Iterative Learning Approach Research Articles

An Active Learning Framework Improves Tumor Variant Interpretation.

A novel machine learning approach predicts the impact of tumor mutations on cellular phenotypes, overcomes limited training data, minimizes costly functional validation, and advances efforts to implement cancer precision medicine.

A learning control scheme for upper-limb exoskeleton via adaptive sliding mode technique

This paper investigates an iterative learning approach integrated with sliding mode control method to accomplish passive rehabilitation therapy tasks for wearable 6 degrees of freedom (DOF) upper-limb exoskeleton. Firstly, the motion data of human body is collected from a healthy subject through VICON motion capture system and reasonable trajectories in joint space are generated by fitting functions. Secondly, an iterative learning controller is developed to estimate the iteration-invariant dynamic parameters which are complicated and difficult to be obtained precisely in practice. Note that the identical initial condition (i.i.c) in traditional iterative learning control (ILC) is released by applying the polynomial reconstruction method. Considering the uncertainties and disturbances which affect the system in the form of friction, backlash and unexpected tissue torques from human body, an adaptive law is proposed to estimate the upper bound of the lumped non-periodic disturbances. Based on that, sliding mode controller is conducted to achieve the robustness over the time domain, while the chattering phenomenon is attenuated by applying tanh function. Afterwards, the stability and convergence of the overall system is rigorously proved with a composite energy function (CEF) composed of tracking and estimating errors. Finally, co-simulation and experiment results are presented to demonstrate the effectiveness of the proposed control scheme.

A fidelity-embedded learning for metal artifact reduction in dental CBCT.

Dental cone-beam computed tomography (CBCT) has been increasingly used for dental and maxillofacial imaging. However, the presence of metallic inserts, such as implants, crowns, and dental braces, violates the CT model assumption, which leads to severe metal artifacts in the reconstructed CBCT image, resulting in the degradation of diagnostic performance. In this study, we used deep learning to reduce metalartifacts. The metal artifacts, appearing as streaks and shadows, are nonlocal and highly associated with various factors, including the geometry of metallic inserts, energy-dependent attenuation, and energy spectrum of the incident X-ray beam, making it difficult to learn their complicated structures directly. To provide a step-by-step environment in which deep learning can be trained, we propose an iterative learning approach in which the network at each iteration step learns the correction error caused by the previous network, while enforcing the data fidelity in the projection domain. To generate a realistic paired training dataset, metal-free CBCT scans were collected from patients without metallic inserts, and then simulated metal projection data were added to generate the corresponding metal-corrupted projectiondata. The feasibility of the proposed method was investigated in clinical metal-affected CBCT scans, as well as simulated metal-affected CBCT scans. The results show that the proposed method significantly reduces metal artifacts while preserving the morphological structures near metallic objects and outperforms direct image domainlearning. The proposed fidelity-embedded learning can effectively reduce metal artifacts in dental CBCT compared with direct image domainlearning.

Maintaining the Status Quo: Simultaneous Estimation and Elimination for Multiple Interference in Transform Domain Vehicular Communication

Wireless communication in vehicular networks is highly vulnerable to electromagnetic interference, which seriously undermines the communication effectiveness. Thus, interference estimation and elimination are essential in vehicular communications. To solve the multitasking problem of interference estimation and cancelation within the transformation domain of vehicular communications where they are divided into the consistency and inconsistency based on the different conditions of spectrum sensing, the present study proposes an iterative Bayesian sparse learning approach for simultaneous interference recovery and elimination which applies multiple Bayesian learning for measured signals, and the structured sparse learning method for simultaneous interference mitigation with signal recovery. Furthermore, the updating optimization based on Bayesian estimation framework is performed to facilitate the learning process and improve the performance of interference estimation and signal reconstruction. Simulation results show that the proposed algorithms achieved more rapid convergence and more accurate recovery (31.84% and 27.88% higher in spectrum sensing with consistency and inconsistency, respectively) than conventional algorithms. During all iterations, the proposed algorithms, because of updating optimization, were more robust than conventional ones, but the computational complexity remained largely the same. To explore the correlation between the parameters and the interference recovery performance, we performed experiments, which revealed linear, logarithmic, and exponential regression correlations for particular forms of interference. We also compared the demodulation performance of the proposed algorithm with existing methods, and it turned out the proposed algorithm also obtained higher gains on average in the bit error rate.

Terminal Iterative Learning Scheme for Consensus Problem in Multi-Agent Systems with State Constraints

In this paper, we investigate the consensus problem of multi-agent systems with state constraints. To achieve the consensus effectively, the terminal iterative learning approach is proposed. This learning strategy is designed without the tracking error. And the consensus state is obtained by the the information interaction between agents. Meanwhile, the constraint condition holds in terms of our learning strategy. It shows the consensus conditions ensure the achievement of the constraints. Finally, a numerical simulation is given to illustrate the effectiveness of the main results.

Development of an innovative noise generation system for turboprop aircraft fuselage testing

This paper presents the development of a cabin noise testing equipment that will be used to evaluate the interior noise of regional aircraft as well as to aid the development of noise reduction techniques. The innovative noise generation system consists of three loudspeaker arrays positioned around the fuselage circumference to synthesize a pressure field that is similar to the pressure field seen by the fuselage during a flight. The acoustic pressure field generated by the loudspeakers is measured by a number of microphones scattered on the fuselage surface. These microphone signals are then fed back to the controller with the purpose of minimising the error between the target pressure field and the measured one by means of an iterative learning approach. The number and location of the microphones used in the control loop are selected through a pre-test optimisation analysis, which aims to reduce the time and cost of the set-up. A small-scale electroacoustic demonstrator has been built to develop the feedback control approach. A frequency domain multi-input multi-output feedback controller is used to replicate the random pressure field generated by the turbulent boundary layer excitation. The multi-harmonics of the propeller induced excitation are then added to the time histories of the broadband noise using a time waveform replication technique. Different arrangements of the driving signal distribution are investigated, and the results are then presented in terms of accuracy of the pressure field reproduction.

Bounding Box-Free Instance Segmentation Using Semi-Supervised Iterative Learning for Vehicle Detection

Vehicle classification is a hot computer vision topic, with studies ranging from ground-view to top-view imagery. Top-view images allow understanding city patterns, traffic management, among others. However, there are some difficulties for pixel-wise classification: most vehicle classification studies use object detection methods, and most publicly available datasets are designed for this task, creating instance segmentation datasets is laborious, and traditional instance segmentation methods underperform on this task since the objects are small. Thus, the present research objectives are as follows: first, propose a novel semisupervised iterative learning approach using the geographic information system software, second, propose a box-free instance segmentation approach, and third, provide a city-scale vehicle dataset. The iterative learning procedure considered the following: first, labeling a few vehicles from the entire scene, second, choosing training samples near those areas, third, training the deep learning model (U-net with efficient-net-B7 backbone), fourth, classifying the whole scene, fifth, converting the predictions into shapefile, sixth, correcting areas with wrong predictions, seventh, including them in the training data, eighth repeating until results are satisfactory. We considered vehicle interior and borders to separate instances using a semantic segmentation model. When removing the borders, the vehicle interior becomes isolated, allowing for unique object identification. Our procedure is very efficient and accurate for generating data iteratively, which resulted in 122 567 mapped vehicles. Metrics-wise, our method presented higher intersection over union when compared to box-based methods (82% against 72%), and per-object metrics surpassed 90% for precision and recall.

Event-triggered data-driven control of discrete-time nonlinear systems with unknown disturbance

This paper is dedicated to event-triggered data-driven control of nonlinear systems with unknown disturbance via model free iterative learning approach. An extended state observer is employed to reconstruct the disturbance in system output. An event-triggered model free iterative learning control strategy is constructed by system input, system output and the reconstructed disturbance. Sufficient conditions are proposed to make the resultant tracking error system be uniform ultimate bounded. Simulation examples are provided to validate the effectiveness of the proposed scheme.

Frequency-Domain Data-Driven Adaptive Iterative Learning Control Approach: With Application to Wafer Stage

The feedforward control is becoming increasingly important in ultra-precision stages. However, the conventional model-based methods cannot achieve expected performance in new-generation stages since it is hard to obtain the accurate plant model due to the complicated stage dynamical properties. To tackle this problem, this article develops a model-free data-driven adaptive iterative learning approach that is designed in the frequency-domain. Explicitly, the proposed method utilizes the frequency-response data to learn and update the output of the feedforward controller online, which has benefits that both the structure and parameters of the plant model are not required. An unbiased estimation method for the frequency response of the closed-loop system is proposed and proved through the theoretical analysis. Comparative experiments on a linear motor confirm the effectiveness and superiority of the proposed method, and show that it has the ability to avoid the performance deterioration caused by the model mismatch with the increasing iterative trials.

Broadening the reach: How effectual entrepreneurs build impact ecosystems across geographical divide

Effectuation theory focuses on entrepreneurs using available means to create artefacts such as firms and markets. In this study, we explore three cases of entrepreneurial innovation processes initiated in developed countries but aimed at low-income markets of developing countries. In these entrepreneurial processes happening simultaneously in two geographical locations, firms often struggle to understand the local institutional and market conditions. We find that the studied entrepreneurs have been able to develop and disseminate solutions for addressing grand challenges in emerging markets by building an “impact ecosystem”. This required the entrepreneurs to work with diverse stakeholders, pioneer, have an iterative learning approach and be deliberately inefficient in their operations. Our results contribute by linking effectuation theory with the literature on ecosystems and demonstrating how entrepreneurs can engage in effectuation processes in building impact ecosystems.

SIR-GN: A Fast Structural Iterative Representation Learning Approach For Graph Nodes

Graph representation learning methods have attracted an increasing amount of attention in recent years. These methods focus on learning a numerical representation of the nodes in a graph. Learning these representations is a powerful instrument for tasks such as graph mining, visualization, and hashing. They are of particular interest because they facilitate the direct use of standard machine learning models on graphs. Graph representation learning methods can be divided into two main categories: methods preserving the connectivity information of the nodes and methods preserving nodes’ structural information. Connectivity-based methods focus on encoding relationships between nodes, with connected nodes being closer together in the resulting latent space. While methods preserving structure generate a latent space where nodes serving a similar structural function in the network are encoded close to each other, independently of them being connected or even close to each other in the graph. While there are a lot of works that focus on preserving node connectivity, only a few works focus on preserving nodes’ structure. Properly encoding nodes’ structural information is fundamental for many real-world applications as it has been demonstrated that this information can be leveraged to successfully solve many tasks where connectivity-based methods usually fail. A typical example is the task of node classification, i.e., the assignment or prediction of a particular label for a node. Current limitations of structural representation methods are their scalability, representation meaning, and no formal proof that guaranteed the preservation of structural properties. We propose a new graph representation learning method, called Structural Iterative Representation learning approach for Graph Nodes ( SIR-GN ). In this work, we propose two variations ( SIR-GN: GMM and SIR-GN: K-Means ) and show how our best variation SIR-GN: K-Means : (1) theoretically guarantees the preservation of graph structural similarities, (2) provides a clear meaning about its representation and a way to interpret it with a specifically designed attribution procedure, and (3) is scalable and fast to compute. In addition, from our experiment, we show that SIR-GN: K-Means is often better or, in the worst-case comparable than the existing structural graph representation learning methods present in the literature. Also, we empirically show its superior scalability and computational performance when compared to other existing approaches.

Iterative learning-based path control for robot-assisted upper-limb rehabilitation

AbstractIn robot-assisted rehabilitation, the performance of robotic assistance is dependent on the human user’s dynamics, which are subject to uncertainties. In order to enhance the rehabilitation performance and in particular to provide a constant level of assistance, we separate the task space into two subspaces where a combined scheme of adaptive impedance control and trajectory learning is developed. Human movement speed can vary from person to person and it cannot be predefined for the robot. Therefore, in the direction of human movement, an iterative trajectory learning approach is developed to update the robot reference according to human movement and to achieve the desired interaction force between the robot and the human user. In the direction normal to the task trajectory, human’s unintentional force may deteriorate the trajectory tracking performance. Therefore, an impedance adaptation method is utilized to compensate for unknown human force and prevent the human user drifting away from the updated robot reference trajectory. The proposed scheme was tested in experiments that emulated three upper-limb rehabilitation modes: zero interaction force, assistive and resistive. Experimental results showed that the desired assistance level could be achieved, despite uncertain human dynamics.

Identifying HIV-related digital social influencers using an iterative deep learning approach.

Community popular opinion leaders have played a critical role in HIV prevention interventions. However, it is often difficult to identify these 'HIV influencers' who are qualified and willing to promote HIV campaigns, especially online, because social media influencers change frequently. We sought to use an iterative deep learning framework to automatically discover HIV-related online social influencers. Out of 1.15 million Twitter users' data from March 2018 to March 2020, we extracted tweets from 1099 Twitter users who had mentioned the keywords 'HIV' or 'AIDS'. Two Twitter users determined to be 'online HIV influencers' based on their conversation topics and engagement were hand-picked by domain experts and used as a seed training dataset. We modelled social influence and discovered new potential influencers based on these seeds using a graph neural network model. We tested the model's precision and recall compared with other baseline model approaches. We validated the results through manual verification. The model identified 23 new (manually verified) HIV-related influencers, including health and research organizations and local HIV advocates across the United States. Our proposed model achieved the highest accuracy/recall, with an average improvement of 38.5% over the other baseline models. Results suggest that iterative deep learning models can be used to automatically identify new and changing key HIV-related influencers online. We discuss the implications and potential of HIV researchers/departments applying this approach across online big data (e.g. hundreds of millions of social media posts per day) to help promote HIV prevention campaigns to affected communities.

Consensus Control for Heterogeneous Multivehicle Systems: An Iterative Learning Approach.

This article investigates the consensus tracking problem of the heterogeneous multivehicle systems (MVSs) under a repeatable control environment. First, a unified iterative learning control (ILC) algorithm is presented for all autonomous vehicles, each of which is governed by both discrete- and continuous-time nonlinear dynamics. Then, several consensus criteria for MVSs with switching topology and external disturbances are established based on our proposed distributed ILC protocols. For discrete-time systems, all vehicles can perfectly track to the common reference trajectory over a specified finite time interval, and the corresponding digraphs may not have spanning trees. Existing approaches dealing with the continuous-time systems generally require that all vehicles have strictly identical initial conditions, being too ideal in practice. We relax this unpractical assumption and propose an extra distributed initial state learning protocol such that vehicles can take different initial states, leading to the fact that the finite time tracking is achieved ultimately regardless of the initial errors. Finally, a numerical example demonstrates the effectiveness of our theoretical results.

The Link Between Organizational Agility And VUCA – An Agile Assessment Model

This paper summarizes the elements of agility listed in contemporary scientific literature and aligns them to the four different forces described by the acronym VUCA. There is a general consensus that organizational resilience and agility are crucial elements to gain sustainable competitive advantage especially since the arresting economic changes during the pandemic year 2020. There is little agreement though how precisely organizational agility responds to volatility, uncertainty, complexity, and ambiguity (VUCA). By systemization of frameworks and models from the literary sources the aim is to create a combined model. Consequently, the purpose of this paper is to develop a model that can practically be employed to evaluate the agile status quo of an organization. It should support practitioners to distinguish between the different requirements each part of the VUCA acronym requires. The design of this study is a conceptual paper which proposes a new corporate agility model with direct linkages to the requirements of VUCA forces. It derives from, and builds on, existing scientific literature while integrating a comprehensive set of existing agile frameworks. The result of this research is a model which defines distinct enablers, sensing capabilities, agile practices, and responses as answers to VUCA forces. It highlights that certain elements reinforce and link to each other to transform an entire organization agile. Surprisingly, leadership and strategy are not incorporated as elements but rather seen as conditions that subordinate the iterative learning approach of this model. Originality is given as the created model provides a set of elements leaders can implement or use to assess organizational agility. Furthermore, it supports how to deal with strategy and leadership in the transitional process. Relevance is given as during the worlds COVID-19 crises the economic and governments changes in terms of volatility and uncertainty were especially challenging for organizations.

A Novel Iterative Learning Approach for Tracking Control of High-Speed Trains Subject to Unknown Time-Varying Delay

In this article, a novel iterative learning control scheme is proposed for high-speed trains, aiming to track the desired reference displacement and velocity, where the Krasovskii function is constructed to compensate for the negative influence of unknown time-varying speed delays. The main feature of the proposed approach is that the hyperbolic tangent function and the command filter are integrated into the learning controller to overcome the singularity problem that may occur during the control process and relax the requirement for the derivability of the desired velocity. The stability of control system is strictly proved through establishing the composite energy function, and the effectiveness is confirmed via numerical simulations. Compared with the existing works, the merits of the proposed control scheme lie in that more general nonlinear uncertainties are imposed on the dynamic model of train instead of the Lipschitz condition, and the reference acceleration assigned by the railway department is not required. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —High-speed train always runs periodically on the same railway, e.g., the same tunnels, slopes, and bridges, according to the scheduling plans developed by the railway department. Due to the repetitive operation pattern, the iterative learning control has the prospect of becoming an inherent method for devising the tracking controller of trains. Nevertheless, the unknown speed delays, which are inevitable due to the damping effect of wheel rails, couplers, and so on as well as the disturbance of external environments, may degrade the performance of control system and even cause instability in severe cases. As a result, this article exploits a compensation method to eliminate the effects of unknown delay under the iterative learning control framework, thus guaranteeing the safety of train operation and the comfort of passengers. To enhance the practicability, the hyperbolic tangent function is introduced to keep the continuity of control signal, and the command filter is synthesized to reduce the complexity of controller implementation. Although the stability analysis and numerical simulations have confirmed the feasibility and effectiveness of the proposed scheme, it is still expected to be verified by experiments in the future.

Computational prediction of species-specific yeast DNA replication origin via iterative feature representation.

Deoxyribonucleic acid replication is one of the most crucial tasks taking place in the cell, and it has to be precisely regulated. This process is initiated in the replication origins (ORIs), and thus it is essential to identify such sites for a deeper understanding of the cellular processes and functions related to the regulation of gene expression. Considering the important tasks performed by ORIs, several experimental and computational approaches have been developed in the prediction of such sites. However, existing computational predictors for ORIs have certain curbs, such as building only single-feature encoding models, limited systematic feature engineering efforts and failure to validate model robustness. Hence, we developed a novel species-specific yeast predictor called yORIpred that accurately identify ORIs in the yeast genomes. To develop yORIpred, we first constructed optimal 40 baseline models by exploring eight different sequence-based encodings and five different machine learning classifiers. Subsequently, the predicted probability of 40 models was considered as the novel feature vector and carried out iterative feature learning approach independently using five different classifiers. Our systematic analysis revealed that the feature representation learned by the support vector machine algorithm (yORIpred) could well discriminate the distribution characteristics between ORIs and non-ORIs when compared with the other four algorithms. Comprehensive benchmarking experiments showed that yORIpred achieved superior and stable performance when compared with the existing predictors on the same training datasets. Furthermore, independent evaluation showcased the best and accurate performance of yORIpred thus underscoring the significance of iterative feature representation. To facilitate the users in obtaining their desired results without undergoing any mathematical, statistical or computational hassles, we developed a web server for the yORIpred predictor, which is available at: http://thegleelab.org/yORIpred.

Iterative learning approach to active noise control of highly autocorrelated signals with applications to machinery noise

This study describes an iterative learning approach to the active control of machinery noise with high autocorrelation properties. In contrast to typical active noise control solutions, which work by adapting the transfer function of the controller, in the iterative learning control one adapts the control signal itself. Special care was taken to develop a generic solution that can handle different sorts of secondary path models including very long and non-minimum phase finite impulse response filters. To achieve that, the authors used spectral factorisation and exploit the fact that, for non-minimum phase systems, a stable inverse can be constructed if the causality constraint is relaxed and later restored by taking advantage of the periodicity of the attenuated signal. The resulting controller can be efficiently implemented on a sample-to-sample calculation basis. The behaviour and the performance of the proposed scheme are studied using computer simulations and real-world experiments on noises from an electric transformer and functional magnetic resonance imaging device. The proposed solution was also compared to normalised feedforward filtered-X least mean squares algorithm and performed much better in terms of attenuation, convergence, and robustness.

Neural Membrane Mutual Coupling Characterisation Using Entropy-Based Iterative Learning Identification

This paper investigates the interaction phenomena of the coupled axons while the mutual coupling factor is presented as a pairwise description. Based on the Hodgkin-Huxley model and the coupling factor matrix, the membrane potentials of the coupled myelinated/unmyelinated axons are quantified which implies that the neural coupling can be characterised by the presented coupling factor. Meanwhile the equivalent electric circuit is supplied to illustrate the physical meaning of this extended model. In order to estimate the coupling factor, a data-based iterative learning identification algorithm is presented where the Rényi entropy of the estimation error has been minimised. The convergence of the presented algorithm is analysed and the learning rate is designed. To verified the presented model and the algorithm, the numerical simulation results indicate the correctness and the effectiveness. Furthermore, the statistical description of the neural coupling, the approximation using ordinary differential equation, the measurement and the conduction of the nerve signals are discussed respectively as advanced topics. The novelties can be summarised as follows: 1) the Hodgkin-Huxley model has been extended considering the mutual interaction between the neural axon membranes, 2) the iterative learning approach has been developed for factor identification using entropy criterion, and 3) the theoretical framework has been established for this class of system identification problems with convergence analysis.

An iterative learning approach to economic model predictive control for an integrated solar thermal system

An iterative learning (IL) approach to disturbance prediction for economic model predictive control (EMPC) is proposed and applied to an integrated solar thermal system (ISTS). The disturbance in the system, which is the user hot water demand, is predicted iteratively by taking advantage of the repetitive nature of hot water consumption and utilized by EMPC for improved ISTS control performance. Various user load scenarios are developed for simulations based on historical data, and the performance of the proposed control method is compared against an idealistic EMPC scheme with perfect load information along with existing EMPC methods and a baseline proportional-integral controller. It is demonstrated that the proposed IL approach to EMPC achieves electrical costs within 0.5% of the idealistic case while outperforming all other methods in both energy savings and output temperature management.