R. James Cotton

Paweł A. Pierzchlewicz, R. James Cotton, Mohammad Bashiri et al.

Due to depth ambiguities and occlusions, lifting 2D poses to 3D is a highly ill-posed problem. Well-calibrated distributions of possible poses can make these ambiguities explicit and preserve the resulting uncertainty for downstream tasks. This study shows that previous attempts, which account for these ambiguities via multiple hypotheses generation, produce miscalibrated distributions. We identify that miscalibration can be attributed to the use of sample-based metrics such as minMPJPE. In a series of simulations, we show that minimizing minMPJPE, as commonly done, should converge to the correct mean prediction. However, it fails to correctly capture the uncertainty, thus resulting in a miscalibrated distribution. To mitigate this problem, we propose an accurate and well-calibrated model called Conditional Graph Normalizing Flow (cGNFs). Our model is structured such that a single cGNF can estimate both conditional and marginal densities within the same model - effectively solving a zero-shot density estimation problem. We evaluate cGNF on the Human~3.6M dataset and show that cGNF provides a well-calibrated distribution estimate while being close to state-of-the-art in terms of overall minMPJPE. Furthermore, cGNF outperforms previous methods on occluded joints while it remains well-calibrated.

R. James Cotton, Anthony Cimorelli, Kunal Shah et al.

Markerless pose estimation allows reconstructing human movement from multiple synchronized and calibrated views, and has the potential to make movement analysis easy and quick, including gait analysis. This could enable much more frequent and quantitative characterization of gait impairments, allowing better monitoring of outcomes and responses to interventions. However, the impact of different keypoint detectors and reconstruction algorithms on markerless pose estimation accuracy has not been thoroughly evaluated. We tested these algorithmic choices on data acquired from a multicamera system from a heterogeneous sample of 25 individuals seen in a rehabilitation hospital. We found that using a top-down keypoint detector and reconstructing trajectories with an implicit function enabled accurate, smooth and anatomically plausible trajectories, with a noise in the step width estimates compared to a GaitRite walkway of only 8mm.

R. James Cotton, Allison DeLillo, Anthony Cimorelli et al.

Markerless motion capture using computer vision and human pose estimation (HPE) has the potential to expand access to precise movement analysis. This could greatly benefit rehabilitation by enabling more accurate tracking of outcomes and providing more sensitive tools for research. There are numerous steps between obtaining videos to extracting accurate biomechanical results and limited research to guide many critical design decisions in these pipelines. In this work, we analyze several of these steps including the algorithm used to detect keypoints and the keypoint set, the approach to reconstructing trajectories for biomechanical inverse kinematics and optimizing the IK process. Several features we find important are: 1) using a recent algorithm trained on many datasets that produces a dense set of biomechanically-motivated keypoints, 2) using an implicit representation to reconstruct smooth, anatomically constrained marker trajectories for IK, 3) iteratively optimizing the biomechanical model to match the dense markers, 4) appropriate regularization of the IK process. Our pipeline makes it easy to obtain accurate biomechanical estimates of movement in a rehabilitation hospital.

R. James Cotton

There has been significant progress in machine learning algorithms for human pose estimation that may provide immense value in rehabilitation and movement sciences. However, there remain several challenges to routine use of these tools for clinical practice and translational research, including: 1) high technical barrier to entry, 2) rapidly evolving space of algorithms, 3) challenging algorithmic interdependencies, and 4) complex data management requirements between these components. To mitigate these barriers, we developed a human pose estimation pipeline that facilitates running state-of-the-art algorithms on data acquired in clinical context. Our system allows for running different implementations of several classes of algorithms and handles their interdependencies easily. These algorithm classes include subject identification and tracking, 2D keypoint detection, 3D joint location estimation, and estimating the pose of body models. The system uses a database to manage videos, intermediate analyses, and data for computations at each stage. It also provides tools for data visualization, including generating video overlays that also obscure faces to enhance privacy. Our goal in this work is not to train new algorithms, but to advance the use of cutting-edge human pose estimation algorithms for clinical and translation research. We show that this tool facilitates analyzing large numbers of videos of human movement ranging from gait laboratories analyses, to clinic and therapy visits, to people in the community. We also highlight limitations of these algorithms when applied to clinical populations in a rehabilitation setting.

R. James Cotton, Emoonah McClerklin, Anthony Cimorelli et al.

Human pose estimation from monocular video is a rapidly advancing field that offers great promise to human movement science and rehabilitation. This potential is tempered by the smaller body of work ensuring the outputs are clinically meaningful and properly calibrated. Gait analysis, typically performed in a dedicated lab, produces precise measurements including kinematics and step timing. Using over 7000 monocular video from an instrumented gait analysis lab, we trained a neural network to map 3D joint trajectories and the height of individuals onto interpretable biomechanical outputs including gait cycle timing and sagittal plane joint kinematics and spatiotemporal trajectories. This task specific layer produces accurate estimates of the timing of foot contact and foot off events. After parsing the kinematic outputs into individual gait cycles, it also enables accurate cycle-by-cycle estimates of cadence, step time, double and single support time, walking speed and step length.

R. James Cotton, J. D. Peiffer, Kunal Shah et al.

Markerless motion capture (MMC) is revolutionizing gait analysis in clinical settings by making it more accessible, raising the question of how to extract the most clinically meaningful information from gait data. In multiple fields ranging from image processing to natural language processing, self-supervised learning (SSL) from large amounts of unannotated data produces very effective representations for downstream tasks. However, there has only been limited use of SSL to learn effective representations of gait and movement, and it has not been applied to gait analysis with MMC. One SSL objective that has not been applied to gait is contrastive learning, which finds representations that place similar samples closer together in the learned space. If the learned similarity metric captures clinically meaningful differences, this could produce a useful representation for many downstream clinical tasks. Contrastive learning can also be combined with causal masking to predict future timesteps, which is an appealing SSL objective given the dynamical nature of gait. We applied these techniques to gait analyses performed with MMC in a rehabilitation hospital from a diverse clinical population. We find that contrastive learning on unannotated gait data learns a representation that captures clinically meaningful information. We probe this learned representation using the framework of biomarkers and show it holds promise as both a diagnostic and response biomarker, by showing it can accurately classify diagnosis from gait and is responsive to inpatient therapy, respectively. We ultimately hope these learned representations will enable predictive and prognostic gait-based biomarkers that can facilitate precision rehabilitation through greater use of MMC to quantify movement in rehabilitation.

Seth Donahue, J. D. Peiffer, R. Tyler Richardson et al.

To validate a clinically accessible approach for quantifying the Upper Extremity Reachable Workspace (UERW) using a single (monocular) camera and Artificial Intelligence (AI)-driven Markerless Motion Capture (MMC) for biomechanical analysis. Objective assessment and validation of these techniques for specific clinically oriented tasks are crucial for their adoption in clinical motion analysis. AI-driven monocular MMC reduces the barriers to adoption in the clinic and has the potential to reduce the overhead for analysis of this common clinical assessment. Nine adult participants with no impairments performed the standardized UERW task, which entails reaching targets distributed across a virtual sphere centered on the torso, with targets displayed in a VR headset. Movements were simultaneously captured using a marker-based motion capture system and a set of eight FLIR cameras. We performed monocular video analysis on two of these video camera views to compare a frontal and offset camera configurations. The frontal camera orientation demonstrated strong agreement with the marker-based reference, exhibiting a minimal mean bias of $0.61 \pm 0.12$ \% reachspace reached per octanct (mean $\pm$ standard deviation). In contrast, the offset camera view underestimated the percent workspace reached ($-5.66 \pm 0.45$ \% reachspace reached). Conclusion: The findings support the feasibility of a frontal monocular camera configuration for UERW assessment, particularly for anterior workspace evaluation where agreement with marker-based motion capture was highest. The overall performance demonstrates clinical potential for practical, single-camera assessments. This study provides the first validation of monocular MMC system for the assessment of the UERW task. By reducing technical complexity, this approach enables broader implementation of quantitative upper extremity mobility assessment.

Divya Joshi, J. D. Peiffer, Colleen Peyton et al.

arly identification of motor impairment in infancy relies on expert visual assessment of spontaneous movement, motivating the development of automated, objective alternatives. One promising approach is using computer vision, which benefits from high quality pose estimation from video. In this study, we systematically evaluated three state-of-the-art pose estimation frameworks (MeTRAbs-ACAE, SAM 3D Body, and Sapiens) on 100 videos over 13 sessions of 8 infants recorded with a multi-view markerless motion capture system. We quantified keypoint detection accuracy using reprojection error, geometric consistency, and Procrustes-aligned 3D position error, and demonstrated proof-of-concept for fitting an inverse kinematic framework to infant data. While Sapiens achieved the lowest reprojection error and highest geometric consistency of the methods evaluated (22.8 pixels and 0.82, respectively), SAM 3D Body provided the most comprehensive 3D information for kinematic reconstruction with Procrustes-aligned position errors of 19 to 28 mm. We demonstrate in a case comparison example that biomechanical models fit to SAM 3D estimates distinguish representative movement patterns in infants related to motor development, as identified by a clinical expert. Together, these findings highlight both the promise and current limitations of 3D pose estimation for infant biomechanics and establish preliminary groundwork for scalable, video-based assessment of early motor development.

R. James Cotton, Colleen Peyton

Easy access to precise 3D tracking of movement could benefit many aspects of rehabilitation. A challenge to achieving this goal is that while there are many datasets and pretrained algorithms for able-bodied adults, algorithms trained on these datasets often fail to generalize to clinical populations including people with disabilities, infants, and neonates. Reliable movement analysis of infants and neonates is important as spontaneous movement behavior is an important indicator of neurological function and neurodevelopmental disability, which can help guide early interventions. We explored the application of dynamic Gaussian splatting to sparse markerless motion capture (MMC) data. Our approach leverages semantic segmentation masks to focus on the infant, significantly improving the initialization of the scene. Our results demonstrate the potential of this method in rendering novel views of scenes and tracking infant movements. This work paves the way for advanced movement analysis tools that can be applied to diverse clinical populations, with a particular emphasis on early detection in infants.

Seth Donahue, Irina Djuraskovic, Kunal Shah et al.

Video-based human movement analysis holds potential for movement assessment in clinical practice and research. However, the clinical implementation and trust of multi-view markerless motion capture (MMMC) require that, in addition to being accurate, these systems produce reliable confidence intervals to indicate how accurate they are for any individual. Building on our prior work utilizing variational inference to estimate joint angle posterior distributions, this study evaluates the calibration and reliability of a probabilistic MMMC method. We analyzed data from 68 participants across two institutions, validating the model against an instrumented walkway and standard marker-based motion capture. We measured the calibration of the confidence intervals using the Expected Calibration Error (ECE). The model demonstrated reliable calibration, yielding ECE values generally < 0.1 for both step and stride length and bias-corrected gait kinematics. We observed a median step and stride length error of ~16 mm and ~12 mm respectively, with median bias-corrected kinematic errors ranging from 1.5 to 3.8 degrees across lower extremity joints. Consistent with the calibrated ECE, the magnitude of the model's predicted uncertainty correlated strongly with observed error measures. These findings indicate that, as designed, the probabilistic model reconstruction quantifies epistemic uncertainty, allowing it to identify unreliable outputs without the need for concurrent ground-truth instrumentation.

R. James Cotton, Pouyan Firouzabadi, Wendy Murray

Accurate hand and finger tracking from video has significant clinical applications for monitoring activities of daily living and measuring range of motion, yet monocular video approaches for obtaining hand biomechanics remain under-developed. We present a method that combines the SAM 3D Body foundation model with inverse kinematics optimization in a full-body biomechanical model to extract anatomically-constrained finger joint angles from single-view video. We port SAM 3D Body from PyTorch to JAX for integration with MuJoCo-MJX, enabling GPU-accelerated optimization, and develop a novel mapping between the Momentum Human Rig (MHR) outputs and biomechanical model markers. Validation against 8-camera multiview reconstruction on 4,590 frames from 7 participants performing a variety of hand poses and object manipulation tasks shows finger joint angle errors of approximately 10 degrees and hand position errors of approximately 6 mm, after Procrustes alignment. Results were consistent across camera viewpoints and robust to different methods for producing reference values from multiview video. This work extends monocular biomechanical analysis to detailed finger tracking, expanding access to quantitative characterization of hand movement from readily available video.

R. James Cotton, Thomas Leonard

Markerless motion capture is making quantitative movement analysis increasingly accessible, yet analyzing the resulting data remains a barrier for clinicians without programming expertise. We present BiomechAgent, a code-generating AI agent that enables biomechanical analysis through natural language and allows users to querying databases, generating visualizations, and even interpret data without requiring users to write code. To evaluate BiomechAgent's capabilities, we developed a systematic benchmark spanning data retrieval, visualization, activity classification, temporal segmentation, and clinical reasoning. BiomechAgent achieved robust accuracy on data retrieval and visualization tasks and demonstrated emerging clinical reasoning capabilities. We used our dataset to systematically evaluate several of our design decisions. Biomechanically-informed, domain-specific instructions significantly improved performance over generic prompts, and integrating validated specialized tools for gait event detection substantially boosted accuracy on challenging spatiotemporal analysis where the base agent struggled. We also tested BiomechAgent using a local open-weight model instead of a frontier cloud based LLM and found that perform was substantially diminished in most domains other than database retrieval. In short, BiomechAgent makes the data from accessible motion capture and much more useful and accessible to end users.

R. James Cotton

Recent developments have created differentiable physics simulators designed for machine learning pipelines that can be accelerated on a GPU. While these can simulate biomechanical models, these opportunities have not been exploited for biomechanics research or markerless motion capture. We show that these simulators can be used to fit inverse kinematics to markerless motion capture data, including scaling the model to fit the anthropomorphic measurements of an individual. This is performed end-to-end with an implicit representation of the movement trajectory, which is propagated through the forward kinematic model to minimize the error from the 3D markers reprojected into the images. The differential optimizer yields other opportunities, such as adding bundle adjustment during trajectory optimization to refine the extrinsic camera parameters or meta-optimization to improve the base model jointly over trajectories from multiple participants. This approach improves the reprojection error from markerless motion capture over prior methods and produces accurate spatial step parameters compared to an instrumented walkway for control and clinical populations.

Tim Unger, Arash Sal Moslehian, J. D. Peiffer et al. · eth-zurich

Marker-based Optical Motion Capture (OMC) paired with biomechanical modeling is currently considered the most precise and accurate method for measuring human movement kinematics. However, combining differentiable biomechanical modeling with Markerless Motion Capture (MMC) offers a promising approach to motion capture in clinical settings, requiring only minimal equipment, such as synchronized webcams, and minimal effort for data collection. This study compares key kinematic outcomes from biomechanically modeled MMC and OMC data in 15 stroke patients performing the drinking task, a functional task recommended for assessing upper limb movement quality. We observed a high level of agreement in kinematic trajectories between MMC and OMC, as indicated by high correlations (median r above 0.95 for the majority of kinematic trajectories) and median RMSE values ranging from 2-5 degrees for joint angles, 0.04 m/s for end-effector velocity, and 6 mm for trunk displacement. Trial-to-trial biases between OMC and MMC were consistent within participant sessions, with interquartile ranges of bias around 1-3 degrees for joint angles, 0.01 m/s in end-effector velocity, and approximately 3mm for trunk displacement. Our findings indicate that our MMC for arm tracking is approaching the accuracy of marker-based methods, supporting its potential for use in clinical settings. MMC could provide valuable insights into movement rehabilitation in stroke patients, potentially enhancing the effectiveness of rehabilitation strategies.

Nikolaos Smyrnakis, Tasos Karakostas, R. James Cotton

Gait analysis from videos obtained from a smartphone would open up many clinical opportunities for detecting and quantifying gait impairments. However, existing approaches for estimating gait parameters from videos can produce physically implausible results. To overcome this, we train a policy using reinforcement learning to control a physics simulation of human movement to replicate the movement seen in video. This forces the inferred movements to be physically plausible, while improving the accuracy of the inferred step length and walking velocity.

R. James Cotton, Fabian Sinz

Advances in multiview markerless motion capture (MMMC) promise high-quality movement analysis for clinical practice and research. While prior validation studies show MMMC performs well on average, they do not provide what is needed in clinical practice or for large-scale utilization of MMMC -- confidence intervals over specific kinematic estimates from a specific individual analyzed using a possibly unique camera configuration. We extend our previous work using an implicit representation of trajectories optimized end-to-end through a differentiable biomechanical model to learn the posterior probability distribution over pose given all the detected keypoints. This posterior probability is learned through a variational approximation and estimates confidence intervals for individual joints at each moment in a trial, showing confidence intervals generally within 10-15 mm of spatial error for virtual marker locations, consistent with our prior validation studies. Confidence intervals over joint angles are typically only a few degrees and widen for more distal joints. The posterior also models the correlation structure over joint angles, such as correlations between hip and pelvis angles. The confidence intervals estimated through this method allow us to identify times and trials where kinematic uncertainty is high.

R. James Cotton, Bryant A. Seamon, Richard L. Segal et al.

Precision rehabilitation offers the promise of an evidence-based approach for optimizing individual rehabilitation to improve long-term functional outcomes. Emerging techniques, including those driven by artificial intelligence, are rapidly expanding our ability to quantify the different domains of function during rehabilitation, other encounters with healthcare, and in the community. While this seems poised to usher rehabilitation into the era of big data and should be a powerful driver of precision rehabilitation, our field lacks a coherent framework to utilize these data and deliver on this promise. We propose a framework that builds upon multiple existing pillars to fill this gap. Our framework aims to identify the Optimal Dynamic Treatment Regimens (ODTR), or the decision-making strategy that takes in the range of available measurements and biomarkers to identify interventions likely to maximize long-term function. This is achieved by designing and fitting causal models, which extend the Computational Neurorehabilitation framework using tools from causal inference. These causal models can learn from heterogeneous data from different silos, which must include detailed documentation of interventions, such as using the Rehabilitation Treatment Specification System. The models then serve as digital twins of patient recovery trajectories, which can be used to learn the ODTR. Our causal modeling framework also emphasizes quantitatively linking changes across levels of the functioning to ensure that interventions can be precisely selected based on careful measurement of impairments while also being selected to maximize outcomes that are meaningful to patients and stakeholders. We believe this approach can provide a unifying framework to leverage growing big rehabilitation data and AI-powered measurements to produce precision rehabilitation treatments that can improve clinical outcomes.

Paweł A. Pierzchlewicz, Caio O. da Silva, R. James Cotton et al.

Single camera 3D pose estimation is an ill-defined problem due to inherent ambiguities from depth, occlusion or keypoint noise. Multi-hypothesis pose estimation accounts for this uncertainty by providing multiple 3D poses consistent with the 2D measurements. Current research has predominantly concentrated on generating multiple hypotheses for single frame static pose estimation or single hypothesis motion estimation. In this study we focus on the new task of multi-hypothesis motion estimation. Multi-hypothesis motion estimation is not simply multi-hypothesis pose estimation applied to multiple frames, which would ignore temporal correlation across frames. Instead, it requires distributions which are capable of generating temporally consistent samples, which is significantly more challenging than multi-hypothesis pose estimation or single-hypothesis motion estimation. To this end, we introduce Platypose, a framework that uses a diffusion model pretrained on 3D human motion sequences for zero-shot 3D pose sequence estimation. Platypose outperforms baseline methods on multiple hypotheses for motion estimation. Additionally, Platypose also achieves state-of-the-art calibration and competitive joint error when tested on static poses from Human3.6M, MPI-INF-3DHP and 3DPW. Finally, because it is zero-shot, our method generalizes flexibly to different settings such as multi-camera inference.

J. D. Peiffer, Kunal Shah, Irina Djuraskovic et al.

The way a person moves is a direct reflection of their neurological and musculoskeletal health, yet it remains one of the most underutilized vital signs in clinical practice. Although clinicians visually observe movement impairments, they lack accessible and validated methods to objectively measure movement in routine care. This gap prevents wider use of biomechanical measurements in practice, which could enable more sensitive outcome measures or earlier identification of impairment. We present our Portable Biomechanics Laboratory (PBL), which includes a secure, cloud-enabled smartphone app for data collection and a novel algorithm for fitting biomechanical models to this data. We extensively validated PBL's biomechanical measures using a large, clinically representative dataset. Next, we tested the usability and utility of our system in neurosurgery and sports medicine clinics. We found joint angle errors within 3 degrees across participants with neurological injury, lower-limb prosthesis users, pediatric inpatients, and controls. In addition to being easy to use, gait metrics computed from the PBL showed high reliability and were sensitive to clinical differences. For example, in individuals undergoing decompression surgery for cervical myelopathy, the mJOA score is a common patient-reported outcome measure; we found that PBL gait metrics correlated with mJOA scores and demonstrated greater responsiveness to surgical intervention than the patient-reported outcomes. These findings support the use of handheld smartphone video as a scalable, low-burden tool for capturing clinically meaningful biomechanical data, offering a promising path toward accessible monitoring of mobility impairments. We release the first clinically validated method for measuring whole-body kinematics from handheld smartphone video at https://intelligentsensingandrehabilitation.github.io/MonocularBiomechanics/ .

Sara Rajaram, R. James Cotton, Fabian H. Sinz

Preference-based Reinforcement Learning (PbRL) entails a variety of approaches for aligning models with human intent to alleviate the burden of reward engineering. However, most previous PbRL work has not investigated the robustness to labeler errors, inevitable with labelers who are non-experts or operate under time constraints. Additionally, PbRL algorithms often target very specific settings (e.g. pairwise ranked preferences or purely offline learning). We introduce Similarity as Reward Alignment (SARA), a simple contrastive framework that is both resilient to noisy labels and adaptable to diverse feedback formats and training paradigms. SARA learns a latent representation of preferred samples and computes rewards as similarities to the learned latent. We demonstrate strong performance compared to baselines on continuous control offline RL benchmarks. We further demonstrate SARA's versatility in applications such as trajectory filtering for downstream tasks, cross-task preference transfer, and reward shaping in online learning.

Ruize Yang, Ann Kennedy, R. James Cotton

Advances in markerless motion capture are expanding access to biomechanical movement analysis, making it feasible to obtain high-quality movement data from outpatient clinics, inpatient hospitals, therapy, and even home. Expanding access to movement data in these diverse contexts makes the challenge of performing downstream analytics all the more acute. Creating separate bespoke analysis code for all the tasks end users might want is both intractable and does not take advantage of the common features of human movement underlying them all. Recent studies have shown that fine-tuning language models to accept tokenized movement as an additional modality enables successful descriptive captioning of movement. Here, we explore whether such a multimodal motion-language model can answer detailed, clinically meaningful questions about movement. We collected over 30 hours of biomechanics from nearly 500 participants, many with movement impairments from a variety of etiologies, performing a range of movements used in clinical outcomes assessments. After tokenizing these movement trajectories, we created a multimodal dataset of motion-related questions and answers spanning a range of tasks. We developed BiomechGPT, a multimodal biomechanics-language model, on this dataset. Our results show that BiomechGPT demonstrates high performance across a range of tasks such as activity recognition, identifying movement impairments, diagnosis, scoring clinical outcomes, and measuring walking. BiomechGPT provides an important step towards a foundation model for rehabilitation movement data.

R. James Cotton

Broader access to high-quality movement analysis could greatly benefit movement science and rehabilitation, such as allowing more detailed characterization of movement impairments and responses to interventions, or even enabling early detection of new neurological conditions or fall risk. While emerging technologies are making it easier to capture kinematics with biomechanical models, or how joint angles change over time, inferring the underlying physics that give rise to these movements, including ground reaction forces, joint torques, or even muscle activations, is still challenging. Here we explore whether imitation learning applied to a biomechanical model from a large dataset of movements from able-bodied and impaired individuals can learn to compute these inverse dynamics. Although imitation learning in human pose estimation has seen great interest in recent years, our work differences in several ways: we focus on using an accurate biomechanical model instead of models adopted for computer vision, we test it on a dataset that contains participants with impaired movements, we reported detailed tracking metrics relevant for the clinical measurement of movement including joint angles and ground contact events, and finally we apply imitation learning to a muscle-driven neuromusculoskeletal model. We show that our imitation learning policy, KinTwin, can accurately replicate the kinematics of a wide range of movements, including those with assistive devices or therapist assistance, and that it can infer clinically meaningful differences in joint torques and muscle activations. Our work demonstrates the potential for using imitation learning to enable high-quality movement analysis in clinical practice.

Kandan Ramakrishnan, R. James Cotton, Xaq Pitkow et al.

Recent work on object-centric world models aim to factorize representations in terms of objects in a completely unsupervised or self-supervised manner. Such world models are hypothesized to be a key component to address the generalization problem. While self-supervision has shown improved performance however, OOD generalization has not been systematically and explicitly tested. In this paper, we conduct an extensive study on the generalization properties of contrastive world model. We systematically test the model under a number of different OOD generalization scenarios such as extrapolation to new object attributes, introducing new conjunctions or new attributes. Our experiments show that the contrastive world model fails to generalize under the different OOD tests and the drop in performance depends on the extent to which the samples are OOD. When visualizing the transition updates and convolutional feature maps, we observe that any changes in object attributes (such as previously unseen colors, shapes, or conjunctions of color and shape) breaks down the factorization of object representations. Overall, our work highlights the importance of object-centric representations for generalization and current models are limited in their capacity to learn such representations required for human-level generalization.

R. James Cotton, Fabian H. Sinz, Andreas S. Tolias

In recent years, artificial neural networks have achieved state-of-the-art performance for predicting the responses of neurons in the visual cortex to natural stimuli. However, they require a time consuming parameter optimization process for accurately modeling the tuning function of newly observed neurons, which prohibits many applications including real-time, closed-loop experiments. We overcome this limitation by formulating the problem as $K$-shot prediction to directly infer a neuron's tuning function from a small set of stimulus-response pairs using a Neural Process. This required us to developed a Factorized Neural Process, which embeds the observed set into a latent space partitioned into the receptive field location and the tuning function properties. We show on simulated responses that the predictions and reconstructed receptive fields from the Factorized Neural Process approach ground truth with increasing number of trials. Critically, the latent representation that summarizes the tuning function of a neuron is inferred in a quick, single forward pass through the network. Finally, we validate this approach on real neural data from visual cortex and find that the predictive accuracy is comparable to -- and for small $K$ even greater than -- optimization based approaches, while being substantially faster. We believe this novel deep learning systems identification framework will facilitate better real-time integration of artificial neural network modeling into neuroscience experiments.