In this paper, we first propose a unified approach for analyzing the stability of the phaseless operator for both amplitude and intensity measurement on an arbitrary geometric set, thus characterizing the robust performance of phase retrieval via the empirical minimization method. We introduce the random embedding of concave lifting operators in tangent space to characterize the unified analysis of any geometric set. Similarly, we investigate the structured matrix recovery problem through the robust injectivity of a linear rank one measurement operator on an arbitrary matrix set. The core of our analysis is to establish a unified empirical chaos process characterization for various matrix sets. Talagrand's $\gamma_{\alpha}$-functionals are introduced to characterize the connection between the geometric constraints and the number of measurements needed to guarantee stability or robust injectivity. Finally, we construct adversarial noise to demonstrate the sharpness of the recovery bounds in the above two scenarios.
In this paper, we address the challenge of unsupervised mistake detection in egocentric procedural video through the analysis of gaze signals. Traditional supervised mistake detection methods rely on manually labeled mistakes, and hence suffer from domain-dependence and scalability issues. We introduce an unsupervised method for detecting mistakes in videos of human activities, overcoming the challenges of domain-specific requirements and the need for annotated data. We postulate that, when a subject is making a mistake in the execution of a procedure, their attention patterns will deviate from normality. We hence propose to detect mistakes by comparing gaze trajectories predicted from input video with ground truth gaze signals collected through a gaze tracker. Since predicting gaze in video is characterized by high uncertainty, we propose a novel \textit{gaze completion task}, which aims to predict gaze from visual observations and partial gaze trajectories. We further contribute a \textit{gaze completion approach} based on a Gaze-Frame Correlation module to explicitly model the correlation between gaze information and each local visual token. Inconsistencies between the predicted and observed gaze trajectories act as an indicator for identifying mistakes. Experiments on the EPIC-Tent, HoloAssist and IndustReal datasets showcase the effectiveness of the proposed approach as compared to unsupervised and one-class techniques. Our method is ranked first on the HoloAssist Mistake Detection challenge.
In this paper, the application of imitation learning in caregiving robotics is explored, aiming at addressing the increasing demand for automated assistance in caring for the elderly and disabled. Leveraging advancements in deep learning and control algorithms, the study focuses on training neural network policies using offline demonstrations. A key challenge addressed is the "Policy Stopping" problem, crucial for enhancing safety in imitation learning-based policies, particularly diffusion policies. Novel solutions proposed include ensemble predictors and adaptations of the normalizing flow-based algorithm for early anomaly detection. Comparative evaluations against anomaly detection methods like VAE and Tran-AD demonstrate superior performance on assistive robotics benchmarks. The paper concludes by discussing the further research in integrating safety models into policy training, crucial for the reliable deployment of neural network policies in caregiving robotics.
In this paper, we address the development of a robotic rehabilitation system for the upper limbs based on collaborative end-effector solutions. The use of commercial collaborative robots offers significant advantages for this task, as they are optimized from an engineering perspective and ensure safe physical interaction with humans. However, they also come with noticeable drawbacks, such as the limited range of sizes available on the market and the standard control modes, which are primarily oriented towards industrial or service applications. To address these limitations, we propose an optimization-based design method to fully exploit the capability of the cobot in performing rehabilitation tasks. Additionally, we introduce a novel control architecture based on an admittance-type Virtual Fixture method, which constrains the motion of the robot along a prescribed path. This approach allows for an intuitive definition of the task to be performed via Programming by Demonstration and enables the system to operate both passively and actively. In passive mode, the system supports the patient during task execution with additional force, while in active mode, it opposes the motion with a braking force. Experimental results demonstrate the effectiveness of the proposed method.
In this paper, we present a new approach to improving the relevance and reliability of medical IR, which builds upon the concept of Level of Evidence (LoE). LoE framework categorizes medical publications into 7 distinct levels based on the underlying empirical evidence. Despite LoE framework's relevance in medical research and evidence-based practice, only few medical publications explicitly state their LoE. Therefore, we develop a classification model for automatically assigning LoE to medical publications, which successfully classifies over 26 million documents in MEDLINE database into LoE classes. The subsequent retrieval experiments on TREC PM datasets show substantial improvements in retrieval relevance, when LoE is used as a search filter.
In this paper, we build a general model of memristors suitable for the simulation of event-based systems, such as hardware spiking neural networks, and more generally, neuromorphic computing systems. We extend an existing general model of memristors - the Generalised Metastable Switch Model - to an event-driven setting, eliminating errors associated discrete time approximation, as well as offering potential improvements in terms of computational efficiency for simulation. We introduce the notion of a volatility state variable, to allow for the modelling of memory-dependent and dynamic switching behaviour, succinctly capturing and unifying a variety of volatile phenomena present in memristive devices, including state relaxation, structural disruption, Joule heating, and drift acceleration phenomena. We supply a drift dataset for titanium dioxide memristors and introduce a linear conductance model to simulate the drift characteristics, motivated by a proposed physical model of filament growth. We then demonstrate an approach for fitting the parameters of the event-based model to the drift model.
In this paper, we introduce a novel and computationally efficient method for vertex embedding, community detection, and community size determination. Our approach leverages a normalized one-hot graph encoder and a rank-based cluster size measure. Through extensive simulations, we demonstrate the excellent numerical performance of our proposed graph encoder ensemble algorithm.
In this paper, we address the problem of enclosing an arbitrarily moving target in three dimensions by a single pursuer, which is an unmanned aerial vehicle (UAV), for maximum coverage while also ensuring the pursuer's safety by preventing collisions with the target. The proposed guidance strategy steers the pursuer to a safe region of space surrounding the target, allowing it to maintain a certain distance from the latter while offering greater flexibility in positioning and converging to any orbit within this safe zone. Our approach is distinguished by the use of nonholonomic constraints to model vehicles with accelerations serving as control inputs and coupled engagement kinematics to craft the pursuer's guidance law meticulously. Furthermore, we leverage the concept of the Lyapunov Barrier Function as a powerful tool to constrain the distance between the pursuer and the target within asymmetric bounds, thereby ensuring the pursuer's safety within the predefined region. To validate the efficacy and robustness of our algorithm, we conduct experimental tests by implementing a high-fidelity quadrotor model within Software-in-the-loop (SITL) simulations, encompassing various challenging target maneuver scenarios. The results obtained showcase the resilience of the proposed guidance law, effectively handling arbitrarily maneuvering targets, vehicle/autopilot dynamics, and external disturbances. Our method consistently delivers stable global enclosing behaviors, even in response to aggressive target maneuvers, and requires only relative information for successful execution.
In this paper, we investigate the extrapolation capabilities of implicit deep learning models in handling unobserved data, where traditional deep neural networks may falter. Implicit models, distinguished by their adaptability in layer depth and incorporation of feedback within their computational graph, are put to the test across various extrapolation scenarios: out-of-distribution, geographical, and temporal shifts. Our experiments consistently demonstrate significant performance advantage with implicit models. Unlike their non-implicit counterparts, which often rely on meticulous architectural design for each task, implicit models demonstrate the ability to learn complex model structures without the need for task-specific design, highlighting their robustness in handling unseen data.
In this paper, we address the issue of estimating and inferring the distributional treatment effects in randomized experiments. The distributional treatment effect provides a more comprehensive understanding of treatment effects by characterizing heterogeneous effects across individual units, as opposed to relying solely on the average treatment effect. To enhance the precision of distributional treatment effect estimation, we propose a regression adjustment method that utilizes the distributional regression and pre-treatment information. Our method is designed to be free from restrictive distributional assumptions. We establish theoretical efficiency gains and develop a practical, statistically sound inferential framework. Through extensive simulation studies and empirical applications, we illustrate the substantial advantages of our method, equipping researchers with a powerful tool for capturing the full spectrum of treatment effects in experimental research.
In this paper, we propose a novel multi-task learning architecture, which incorporates recent advances in attention mechanisms. Our approach, the Multi-Task Attention Network (MTAN), consists of a single shared network containing a global feature pool, together with task-specific soft-attention modules, which are trainable in an end-to-end manner. These attention modules allow for learning of task-specific features from the global pool, whilst simultaneously allowing for features to be shared across different tasks. The architecture can be built upon any feed-forward neural network, is simple to implement, and is parameter efficient. Experiments on the CityScapes dataset show that our method outperforms several baselines in both single-task and multi-task learning, and is also more robust to the various weighting schemes in the multi-task loss function. We further explore the effectiveness of our method through experiments over a range of task complexities, and show how our method scales well with task complexity compared to baselines.