Sequential Task Research Articles

Online Linearized Confidence‐Weighted Learning on a Budget

Online learning aims to solve a sequence of consecutive prediction tasks by leveraging the knowledge gained from previous tasks. Linearized confidence‐weighted (LCW) learning is the first online learning algorithm introducing the concept of weight confidence into the prediction model through distributions over weights. It provides the flexibility for weights to update their values at different scales. The kernel trick in machine learning can be applied to LCW for a better prediction performance. However, the kernel‐based LCW algorithm is subject to the curse of kernelization which makes it vulnerable to the unlimited growth of the prediction model in runtime and memory consumption. In this study, we present the budgeted LCW (BLCW) algorithm which puts a limit on the growth by a predefined budget with optimization. Consequently, BLCW performs the LCW update and then reduces the information loss by projection. Based on the resource perspective that reinterprets LCW in terms of resources and utilization degrees, we demonstrated that BLCW approximates the kernel‐based LCW algorithm. We evaluate four budget maintenance strategies and suggest that the mean removal is the most stable. By various numerical experiments on real datasets, we demonstrate that BLCW performs competitively and effectively when compared to leading budgeted online algorithms.

Task Allocation and Sequence Planning for Human–Robot Collaborative Disassembly of End-of-Life Products Using the Bees Algorithm

Remanufacturing, which benefits the environment and saves resources, is attracting increasing attention. Disassembly is arguably the most critical step in the remanufacturing of end-of-life (EoL) products. Human–robot collaborative disassembly as a flexible semi-automated approach can increase productivity and relieve people of tedious, laborious, and sometimes hazardous jobs. Task allocation in human–robot collaborative disassembly involves methodically assigning disassembly tasks to human operators or robots. However, the schemes for task allocation in recent studies have not been sufficiently refined and the issue of component placement after disassembly has not been fully addressed in recent studies. This paper presents a method of task allocation and sequence planning for human–robot collaborative disassembly of EoL products. The adopted criteria for human–robot disassembly task allocation are introduced. The disassembly of each component includes dismantling and placing. The performance of a disassembly plan is evaluated according to the time, cost, and utility value. A discrete Bees Algorithm using genetic operators is employed to optimise the generated human–robot collaborative disassembly solutions. The proposed task allocation and sequence planning method is validated in two case studies involving an electric motor and a power battery from an EoL vehicle. The results demonstrate the feasibility of the proposed method for planning and optimising human–robot collaborative disassembly solutions.

End-To-End Latency of Cause-Effect Chains: A Tutorial

In many applications of cyber-physical systems, a sequence of tasks is necessary to perform a certain functionality. For example, from a sensor to an actuator, the first task reads the sensor value (cause), the second task processes the data, and the third task produces an output for the actuator (an effect is triggered). For such scenarios, the end-to-end timing properties (the so-called end-to-end latency) of the sequence of tasks (the so-called cause-effect chain) are of importance. This tutorial recaps different metrics for the end-to-end latency of cause-effect chains, and summarizes fundamental properties and existing analytical results in a systematic manner. To that end, this tutorial has a special focus on the reaction time (how fast can a reaction be in the worst case) and the data age (how old is the data source of an actuation in the worst case). The goal of this tutorial is to provide a systematic view of the fundamental end-to-end timing properties of cause-effect chains and offer an outlook of possible research directions in the near future. Furthermore, we extend the proof of one fundamental property in the literature to comply with the current state-of-the-art definition of end-to-end latencies.

Self vs. Other in Affective Forecasting: The Role of Psychological Distance and Decision from Experience

This work tests self–other differences in the impact bias using the perspectives of psychological distance and decision from description vs. experience. Two studies compared the bias participants made for themselves and for others in a sequential gambling task. The task involved two identical gambles where the first gamble was mandatory and participants made decisions (accept or reject) for the second gamble. Planned decisions were made anticipating a gain or loss in the first gamble, and revised decisions were made following the actual experience of gain or loss. Study 1 compared decisions for self, abstract other, and a close friend. Study 2 replicated the comparison between the self and a close friend and added a measure of empathy. Both studies demonstrated an impact bias indicating that participants tended to overestimate the impact of anticipated outcomes on their tendency towards risk. Specifically, revised decisions indicated risk-aversion shifts after experienced gain and risk-seeking shifts after experienced loss. A reversed pattern emerged for close friends, indicating risky shifts after gain and cautious shifts after loss in Study 1 and for highly empathetic participants in Study 2. Assessing the utility functions that underlie participants’ decisions revealed a qualitative difference. The utility function for the self was consistent with prospect theory (with moderate intensity and diminishing sensitivity), while the utility function for others was more intense with little or no diminishing sensitivity. This research offers new insights regarding the roles of psychological distance and description vs. experience in affective forecasting and impact bias for self vs. other.

Decrease in decision noise from adolescence into adulthood mediates an increase in more sophisticated choice behaviors and performance gain.

Learning and decision-making undergo substantial developmental changes, with adolescence being a particular vulnerable window of opportunity. In adolescents, developmental changes in specific choice behaviors have been observed (e.g., goal-directed behavior, motivational influences over choice). Elevated levels of decision noise, i.e., choosing suboptimal options, were reported consistently in adolescents. However, it remains unknown whether these observations, the development of specific and more sophisticated choice processes and higher decision noise, are independent or related. It is conceivable, but has not yet been investigated, that the development of specific choice processes might be impacted by age-dependent changes in decision noise. To answer this, we examined 93 participants (12 to 42 years) who completed 3 reinforcement learning (RL) tasks: a motivational Go/NoGo task assessing motivational influences over choices, a reversal learning task capturing adaptive decision-making in response to environmental changes, and a sequential choice task measuring goal-directed behavior. This allowed testing of (1) cross-task generalization of computational parameters focusing on decision noise; and (2) assessment of mediation effects of noise on specific choice behaviors. Firstly, we found only noise levels to be strongly correlated across RL tasks. Second, and critically, noise levels mediated age-dependent increases in more sophisticated choice behaviors and performance gain. Our findings provide novel insights into the computational processes underlying developmental changes in decision-making: namely a vital role of seemingly unspecific changes in noise in the specific development of more complex choice components. Studying the neurocomputational mechanisms of how varying levels of noise impact distinct aspects of learning and decision processes may also be key to better understand the developmental onset of psychiatric diseases.

On the Performance of the Minimax Optimal Strategy in the Stochastic Case of Logistic Bandits

The multi-armed bandit problem is a well-established model for examining the exploration/exploitation trade-offs in sequential decision-making tasks. This study focuses on the logistic bandit, where rewards are derived from two distinct datasets of movie ratings, ranging from 1 to 5, each characterized by different variances. Previous research has shown that regret bounds for multi-armed bandit algorithms can be unstable across varying environments. However, this paper provides new insights by demonstrating the robustness of the Minimax Optimal Strategy in the Stochastic (MOSS) algorithm across environments with differing reward variances. Unlike prior studies, this research shows that MOSS maintains superior performance in both dense and sparse reward settings, consistently outperforming widely used algorithms like UCB and TS, particularly in high variance conditions and over sufficient trials. The findings indicate that MOSS achieves logarithmic expected regret for both types of environments, effectively balancing exploration and exploitation. Specifically, with K arms and T time steps, the regret R(T) of MOSS is bounded by O((KT logT )). This work highlights MOSS as a robust solution for handling diverse stochastic conditions, filling a crucial gap in the understanding of its practical adaptability across different reward distributions.

PENERAPAN CRITICAL PATH METHOD DALAM PENJADWALAN PROYEK PEMASANGAN TANGKI STORAGE ST-1812, STUDI KASUS PADA PT. MSM

The installation of the ST-1812 storage tank aims to improve production efficiency at PT. MSM by saving costs in the use of steam from 16 hours to only 4 hours, which is used to heat the final product to temperature of 63°C according to work instructions. Critical Path Method is used to determine the optimal construction time, because this method can identify the sequence of tasks that are most critical or that have the most impact on the duration of the entire project. The purpose of this study is to determine the optimal duration for implementation and completion of the ST-1812 storage tank installation project at PT MSM. The analysis of the CPM application showed that the project took 90 days to complete, with 19 activities included in the critical path. The addition of working hours (overtime) on activities on the critical path makes the project can be completed in 73 days, 17 days faster than the initial duration of the project. The initial cost of the project was IDR 66,555,000, and the cost after accelerating the project with crashing calculations was IDR 74,886,429, with a cost slope of IDR 8,331,429.

Feature Selection as Deep Sequential Generative Learning

Feature selection aims to identify the most pattern-discriminative feature subset. In prior literature, filter (e.g., backward elimination) and embedded (e.g., LASSO) methods have hyperparameters (e.g., top- k , score thresholding) and tie to specific models, thus, hard to generalize; wrapper methods search a feature subset in a huge discrete space and is computationally costly. To transform the way of feature selection, we regard a selected feature subset as a selection decision token sequence and reformulate feature selection as a deep sequential generative learning task that distills feature knowledge and generates decision sequences. Our method includes three steps: (1) We develop a deep variational transformer model over a joint of sequential reconstruction, variational, and performance evaluator losses. Our model can distill feature selection knowledge and learn a continuous embedding space to map feature selection decision sequences into embedding vectors associated with utility scores. (2) We leverage the trained feature subset utility evaluator as a gradient provider to guide the identification of the optimal feature subset embedding; (3) We decode the optimal feature subset embedding to autoregressively generate the best feature selection decision sequence with autostop. Extensive experimental results show this generative perspective is effective and generic, without large discrete search space and expert-specific hyperparameters. The code is available at http://tinyurl.com/FSDSGL .

The Suppression Mechanisms of Passive Memory in Visual Working Memory: The Evidence from Electrophysiology.

Recent studies of visual working memory (VWM) underscore a structured hierarchy of storage states. Memories that are not immediately relevant to the task at hand but are essential for later use are transferred to a passive state, which operates independently of actively maintaining and manipulating current memories. Note that stimulating passive memory forcefully can reactivate it into an active state, resulting in a competition with active memory. Thus, to remain stable representations for both states within VWM, passive memory might involve sustained suppression during activity-silent maintenance to prevent reactivation from disrupting the current active storage. To investigate this, we analyzed lateralized electrophysiology signals while human participants (both women and men) were engaged in a sequential presentation memory task across two experiments. The results revealed positive contralateral delayed activity components and lateralized alpha enhancement for passive memory, neural indicative of suppression on passive storage. In addition, the suppression effect was independent of the memory load in both the active and the passive states. These findings support the notion of sustained suppression during activity-silent maintenance of passive memory, facilitating the stable maintenance of distinct storage states and advancing our understanding of the dynamic coding framework in VWM.

Sequential Task Allocation of More Scalable Artificial Dragonfly Swarms Considering Dubins Trajectory

With the rapid advancement of UAV technology and the increasing complexity of tasks, multi-UAV systems face growing challenges in task execution. Traditional task allocation algorithms often perform poorly when dealing with issues such as local optima, slow convergence speed, and low convergence accuracy, making it difficult to meet the demands for efficiency and practicality in real-world applications. To address these problems, this paper focuses on collaborative task allocation technology for multi-UAV. It proposes a collaborative task allocation strategy for multi-UAV in a multi-target environment, which comprehensively considers various complex constraints in practical application scenarios. The strategy utilizes Dubins curves for trajectory planning and constructs a multi-UAV collaborative task allocation model, with targets including the shortest total distance index, the minimum time index, and the trajectory coordination index. Each UAV is set as an artificial dragonfly by modifying the traditional dragonfly algorithm, incorporating differential evolution algorithms and their crossover, mutation, and selection operations to bring UAV swarms closer to the characteristics of biological dragonflies. The modifications can enhance the global scalability of artificial dragonfly swarms (ADSs), including wider search capacity, wider speed range, and more diverse search accuracy. Meanwhile, potential solutions with global convergence properties are stored to better support real-time adjustments to task allocation. The simulation results show that the proposed strategy can generate a conflict-free task execution scheme and plan the trajectory, which has advantages in changing the data scale of the UAV and the target and improves the reliability of the system to a certain extent.

Distinct functions for beta and alpha bursts in gating of human working memory

Multiple neural mechanisms underlying gating to working memory have been proposed with divergent results obtained in human and animal studies. Previous findings from non-human primates suggest prefrontal beta frequency bursts as a correlate of transient inhibition during selective encoding. Human studies instead suggest a similar role for sensory alpha power fluctuations. To cast light on these discrepancies we employed a sequential working memory task with distractors for human participants. In particular, we examined their whole-brain electrophysiological activity in both alpha and beta bands with the same single-trial burst analysis earlier performed on non-human primates. Our results reconcile earlier findings by demonstrating that both alpha and beta bursts in humans correlate with the filtering and control of memory items, but with region and task-specific differences between the two rhythms. Occipital beta burst patterns were selectively modulated during the transition from sensory processing to memory retention whereas prefrontal and parietal beta bursts tracked sequence order and were proactively upregulated prior to upcoming target encoding. Occipital alpha bursts instead increased during the actual presentation of unwanted sensory stimuli. Source reconstruction additionally suggested the involvement of striatal and thalamic alpha and beta. Thus, specific whole-brain burst patterns correlate with different aspects of working memory control.

Predicting and Understanding Human Action Decisions: Insights from Large Language Models and Cognitive Instance-Based Learning

Large Language Models (LLMs) excel in tasks from translation to complex reasoning. For AI systems to help effectively, understanding and predicting human behavior and biases is essential. However, it remains an open question whether LLMs can achieve this goal. This paper addresses this gap by leveraging the reasoning and generative capabilities of LLMs to predict human behavior in two sequential decision-making tasks. These tasks involve balancing between exploratory and exploitative actions and handling delayed feedback, which is essential for simulating real-life decision processes. We compare the performance of LLMs with a cognitive instance-based learning (IBL) model, which imitates human experiential decision-making. Our findings indicate that LLMs excel at rapidly incorporating feedback to enhance prediction accuracy. In contrast, the IBL model better accounts for human exploratory behaviors and effectively captures loss aversion bias — the tendency to choose a sub-optimal goal with fewer step-cost penalties rather than exploring to find the optimal choice, even with limited experience. The results highlight the benefits of integrating LLMs with cognitive architectures, suggesting that this synergy could enhance the modeling and understanding of complex human decision-making patterns.

An Exploratory Study of the Impact of Task Selection Strategies on Worker Performance in Crowdsourcing Microtasks

In microtask crowdsourcing systems like Amazon Mechanical Turk (AMT) and Appen Figure-Eight, workers often employ task selection strategies, completing sequences of tasks to maximize earnings. While previous literature has explored the effects of sequential tasks with varying complexities of the same type, there is a lack of knowledge on the consequences when multiple types of tasks with similar levels of difficulty are performed. This study examines the impact of sequences of three frequently employed task types, namely image classification, text classification, and surveys, on workers' engagement, accuracy, and perceived workloads. In addition, we analyze the influence of workers' personality traits on their strategies for selecting tasks. Our study, which involved 558 participants using AMT, found that engaging in sequences of distinct task types had a detrimental effect on classification task engagement and accuracy. It also increases the perceived task load and the worker's frustration. Nevertheless, the precise order of tasks does not significantly impact these results. Moreover, we showed a slight association between personality traits and the workers' selection strategy for the tasks. The results offered valuable knowledge for designing an efficient and inclusive crowdsourcing platform.

Long-term motor learning in focal hand dystonia

ObjectiveBecause focal hand dystonia usually occurs in the over-learned stage, it would be valuable to know long-term motor learning characteristics and underlying pathophysiological features that might predispose to dystonia. MethodsWe conducted a case-control exploratory study of 15 visits over 12 weeks in the non-affected hand of a 4-finger sequence of 8 key presses in eight patients with FHD compared with eight age- and sex-matched, healthy volunteers (HVs). We studied the behavioral data and the physiological changes of the brain, including motor cortical excitability and cortical oscillations. ResultsThere was no significant difference in the time to reach 100 % accuracy between FHD and HV during the 80-day follow-up period. There was a statistically significant difference in the accuracy of sequential finger movement tasks between patients with FHD compared with HVs over 12 weeks, but post-hoc analysis with multiple comparion correction did not show difference. There were no significant differences in recruitment curve changes and task-related power changes of alpha and beta bands. ConclusionOver 12 weeks, FHD have motor learning capacity comparable to HVs and do not show pathophysiological abnormalities. SignificanceFurther studies would be valuable with more patients, more extended periods of practice, and more detailed electrophysiological explorations.

Implicit motor sequence learning using three-dimensional reaching movements with the non-dominant left arm.

Interlimb differences in reach control could impact the learning of a motor sequence that requires whole-arm movements. The purpose of this study was to investigate the learning of an implicit, 3-dimensional whole-arm sequence task with the non-dominant left arm compared to the dominant right arm. Thirty-one right-hand dominant adults completed two consecutive days of practice of a motor sequence task presented in a virtual environment with either their dominant right or non-dominant left arm. Targets were presented one-at-a-time alternating between Random and Repeated sequences. Task performance was indicated by the time to complete the sequence (response time), and kinematic measures (hand path distance, peak velocity) were used to examine how movements changed over time. While the Left Arm group was slower than the Right Arm group at baseline, both groups significantly improved response time with practice with the Left Arm group demonstrating greater gains. The Left Arm group improved performance by decreasing hand path distance (straighter path to targets) while the Right Arm group improved performance through a smaller decrease in hand path distance combined with increasing peak velocity. Gains made during practice on Day 1 were retained on Day 2 for both groups. Overall, individuals reaching with the non-dominant left arm learned the whole-arm motor sequence task but did so through a different strategy than individuals reaching with the dominant right arm. The strategy adopted for the learning of movement sequences that require whole-arm movements may be impacted by differences in reach control between the nondominant and dominant arms.

TS‐Mixer: A lightweight text representation model based on context awareness

AbstractLarge pre‐trained models (PTMs) have shown their powerful ability in multiple natural language processing tasks. However, using them in practical application remains a challenge due to the significant computational cost and memory requirements. In order to achieve the balance of computational cost and accuracy, MLP architecture can be used as an alternative to the self‐attention module, such as pNLP‐Mixer and Hyper‐Mixer. Experiments indicate that, MLP‐based models can attain competitive performance with low cost. They maintain the balance of computation cost and accuracy successfully, yet, this is at the expense of not being able to capture short‐range dependencies. In this paper, a novel MLP‐based model, termed TS‐Mixer, is proposed which can capture local dependencies by shifting operation. Compared with other MLP‐based models, the parameters of TS‐Mixer are decoupled from the sequence length, hence it has a smaller model size in long sequence tasks. In addition, TS‐Mixer has linear computational complexity, therefore it can be used as a lightweight alternative to the self‐attention model. Experiments show that the TS‐Mixer outperforms other MLP‐based models, which achieves higher accuracy with fewer parameters in multiple downstream tasks. Notably, compared with pre‐trained models, TS‐Mixer can reach more than 90% of their accuracy with 1% or even one thousandth of the parameters (0.174 ~ 1.2 M). These results demonstrate that TS‐Mixer can achieve a better balance between the computing resources and accuracy. Code is available at: https://github.com/wyl-privacy-project/TS-Mixer.

Understanding the SSARC Model of Task Sequencing: Assessing L2 Writing Development

This study aimed to explore the impact of task sequencing on the development of second language (L2) writing and investigate how L2 learners performed on three decision-making writing tasks completed in different orders over nine weeks. 120 advanced-high EFL students were randomly assigned to one of three groups and given different task sequences: 1) a simple-medium-complex (SMC) sequence, 2) a complex-medium-simple sequence (CMS), or 3) a random sequence (RDM). Essays were analyzed using measures of syntactic complexity, accuracy, lexical complexity, and fluency (CALF). Results showed that the CALF of L2 writing demonstrated longitudinal development over time in all three task sequencing groups. CALF development was not immediately apparent in the first six weeks, with most measures displaying a significant increase by the end of the ninth week. Furthermore, different task sequences resulted in varying patterns and magnitudes of CALF growth, but no specific sequence was found to be superior overall.

A simple method and case study of task planning for collaborative design based on FDSM

In multiperson collaborative mechanical product design, unreasonable task allocation can result in excessive coupling between tasks, complex design communication, and rework, particularly in small enterprises and teams that rely primarily on experience. Hence, this study proposes a simple and efficient task-planning method based on a fuzzy design structure matrix (FDSM), which involves single-level task decomposition, three-level number assignment of the coupling relationship, decoupling of the structural tree algorithm, and validation of the optimization results based on the G-value. Design tasks are promptly allocated into a single level based on the functional structure module of the product. The task-correlation degree is quantified based on three levels of numerical values through personal experience to construct an FDSM, the latter of which is decoupled by the correlation degree of the structural tree algorithm to obtain the optimized sequence and allocation of the collaborative design task. Finally, the effectiveness of the planning is verified by calculating the tightness of the corresponding matrix before and after the planning. A task-planning case of a riveting-machine design for capacitor cover plates is established to validate this method.

Neural mechanisms associated with sleep-dependent enhancement of motor learning after brain injury.

This study aimed to examine the improvement in performance and functional magnetic resonance imaging correlates of motor learning after a daytime nap versus a period of resting wakefulness among individuals with traumatic brain injury. A sample of 32 individuals with traumatic brain injury was randomly assigned to a Sleep (N = 17) or Wake (N = 15) group after a period of training on a motor sequential learning task. A 45-min nap opportunity was provided for the Sleep group, while the Wake group watched a documentary for 45 min. Performance at the end of training was compared with their performance after the nap or wake intervention. Before and after the intervention, the motor sequential learning task was completed in the magnetic resonance imaging scanner to examine the relationship between change in performance and neural activation. Participants in the Sleep group showed significant gains from the end of training to after the intervention, whereas the Wake group did not. The functional magnetic resonance imaging results showed that relative to the Wake group, the Sleep group showed significantly decreased activation post-intervention in the anterior cingulate/paracingulate, cerebellum, cuneus/precuneus, and inferior parietal lobule including angular and supramarginal gyri. Importantly, across both groups, increased task performance at post-intervention was associated with decreased activation in the anterior cingulate/paracingulate and cerebellum. This study demonstrated the enhancing effect of a nap on motor learning performance in a sample of individuals with traumatic brain injury, with patterns of neural activation suggesting that the sequence was more automatized in the Sleep group. Strategic placement of a nap after an intense period of motor learning in the medical rehabilitation setting may have important implications for maximizing recovery after traumatic brain injury.

Unraveling the time-of-day influences on motor consolidation through the motor-declarative memory conflict.

Competition between motor and declarative memory systems, both involved simultaneously in motor learning, has been shown to reduce motor consolidation. Here, we investigated this conflict during the learning of a sequential finger-tapping task (SFTT) scheduled for either the morning or the afternoon. Sixty participants, divided into four groups, trained on SFTT at either 10 a.m. or 3 p.m., and retested five hours later. To disrupt the conflict between the two memories, two groups underwent declarative learning immediately after SFTT training, involving word list training (G10DL and G3DL), while the two other groups (G10CTR and G3CTR) experienced no additional learning. The results revealed that after morning training without additional learning (C10CTR), skill consolidation deteriorated, while the addition of declarative learning (G10DL) significantly attenuated this decay, stabilizing consolidation. Afternoon training showed skill stabilization for both groups (G3CTR and G3DL). These results suggest that weaker consolidation after morning training may be due to an important competition between motor and declarative memories within the same motor task.