David Fouhey

Ahmad Darkhalil, Dandan Shan, Bin Zhu et al.

We introduce VISOR, a new dataset of pixel annotations and a benchmark suite for segmenting hands and active objects in egocentric video. VISOR annotates videos from EPIC-KITCHENS, which comes with a new set of challenges not encountered in current video segmentation datasets. Specifically, we need to ensure both short- and long-term consistency of pixel-level annotations as objects undergo transformative interactions, e.g. an onion is peeled, diced and cooked - where we aim to obtain accurate pixel-level annotations of the peel, onion pieces, chopping board, knife, pan, as well as the acting hands. VISOR introduces an annotation pipeline, AI-powered in parts, for scalability and quality. In total, we publicly release 272K manual semantic masks of 257 object classes, 9.9M interpolated dense masks, 67K hand-object relations, covering 36 hours of 179 untrimmed videos. Along with the annotations, we introduce three challenges in video object segmentation, interaction understanding and long-term reasoning. For data, code and leaderboards: http://epic-kitchens.github.io/VISOR

Vadim Tschernezki, Ahmad Darkhalil, Zhifan Zhu et al.

Neural rendering is fuelling a unification of learning, 3D geometry and video understanding that has been waiting for more than two decades. Progress, however, is still hampered by a lack of suitable datasets and benchmarks. To address this gap, we introduce EPIC Fields, an augmentation of EPIC-KITCHENS with 3D camera information. Like other datasets for neural rendering, EPIC Fields removes the complex and expensive step of reconstructing cameras using photogrammetry, and allows researchers to focus on modelling problems. We illustrate the challenge of photogrammetry in egocentric videos of dynamic actions and propose innovations to address them. Compared to other neural rendering datasets, EPIC Fields is better tailored to video understanding because it is paired with labelled action segments and the recent VISOR segment annotations. To further motivate the community, we also evaluate two benchmark tasks in neural rendering and segmenting dynamic objects, with strong baselines that showcase what is not possible today. We also highlight the advantage of geometry in semi-supervised video object segmentations on the VISOR annotations. EPIC Fields reconstructs 96% of videos in EPICKITCHENS, registering 19M frames in 99 hours recorded in 45 kitchens.

Nilesh Kulkarni, Davis Rempe, Kyle Genova et al.

We address the problem of generating realistic 3D motions of humans interacting with objects in a scene. Our key idea is to create a neural interaction field attached to a specific object, which outputs the distance to the valid interaction manifold given a human pose as input. This interaction field guides the sampling of an object-conditioned human motion diffusion model, so as to encourage plausible contacts and affordance semantics. To support interactions with scarcely available data, we propose an automated synthetic data pipeline. For this, we seed a pre-trained motion model, which has priors for the basics of human movement, with interaction-specific anchor poses extracted from limited motion capture data. Using our guided diffusion model trained on generated synthetic data, we synthesize realistic motions for sitting and lifting with several objects, outperforming alternative approaches in terms of motion quality and successful action completion. We call our framework NIFTY: Neural Interaction Fields for Trajectory sYnthesis.

Payel Mukhopadhyay, Stefan S. Nixon, Romain Watteaux et al.

Whether physics foundation models can be usefully deployed on laboratory experiments remains an open question for scientific machine learning (ML). We test this question on the Rayleigh-Taylor instability (RTI), a ubiquitous and demanding fluid instability seen from tabletop flows to supernova explosions, in which small perturbations at a density interface grow into chaotic, multiscale mixing as a lighter fluid accelerates into a heavier one. Standard ML models struggle with RTI, and despite over a century of theoretical, numerical, and experimental work, it carries an unresolved discrepancy between simulation and experiment: the late-time mixing growth rate, $α$, measured in most laboratory experiments ($\sim$ 0.06-0.07), is roughly three times the value from idealized direct numerical simulations (DNS, $\sim$ 0.02). The gap's origin remains debated. These properties make RTI a stringent test for a question that matters well beyond RTI: can foundation models trained only on simulations generalise to sparse, messy, and noisy laboratory settings? We finetune Walrus, a foundation model for continuum dynamics, on three or fewer DNS realizations and recover key RTI physics over long rollouts. Applied zero-shot to sliding-barrier laboratory data, the finetuned model leaves the DNS-like regime and enters the observed growth band, having never seen a single experimental sample. These results provide independent, data-driven evidence that initial conditions play a crucial role in the longstanding sim-experiment gap in $α$. The model also generalises zero-shot to stable stratification, a buoyancy regime absent from training, correctly slowing mixing-layer growth. Together, our results show that foundation models can generalise well beyond their training data, predicting laboratory behavior and unseen physical regimes, opening new ways to probe longstanding simulation-experiment gaps.

Georgios Pavlakos, Dandan Shan, Ilija Radosavovic et al.

We present an approach that can reconstruct hands in 3D from monocular input. Our approach for Hand Mesh Recovery, HaMeR, follows a fully transformer-based architecture and can analyze hands with significantly increased accuracy and robustness compared to previous work. The key to HaMeR's success lies in scaling up both the data used for training and the capacity of the deep network for hand reconstruction. For training data, we combine multiple datasets that contain 2D or 3D hand annotations. For the deep model, we use a large scale Vision Transformer architecture. Our final model consistently outperforms the previous baselines on popular 3D hand pose benchmarks. To further evaluate the effect of our design in non-controlled settings, we annotate existing in-the-wild datasets with 2D hand keypoint annotations. On this newly collected dataset of annotations, HInt, we demonstrate significant improvements over existing baselines. We make our code, data and models available on the project website: https://geopavlakos.github.io/hamer/.

Sarah Jabbour, Gregory Kondas, Ella Kazerooni et al.

We propose a permutation-based explanation method for image classifiers. Current image-model explanations like activation maps are limited to instance-based explanations in the pixel space, making it difficult to understand global model behavior. In contrast, permutation based explanations for tabular data classifiers measure feature importance by comparing model performance on data before and after permuting a feature. We propose an explanation method for image-based models that permutes interpretable concepts across dataset images. Given a dataset of images labeled with specific concepts like captions, we permute a concept across examples in the text space and then generate images via a text-conditioned diffusion model. Feature importance is then reflected by the change in model performance relative to unpermuted data. When applied to a set of concepts, the method generates a ranking of feature importance. We show this approach recovers underlying model feature importance on synthetic and real-world image classification tasks.

Linyi Jin, Richard Tucker, Zhengqi Li et al. · deepmind

Learning to understand dynamic 3D scenes from imagery is crucial for applications ranging from robotics to scene reconstruction. Yet, unlike other problems where large-scale supervised training has enabled rapid progress, directly supervising methods for recovering 3D motion remains challenging due to the fundamental difficulty of obtaining ground truth annotations. We present a system for mining high-quality 4D reconstructions from internet stereoscopic, wide-angle videos. Our system fuses and filters the outputs of camera pose estimation, stereo depth estimation, and temporal tracking methods into high-quality dynamic 3D reconstructions. We use this method to generate large-scale data in the form of world-consistent, pseudo-metric 3D point clouds with long-term motion trajectories. We demonstrate the utility of this data by training a variant of DUSt3R to predict structure and 3D motion from real-world image pairs, showing that training on our reconstructed data enables generalization to diverse real-world scenes. Project page and data at: https://stereo4d.github.io

Siavash Golkar, Jake Kovalic, Irina Espejo Morales et al.

Biological function emerges from coupled constraints across sequence, structure, regulation, evolution, and cellular context, yet most foundation models in biology are trained within one modality or for a fixed forward task. We present MIMIC, a generative multimodal foundation model trained on our newly curated and aligned dataset, LORE, linking nucleic acid, protein, evolutionary, structural, regulatory, and semantic/contextual modalities within partially observed biomolecular states. MIMIC uses a split-track encoder-decoder architecture to condition on arbitrary subsets of observed modalities and reconstruct or generate missing components of molecular state across the genome, transcriptome, and proteome. Multimodal conditioning consistently improves MIMIC's sequence reconstruction relative to sequence-only inputs, while its learned representations enable state-of-the-art performance on RNA and protein downstream tasks. MIMIC achieves state-of-the-art splicing prediction, and its joint generative formulation enables isoform-aware inference that further improves performance. Beyond prediction, the same generative framework supports constrained design. For RNA, MIMIC identifies corrective edits in a clinically relevant HBB splice-disrupting mutation without reverting it by using evolutionary and structural signals. For proteins, jointly conditioning on shape and surface chemistry of PD-L1 and hACE2 binding sites produces diverse, high-confidence sequences with strong in silico support for target binding. Finally, MIMIC uses experimental context as semantic conditioning to model assay-dependent RNA chemical probing, rather than treating context as a fixed output. Together, these results position MIMIC's aligned multimodal generative modeling as a strong foundation for unifying representation learning, conditional prediction, and constrained biomolecular design within a single model.

Linyi Jin, Nilesh Kulkarni, David Fouhey · deepmind

This paper introduces 3DFIRES, a novel system for scene-level 3D reconstruction from posed images. Designed to work with as few as one view, 3DFIRES reconstructs the complete geometry of unseen scenes, including hidden surfaces. With multiple view inputs, our method produces full reconstruction within all camera frustums. A key feature of our approach is the fusion of multi-view information at the feature level, enabling the production of coherent and comprehensive 3D reconstruction. We train our system on non-watertight scans from large-scale real scene dataset. We show it matches the efficacy of single-view reconstruction methods with only one input and surpasses existing techniques in both quantitative and qualitative measures for sparse-view 3D reconstruction.

Dandan Shan, Kaichun Mo, Wei Yang et al.

The majority of modern robot learning methods focus on learning a set of pre-defined tasks with limited or no generalization to new tasks. Extending the robot skillset to novel tasks involves gathering an extensive amount of training data for additional tasks. In this paper, we address the problem of teaching new tasks to robots using human demonstration videos for repetitive tasks (e.g., packing). This task requires understanding the human video to identify which object is being manipulated (the pick object) and where it is being placed (the placement slot). In addition, it needs to re-identify the pick object and the placement slots during inference along with the relative poses to enable robot execution of the task. To tackle this, we propose SLeRP, a modular system that leverages several advanced visual foundation models and a novel slot-level placement detector Slot-Net, eliminating the need for expensive video demonstrations for training. We evaluate our system using a new benchmark of real-world videos. The evaluation results show that SLeRP outperforms several baselines and can be deployed on a real robot.

Yiwen Zhang, Joseph Tung, Ruojin Cai et al.

3D foundation models (3DFMs) have recently transformed 3D vision, enabling joint prediction of depths, poses, and point maps directly from images. Yet their ability to reason under extreme, non-overlapping views remains largely unexplored. In this work, we study their internal representations and find that 3DFMs exhibit an emergent understanding of extreme-view geometry, despite never being trained for such conditions. To further enhance these capabilities, we introduce a lightweight alignment scheme that refines their internal 3D representation by tuning only a small subset of backbone bias terms, leaving all decoder heads frozen. This targeted adaptation substantially improves relative pose estimation under extreme viewpoints without degrading per-image depth or point quality. Additionally, we contribute MegaUnScene, a new benchmark of Internet scenes unseen by existing 3DFMs, with dedicated test splits for both relative pose estimation and dense 3D reconstruction. All code and data will be released.

Rudy Morel, Francesco Pio Ramunno, Jeff Shen et al.

Conditional diffusion models provide a natural framework for probabilistic prediction of dynamical systems and have been successfully applied to fluid dynamics and weather prediction. However, in many settings, the available information at a given time represents only a small fraction of what is needed to predict future states, either due to measurement uncertainty or because only a small fraction of the state can be observed. This is true for example in solar physics, where we can observe the Sun's surface and atmosphere, but its evolution is driven by internal processes for which we lack direct measurements. In this paper, we tackle the probabilistic prediction of partially observable, long-memory dynamical systems, with applications to solar dynamics and the evolution of active regions. We show that standard inference schemes, such as autoregressive rollouts, fail to capture long-range dependencies in the data, largely because they do not integrate past information effectively. To overcome this, we propose a multiscale inference scheme for diffusion models, tailored to physical processes. Our method generates trajectories that are temporally fine-grained near the present and coarser as we move farther away, which enables capturing long-range temporal dependencies without increasing computational cost. When integrated into a diffusion model, we show that our inference scheme significantly reduces the bias of the predicted distributions and improves rollout stability.

Sarah Jabbour, David Fouhey, Nikola Banovic et al.

AI has the potential to augment human decision making. However, even high-performing models can produce inaccurate predictions when deployed. These inaccuracies, combined with automation bias, where humans overrely on AI predictions, can result in worse decisions. Selective prediction, in which potentially unreliable model predictions are hidden from users, has been proposed as a solution. This approach assumes that when AI abstains and informs the user so, humans make decisions as they would without AI involvement. To test this assumption, we study the effects of selective prediction on human decisions in a clinical context. We conducted a user study of 259 clinicians tasked with diagnosing and treating hospitalized patients. We compared their baseline performance without any AI involvement to their AI-assisted accuracy with and without selective prediction. Our findings indicate that selective prediction mitigates the negative effects of inaccurate AI in terms of decision accuracy. Compared to no AI assistance, clinician accuracy declined when shown inaccurate AI predictions (66% [95% CI: 56%-75%] vs. 56% [95% CI: 46%-66%]), but recovered under selective prediction (64% [95% CI: 54%-73%]). However, while selective prediction nearly maintains overall accuracy, our results suggest that it alters patterns of mistakes: when informed the AI abstains, clinicians underdiagnose (18% increase in missed diagnoses) and undertreat (35% increase in missed treatments) compared to no AI input at all. Our findings underscore the importance of empirically validating assumptions about how humans engage with AI within human-AI systems.

Giancarlo Pereira, Yidan Gao, Yurii Piadyk et al.

High speed, high-resolution, and accurate 3D scanning would open doors to many new applications in graphics, robotics, science, and medicine by enabling the accurate scanning of deformable objects during interactions. Past attempts to use structured light, time-of-flight, and stereo in high-speed settings have usually required tradeoffs in resolution or inaccuracy. In this paper, we introduce a method that enables, for the first time, 3D scanning at 450 frames per second at 1~Megapixel, or 1,450 frames per second at 0.4~Megapixel in an environment with controlled lighting. The key idea is to use a per-pixel lookup table that maps colors to depths, which is built using a linear stage. Imperfections, such as lens-distortion and sensor defects are baked into the calibration. We describe our method and test it on a novel hardware prototype. We compare the system with both ground-truth geometry as well as commercially available dynamic sensors like the Microsoft Kinect and Intel Realsense. Our results show the system acquiring geometry of objects undergoing high-speed deformations and oscillations and demonstrate the ability to recover physical properties from the reconstructions.

Sarah Jabbour, David Fouhey, Ella Kazerooni et al.

Objective: When patients develop acute respiratory failure, accurately identifying the underlying etiology is essential for determining the best treatment. However, differentiating between common medical diagnoses can be challenging in clinical practice. Machine learning models could improve medical diagnosis by aiding in the diagnostic evaluation of these patients. Materials and Methods: Machine learning models were trained to predict the common causes of acute respiratory failure (pneumonia, heart failure, and/or COPD). Models were trained using chest radiographs and clinical data from the electronic health record (EHR) and applied to an internal and external cohort. Results: The internal cohort of 1,618 patients included 508 (31%) with pneumonia, 363 (22%) with heart failure, and 137 (8%) with COPD based on physician chart review. A model combining chest radiographs and EHR data outperformed models based on each modality alone. Models had similar or better performance compared to a randomly selected physician reviewer. For pneumonia, the combined model area under the receiver operating characteristic curve (AUROC) was 0.79 (0.77-0.79), image model AUROC was 0.74 (0.72-0.75), and EHR model AUROC was 0.74 (0.70-0.76). For heart failure, combined: 0.83 (0.77-0.84), image: 0.80 (0.71-0.81), and EHR: 0.79 (0.75-0.82). For COPD, combined: AUROC = 0.88 (0.83-0.91), image: 0.83 (0.77-0.89), and EHR: 0.80 (0.76-0.84). In the external cohort, performance was consistent for heart failure and increased for COPD, but declined slightly for pneumonia. Conclusions: Machine learning models combining chest radiographs and EHR data can accurately differentiate between common causes of acute respiratory failure. Further work is needed to determine how these models could act as a diagnostic aid to clinicians in clinical settings.

Sarah Jabbour, David Fouhey, Ella Kazerooni et al.

While deep learning has shown promise in improving the automated diagnosis of disease based on chest X-rays, deep networks may exhibit undesirable behavior related to shortcuts. This paper studies the case of spurious class skew in which patients with a particular attribute are spuriously more likely to have the outcome of interest. For instance, clinical protocols might lead to a dataset in which patients with pacemakers are disproportionately likely to have congestive heart failure. This skew can lead to models that take shortcuts by heavily relying on the biased attribute. We explore this problem across a number of attributes in the context of diagnosing the cause of acute hypoxemic respiratory failure. Applied to chest X-rays, we show that i) deep nets can accurately identify many patient attributes including sex (AUROC = 0.96) and age (AUROC >= 0.90), ii) they tend to exploit correlations between such attributes and the outcome label when learning to predict a diagnosis, leading to poor performance when such correlations do not hold in the test population (e.g., everyone in the test set is male), and iii) a simple transfer learning approach is surprisingly effective at preventing the shortcut and promoting good generalization performance. On the task of diagnosing congestive heart failure based on a set of chest X-rays skewed towards older patients (age >= 63), the proposed approach improves generalization over standard training from 0.66 (95% CI: 0.54-0.77) to 0.84 (95% CI: 0.73-0.92) AUROC. While simple, the proposed approach has the potential to improve the performance of models across populations by encouraging reliance on clinically relevant manifestations of disease, i.e., those that a clinician would use to make a diagnosis.

Dimitri Zhukov, Jean-Baptiste Alayrac, Ramazan Gokberk Cinbis et al.

In this paper we investigate learning visual models for the steps of ordinary tasks using weak supervision via instructional narrations and an ordered list of steps instead of strong supervision via temporal annotations. At the heart of our approach is the observation that weakly supervised learning may be easier if a model shares components while learning different steps: `pour egg' should be trained jointly with other tasks involving `pour' and `egg'. We formalize this in a component model for recognizing steps and a weakly supervised learning framework that can learn this model under temporal constraints from narration and the list of steps. Past data does not permit systematic studying of sharing and so we also gather a new dataset, CrossTask, aimed at assessing cross-task sharing. Our experiments demonstrate that sharing across tasks improves performance, especially when done at the component level and that our component model can parse previously unseen tasks by virtue of its compositionality.

Ashish Kumar, Saurabh Gupta, David Fouhey et al.

Humans routinely retrace paths in a novel environment both forwards and backwards despite uncertainty in their motion. This paper presents an approach for doing so. Given a demonstration of a path, a first network generates a path abstraction. Equipped with this abstraction, a second network observes the world and decides how to act to retrace the path under noisy actuation and a changing environment. The two networks are optimized end-to-end at training time. We evaluate the method in two realistic simulators, performing path following and homing under actuation noise and environmental changes. Our experiments show that our approach outperforms classical approaches and other learning based baselines.

Saurabh Gupta, David Fouhey, Sergey Levine et al.

This works presents a formulation for visual navigation that unifies map based spatial reasoning and path planning, with landmark based robust plan execution in noisy environments. Our proposed formulation is learned from data and is thus able to leverage statistical regularities of the world. This allows it to efficiently navigate in novel environments given only a sparse set of registered images as input for building representations for space. Our formulation is based on three key ideas: a learned path planner that outputs path plans to reach the goal, a feature synthesis engine that predicts features for locations along the planned path, and a learned goal-driven closed loop controller that can follow plans given these synthesized features. We test our approach for goal-driven navigation in simulated real world environments and report performance gains over competitive baseline approaches.

Shubham Tulsiani, Saurabh Gupta, David Fouhey et al.

The goal of this paper is to take a single 2D image of a scene and recover the 3D structure in terms of a small set of factors: a layout representing the enclosing surfaces as well as a set of objects represented in terms of shape and pose. We propose a convolutional neural network-based approach to predict this representation and benchmark it on a large dataset of indoor scenes. Our experiments evaluate a number of practical design questions, demonstrate that we can infer this representation, and quantitatively and qualitatively demonstrate its merits compared to alternate representations.