer vision models often fail when they encounter adversarial examples. Video perception models tend to be more fragile under attacks, because the adversary has more places to manipulate in high-dimensional data. In this paper, we find one reason for video models' vulnerability is that they fail to perceive the correct motion under adversar… ▽ More Despite their excellent performance, state-of-the-art computer vision models often fail when they encounter adversarial examples. Video perception models tend to be more fragile under attacks, because the adversary has more places to manipulate in high-dimensional data. In this paper, we find one reason for video models' vulnerability is that they fail to perceive the correct motion under adversarial perturbations. Inspired by the extensive evidence that motion is a key factor for the human visual system, we propose to correct what the model sees by restoring the perceived motion information. Since motion information is an intrinsic structure of the video data, recovering motion signals can be done at inference time without any human annotation, which allows the model to adapt to unforeseen, worst-case inputs. Visualizations and empirical experiments on UCF-101 and HMDB-51 datasets show that restoring motion information in deep vision models improves adversarial robustness. Even under adaptive attacks where the adversary knows our defense, our algorithm is still effective. Our work provides new insight into robust video perception algorithms by using intrinsic structures from the data. Our webpage is available at https://motion4robust.cs.columbia.edu. △ Less

Submitted 12 December, 2022; originally announced December 2022.

Understanding Zero-Shot Adversarial Robustness for Large-Scale Models

Authors: Chengzhi Mao, Scott Geng, Junfeng Yang, Xin Wang, Carl Vondrick

Abstract: Pretrained large-scale vision-language models like CLIP have exhibited strong generalization over unseen tasks. Yet imperceptible adversarial perturbations can significantly reduce CLIP's performance on new tasks. In this work, we identify and explore the problem of \emph{adapting large-scale models for zero-shot adversarial robustness}. We first identify two key factors during model adaption -- t… ▽ More Pretrained large-scale vision-language models like CLIP have exhibited strong generalization over unseen tasks. Yet imperceptible adversarial perturbations can significantly reduce CLIP's performance on new tasks. In this work, we identify and explore the problem of \emph{adapting large-scale models for zero-shot adversarial robustness}. We first identify two key factors during model adaption -- training losses and adaptation methods -- that affect the model's zero-shot adversarial robustness. We then propose a text-guided contrastive adversarial training loss, which aligns the text embeddings and the adversarial visual features with contrastive learning on a small set of training data. We apply this training loss to two adaption methods, model finetuning and visual prompt tuning. We find that visual prompt tuning is more effective in the absence of texts, while finetuning wins in the existence of text guidance. Overall, our approach significantly improves the zero-shot adversarial robustness over CLIP, seeing an average improvement of over 31 points over ImageNet and 15 zero-shot datasets. We hope this work can shed light on understanding the zero-shot adversarial robustness of large-scale models. △ Less

Submitted 21 April, 2023; v1 submitted 13 December, 2022; originally announced December 2022.

Doubly Right Object Recognition: A Why Prompt for Visual Rationales

Authors: Chengzhi Mao, Revant Teotia, Amrutha Sundar, Sachit Menon, Junfeng Yang, Xin Wang, Carl Vondrick

Abstract: Many visual recognition models are evaluated only on their classification accuracy, a metric for which they obtain strong performance. In this paper, we investigate whether computer vision models can also provide correct rationales for their predictions. We propose a ``doubly right'' object recognition benchmark, where the metric requires the model to simultaneously produce both the right labels a… ▽ More Many visual recognition models are evaluated only on their classification accuracy, a metric for which they obtain strong performance. In this paper, we investigate whether computer vision models can also provide correct rationales for their predictions. We propose a ``doubly right'' object recognition benchmark, where the metric requires the model to simultaneously produce both the right labels as well as the right rationales. We find that state-of-the-art visual models, such as CLIP, often provide incorrect rationales for their categorical predictions. However, by transferring the rationales from language models into visual representations through a tailored dataset, we show that we can learn a ``why prompt,'' which adapts large visual representations to produce correct rationales. Visualizations and empirical experiments show that our prompts significantly improve performance on doubly right object recognition, in addition to zero-shot transfer to unseen tasks and datasets. △ Less

Submitted 22 March, 2023; v1 submitted 12 December, 2022; originally announced December 2022.

Comments: Accepted at CVPR 2023

Robust Perception through Equivariance

Authors: Chengzhi Mao, Lingyu Zhang, Abhishek Joshi, Junfeng Yang, Hao Wang, Carl Vondrick

Abstract: Deep networks for computer vision are not reliable when they encounter adversarial examples. In this paper, we introduce a framework that uses the dense intrinsic constraints in natural images to robustify inference. By introducing constraints at inference time, we can shift the burden of robustness from training to the inference algorithm, thereby allowing the model to adjust dynamically to each… ▽ More Deep networks for computer vision are not reliable when they encounter adversarial examples. In this paper, we introduce a framework that uses the dense intrinsic constraints in natural images to robustify inference. By introducing constraints at inference time, we can shift the burden of robustness from training to the inference algorithm, thereby allowing the model to adjust dynamically to each individual image's unique and potentially novel characteristics at inference time. Among different constraints, we find that equivariance-based constraints are most effective, because they allow dense constraints in the feature space without overly constraining the representation at a fine-grained level. Our theoretical results validate the importance of having such dense constraints at inference time. Our empirical experiments show that restoring feature equivariance at inference time defends against worst-case adversarial perturbations. The method obtains improved adversarial robustness on four datasets (ImageNet, Cityscapes, PASCAL VOC, and MS-COCO) on image recognition, semantic segmentation, and instance segmentation tasks. Project page is available at equi4robust.cs.columbia.edu. △ Less

Submitted 3 June, 2023; v1 submitted 12 December, 2022; originally announced December 2022.

Comments: Published in ICML 2023

Task Bias in Vision-Language Models

Authors: Sachit Menon, Ishaan Preetam Chandratreya, Carl Vondrick

Abstract: Incidental supervision from language has become a popular approach for learning generic visual representations that can be prompted to perform many recognition tasks in computer vision. We conduct an in-depth exploration of the CLIP model and show that its visual representation is often strongly biased towards solving some tasks more than others. Moreover, which task the representation will be bia… ▽ More Incidental supervision from language has become a popular approach for learning generic visual representations that can be prompted to perform many recognition tasks in computer vision. We conduct an in-depth exploration of the CLIP model and show that its visual representation is often strongly biased towards solving some tasks more than others. Moreover, which task the representation will be biased towards is unpredictable, with little consistency across images. To resolve this task bias, we show how to learn a visual prompt that guides the representation towards features relevant to their task of interest. Our results show that these visual prompts can be independent of the input image and still effectively provide a conditioning mechanism to steer visual representations towards the desired task. △ Less

Submitted 8 December, 2022; originally announced December 2022.

Comments: First two authors contributed equally

Muscles in Action

Authors: Mia Chiquier, Carl Vondrick

Abstract: Human motion is created by, and constrained by, our muscles. We take a first step at building computer vision methods that represent the internal muscle activity that causes motion. We present a new dataset, Muscles in Action (MIA), to learn to incorporate muscle activity into human motion representations. The dataset consists of 12.5 hours of synchronized video and surface electromyography (sEMG)… ▽ More Human motion is created by, and constrained by, our muscles. We take a first step at building computer vision methods that represent the internal muscle activity that causes motion. We present a new dataset, Muscles in Action (MIA), to learn to incorporate muscle activity into human motion representations. The dataset consists of 12.5 hours of synchronized video and surface electromyography (sEMG) data of 10 subjects performing various exercises. Using this dataset, we learn a bidirectional representation that predicts muscle activation from video, and conversely, reconstructs motion from muscle activation. We evaluate our model on in-distribution subjects and exercises, as well as on out-of-distribution subjects and exercises. We demonstrate how advances in modeling both modalities jointly can serve as conditioning for muscularly consistent motion generation. Putting muscles into computer vision systems will enable richer models of virtual humans, with applications in sports, fitness, and AR/VR. △ Less

Submitted 20 March, 2023; v1 submitted 5 December, 2022; originally announced December 2022.

Private Multiparty Perception for Navigation

Authors: Hui Lu, Mia Chiquier, Carl Vondrick

Abstract: We introduce a framework for navigating through cluttered environments by connecting multiple cameras together while simultaneously preserving privacy. Occlusions and obstacles in large environments are often challenging situations for navigation agents because the environment is not fully observable from a single camera view. Given multiple camera views of an environment, our approach learns to p… ▽ More We introduce a framework for navigating through cluttered environments by connecting multiple cameras together while simultaneously preserving privacy. Occlusions and obstacles in large environments are often challenging situations for navigation agents because the environment is not fully observable from a single camera view. Given multiple camera views of an environment, our approach learns to produce a multiview scene representation that can only be used for navigation, provably preventing one party from inferring anything beyond the output task. On a new navigation dataset that we will publicly release, experiments show that private multiparty representations allow navigation through complex scenes and around obstacles while jointly preserving privacy. Our approach scales to an arbitrary number of camera viewpoints. We believe developing visual representations that preserve privacy is increasingly important for many applications such as navigation. △ Less

Submitted 1 December, 2022; originally announced December 2022.

FLEX: Full-Body Grasping Without Full-Body Grasps

Authors: Purva Tendulkar, Dídac Surís, Carl Vondrick

Abstract: Synthesizing 3D human avatars interacting realistically with a scene is an important problem with applications in AR/VR, video games and robotics. Towards this goal, we address the task of generating a virtual human -- hands and full body -- grasping everyday objects. Existing methods approach this problem by collecting a 3D dataset of humans interacting with objects and training on this data. How… ▽ More Synthesizing 3D human avatars interacting realistically with a scene is an important problem with applications in AR/VR, video games and robotics. Towards this goal, we address the task of generating a virtual human -- hands and full body -- grasping everyday objects. Existing methods approach this problem by collecting a 3D dataset of humans interacting with objects and training on this data. However, 1) these methods do not generalize to different object positions and orientations, or to the presence of furniture in the scene, and 2) the diversity of their generated full-body poses is very limited. In this work, we address all the above challenges to generate realistic, diverse full-body grasps in everyday scenes without requiring any 3D full-body grasping data. Our key insight is to leverage the existence of both full-body pose and hand grasping priors, composing them using 3D geometrical constraints to obtain full-body grasps. We empirically validate that these constraints can generate a variety of feasible human grasps that are superior to baselines both quantitatively and qualitatively. See our webpage for more details: https://flex.cs.columbia.edu/. △ Less

Submitted 28 March, 2023; v1 submitted 21 November, 2022; originally announced November 2022.

Comments: CVPR 2023 Camera-ready

Visual Classification via Description from Large Language Models

Authors: Sachit Menon, Carl Vondrick

Abstract: Vision-language models (VLMs) such as CLIP have shown promising performance on a variety of recognition tasks using the standard zero-shot classification procedure -- computing similarity between the query image and the embedded words for each category. By only using the category name, they neglect to make use of the rich context of additional information that language affords. The procedure gives… ▽ More Vision-language models (VLMs) such as CLIP have shown promising performance on a variety of recognition tasks using the standard zero-shot classification procedure -- computing similarity between the query image and the embedded words for each category. By only using the category name, they neglect to make use of the rich context of additional information that language affords. The procedure gives no intermediate understanding of why a category is chosen, and furthermore provides no mechanism for adjusting the criteria used towards this decision. We present an alternative framework for classification with VLMs, which we call classification by description. We ask VLMs to check for descriptive features rather than broad categories: to find a tiger, look for its stripes; its claws; and more. By basing decisions on these descriptors, we can provide additional cues that encourage using the features we want to be used. In the process, we can get a clear idea of what features the model uses to construct its decision; it gains some level of inherent explainability. We query large language models (e.g., GPT-3) for these descriptors to obtain them in a scalable way. Extensive experiments show our framework has numerous advantages past interpretability. We show improvements in accuracy on ImageNet across distribution shifts; demonstrate the ability to adapt VLMs to recognize concepts unseen during training; and illustrate how descriptors can be edited to effectively mitigate bias compared to the baseline. △ Less

Submitted 1 December, 2022; v1 submitted 13 October, 2022; originally announced October 2022.

Representing Spatial Trajectories as Distributions

Authors: Dídac Surís, Carl Vondrick

Abstract: We introduce a representation learning framework for spatial trajectories. We represent partial observations of trajectories as probability distributions in a learned latent space, which characterize the uncertainty about unobserved parts of the trajectory. Our framework allows us to obtain samples from a trajectory for any continuous point in time, both interpolating and extrapolating. Our flexib… ▽ More We introduce a representation learning framework for spatial trajectories. We represent partial observations of trajectories as probability distributions in a learned latent space, which characterize the uncertainty about unobserved parts of the trajectory. Our framework allows us to obtain samples from a trajectory for any continuous point in time, both interpolating and extrapolating. Our flexible approach supports directly modifying specific attributes of a trajectory, such as its pace, as well as combining different partial observations into single representations. Experiments show our method's advantage over baselines in prediction tasks. △ Less

Submitted 3 October, 2022; originally announced October 2022.

Comments: Accepted to NeurIPS 2022