In this paper, we address the issue of modeling and estimating changes in the state of the spatio-temporal dynamical systems based on a sequence of observations like video frames. Traditional numerical simulation systems depend largely on the initial settings and correctness of the constructed partial differential equations (PDEs). Despite recent efforts yielding significant success in discovering data-driven PDEs with neural networks, the limitations posed by singular scenarios and the absence of local insights prevent them from performing effectively in a broader real-world context. To this end, this paper propose the universal expert module -- that is, optical flow estimation component, to capture the evolution laws of general physical processes in a data-driven fashion. To enhance local insight, we painstakingly design a finer-grained physical pipeline, since local characteristics may be influenced by various internal contextual information, which may contradict the macroscopic properties of the whole system. Further, we harness currently popular neural discrete learning to unveil the underlying important features in its latent space, this process better injects interpretability, which can help us obtain a powerful prior over these discrete random variables. We conduct extensive experiments and ablations to demonstrate that the proposed framework achieves large performance margins, compared with the existing SOTA baselines.
In supervised learning, models are trained to extract correlations from a static dataset. This often leads to models that rely on high-level misconceptions. To prevent such misconceptions, we must necessarily provide additional information beyond the training data. Existing methods incorporate forms of additional instance-level supervision, such as labels for spurious features or additional labeled data from a balanced distribution. Such strategies can become prohibitively costly for large-scale datasets since they require additional annotation at a scale close to the original training data. We hypothesize that targeted natural language feedback about a model's misconceptions is a more efficient form of additional supervision. We introduce Clarify, a novel interface and method for interactively correcting model misconceptions. Through Clarify, users need only provide a short text description to describe a model's consistent failure patterns. Then, in an entirely automated way, we use such descriptions to improve the training process by reweighting the training data or gathering additional targeted data. Our user studies show that non-expert users can successfully describe model misconceptions via Clarify, improving worst-group accuracy by an average of 17.1% in two datasets. Additionally, we use Clarify to find and rectify 31 novel hard subpopulations in the ImageNet dataset, improving minority-split accuracy from 21.1% to 28.7%.
Current state-of-the-art Video-based Person Re-Identification (Re-ID) primarily relies on appearance features extracted by deep learning models. These methods are not applicable for long-term analysis in real-world scenarios where persons have changed clothes, making appearance information unreliable. In this work, we deal with the practical problem of Video-based Cloth-Changing Person Re-ID (VCCRe-ID) by proposing "Attention-based Shape and Gait Representations Learning" (ASGL) for VCCRe-ID. Our ASGL framework improves Re-ID performance under clothing variations by learning clothing-invariant gait cues using a Spatial-Temporal Graph Attention Network (ST-GAT). Given the 3D-skeleton-based spatial-temporal graph, our proposed ST-GAT comprises multi-head attention modules, which are able to enhance the robustness of gait embeddings under viewpoint changes and occlusions. The ST-GAT amplifies the important motion ranges and reduces the influence of noisy poses. Then, the multi-head learning module effectively reserves beneficial local temporal dynamics of movement. We also boost discriminative power of person representations by learning body shape cues using a GAT. Experiments on two large-scale VCCRe-ID datasets demonstrate that our proposed framework outperforms state-of-the-art methods by 12.2% in rank-1 accuracy and 7.0% in mAP.
Backhaul traffic congestion caused by the video traffic of a few popular files can be alleviated by storing the to-be-requested content at various levels in wireless video caching networks. Typically, content service providers (CSPs) own the content, and the users request their preferred content from the CSPs using their (wireless) internet service providers (ISPs). As these parties do not reveal their private information and business secrets, traditional techniques may not be readily used to predict the dynamic changes in users' future demands. Motivated by this, we propose a novel resource-aware hierarchical federated learning (RawHFL) solution for predicting user's future content requests. A practical data acquisition technique is used that allows the user to update its local training dataset based on its requested content. Besides, since networking and other computational resources are limited, considering that only a subset of the users participate in the model training, we derive the convergence bound of the proposed algorithm. Based on this bound, we minimize a weighted utility function for jointly configuring the controllable parameters to train the RawHFL energy efficiently under practical resource constraints. Our extensive simulation results validate the proposed algorithm's superiority, in terms of test accuracy and energy cost, over existing baselines.
In the control of lower-limb exoskeletons with feet, the phase in the gait cycle can be identified by monitoring the weight distribution at the feet. This phase information can be used in the exoskeleton's controller to compensate the dynamics of the exoskeleton and to assign impedance parameters. Typically the weight distribution is calculated using data from sensors such as treadmill force plates or insole force sensors. However, these solutions increase both the setup complexity and cost. For this reason, we propose a deep-learning approach that uses a short time window of joint kinematics to predict the weight distribution of an exoskeleton in real time. The model was trained on treadmill walking data from six users wearing a four-degree-of-freedom exoskeleton and tested in real time on three different users wearing the same device. This test set includes two users not present in the training set to demonstrate the model's ability to generalize across individuals. Results show that the proposed method is able to fit the actual weight distribution with R2=0.9 and is suitable for real-time control with prediction times less than 1 ms. Experiments in closed-loop exoskeleton control show that deep-learning-based weight distribution estimation can be used to replace force sensors in overground and treadmill walking.
Infrared and visible image fusion has emerged as a prominent research in computer vision. However, little attention has been paid on complex scenes fusion, causing existing techniques to produce sub-optimal results when suffers from real interferences. To fill this gap, we propose a decomposition-based and interference perception image fusion method. Specifically, we classify the pixels of visible image from the degree of scattering of light transmission, based on which we then separate the detail and energy information of the image. This refined decomposition facilitates the proposed model in identifying more interfering pixels that are in complex scenes. To strike a balance between denoising and detail preservation, we propose an adaptive denoising scheme for fusing detail components. Meanwhile, we propose a new weighted fusion rule by considering the distribution of image energy information from the perspective of multiple directions. Extensive experiments in complex scenes fusions cover adverse weathers, noise, blur, overexposure, fire, as well as downstream tasks including semantic segmentation, object detection, salient object detection and depth estimation, consistently indicate the effectiveness and superiority of the proposed method compared with the recent representative methods.
Active visual SLAM finds a wide array of applications in GNSS-Denied sub-terrain environments and outdoor environments for ground robots. To achieve robust localization and mapping accuracy, it is imperative to incorporate the perception considerations in the goal selection and path planning towards the goal during an exploration mission. Through this work, we propose FIT-SLAM (Fisher Information and Traversability estimation-based Active SLAM), a new exploration method tailored for unmanned ground vehicles (UGVs) to explore 3D environments. This approach is devised with the dual objectives of sustaining an efficient exploration rate while optimizing SLAM accuracy. Initially, an estimation of a global traversability map is conducted, which accounts for the environmental constraints pertaining to traversability. Subsequently, we propose a goal candidate selection approach along with a path planning method towards this goal that takes into account the information provided by the landmarks used by the SLAM backend to achieve robust localization and successful path execution . The entire algorithm is tested and evaluated first in a simulated 3D world, followed by a real-world environment and is compared to pre-existing exploration methods. The results obtained during this evaluation demonstrate a significant increase in the exploration rate while effectively minimizing the localization covariance.
Neurons in the brain communicate information via punctual events called spikes. The timing of spikes is thought to carry rich information, but it is not clear how to leverage this in digital systems. We demonstrate that event-based encoding is efficient for audio compression. To build this event-based representation we use a deep binary auto-encoder, and under high sparsity pressure, the model enters a regime where the binary event matrix is stored more efficiently with sparse matrix storage algorithms. We test this on the large MAESTRO dataset of piano recordings against vector quantized auto-encoders. Not only does our "Spiking Music compression" algorithm achieve a competitive compression/reconstruction trade-off, but selectivity and synchrony between encoded events and piano key strikes emerge without supervision in the sparse regime.
Driven by the rapid ascent of artificial intelligence (AI), organizations are at the epicenter of a seismic shift, facing a crucial question: How can AI be successfully integrated into existing operations? To help answer it, manage expectations and mitigate frustration, this article introduces Computational Management, a systematic approach to task automation for enhancing the ability of organizations to harness AI's potential within existing workflows. Computational Management acts as a bridge between the strategic insights of management science with the analytical rigor of computational thinking. The article offers three easy step-by-step procedures to begin the process of implementing AI within a workflow. Such procedures focus on task (re)formulation, on the assessment of the automation potential of tasks, on the completion of task specification templates for AI selection and adaptation. Included in the article there are manual and automated methods, with prompt suggestions for publicly available LLMs, to complete these three procedures. The first procedure, task (re)formulation, focuses on breaking down work activities into basic units, so they can be completed by one agent, involve a single well-defined action, and produce a distinct outcome. The second, allows the assessment of the granular task and its suitability for automation, using the Task Automation Index to rank tasks based on whether they have standardized input, well-defined rules, repetitiveness, data dependency, and objective outputs. The third, focuses on a task specification template which details information on 16 critical components of tasks, and can be used as a checklist to select or adapt the most suitable AI solution for integration into existing workflows. Computational Management provides a roadmap and a toolkit for humans and AI to thrive together, while enhancing organizational efficiency and innovation.
Continuous physical interaction between robots and their environment is a requirement in many industrial and household tasks, such as sanding and cleaning. Due to the complex tactile information, these tasks are notoriously difficult to model and to sense. In this article, we introduce a closed-loop control method that is constrained to surfaces. The applications that we target have in common that they can be represented by probability distributions on the surface that correlate to the time the robot should spend in a region. These surfaces can easily be captured jointly with the target distributions using coloured point clouds. We present the extension of an ergodic control approach that can be used with point clouds, based on heat equation-driven area coverage (HEDAC). Our method enables closed-loop exploration by measuring the actual coverage using vision. Unlike existing approaches, we approximate the potential field from non-stationary diffusion using spectral acceleration, which does not require complex preprocessing steps and achieves real-time closed-loop control frequencies. We exploit geometric algebra to stay in contact with the target surface by tracking a line while simultaneously exerting a desired force along that line. Our approach is suitable for fully autonomous and human-robot interaction settings where the robot can either directly measure the coverage of the target with its sensors or by being guided online by markings or annotations of a human expert. We tested the performance of the approach in kinematic simulation using point clouds, ranging from the Stanford bunny to a variety of kitchen utensils. Our real-world experiments demonstrate that the proposed approach can successfully be used to wash kitchenware with curved surfaces, by cleaning the dirt detected by vision in an online manner. Website: https://geometric-algebra.tobiloew.ch/tactile_ergodic_control