Learning Task Research Articles

Fostering Learner Autonomy and Best Approaches to Implement the Practice

As per the definition, “The ability to take charge of one’s own learning”, Learner Autonomy (LA) has been a significant theme of interest in second and foreign language learning over the last five decades. Yet this concept is not widely practiced among local learners in the English as a Second Language (ESL) pedagogy. Implementing autonomous learning behaviours benefits the education system, of a country like Sri Lanka, to minimize most obstacles the ESL learners encounter. For a successful implementation of the concept, the contribution of both teachers and students is essential. Therefore, this research aims to identify sixty ESL teachers’ understanding of how much they support in fostering LA, and the best approaches to implement the practice. Using the simple random sampling method, 10% out of 600 English teachers from the Galle zonal division were given a researcher-made questionnaire that included four sections. A 5-point Likert scale was used to measure the collected data, and they were quantitatively analyzed using descriptive statistics. The SPSS version 21 was used, while Cronbach’s alpha determined the reliability of the Likert scale. The questions were composed of identifying the teachers’ understanding upon the LA concept, the teachers’ roles and responsibilities in planning, implementing, monitoring, and evaluating the class; and how the teachers view their learners’ abilities to take responsibility in planning, implementing, monitoring, and assessing their learning tasks, while suggestions were forwarded to get the teachers’ preferences as the best approaches to foster autonomy among ESL learners. Significantly, 98.3% believe that implementing LA is essential, and the study concluded with the understanding that teachers are willing to cultivate LA if a suitable environment is created.

L’impact de l'utilisation de la reconnaissance vocale dans l'enseignement du FLE à l'oral dans les classes du secondaire en Algérie

The objective of this research is to assess the impact of using speech recognition tools to enhance pronunciation in the teaching of French as a Foreign Language (FLE) during an oral session in a secondary school classroom in Algeria. We conducted an experiment to measure the difference in acquisition, interaction, and motivation among students in two different learning situations, integrating ICT (Information and Communication Technology) and tasks to be performed.

Methamphetamine-induced adaptation of learning rate dynamics depend on baseline performance.

The ability to calibrate learning according to new information is a fundamental component of an organism's ability to adapt to changing conditions. Yet, the exact neural mechanisms guiding dynamic learning rate adjustments remain unclear. Catecholamines appear to play a critical role in adjusting the degree to which we use new information over time, but individuals vary widely in the manner in which they adjust to changes. Here, we studied the effects of a low dose of methamphetamine (MA), and individual differences in these effects, on probabilistic reversal learning dynamics in a within-subject, double-blind, randomized design. Participants first completed a reversal learning task during a drug-free baseline session to provide a measure of baseline performance. Then they completed the task during two sessions, one with MA (20 mg oral) and one with placebo (PL). First, we showed that, relative to PL, MA modulates the ability to dynamically adjust learning from prediction errors. Second, this effect was more pronounced in participants who performed poorly at baseline. These results present novel evidence for the involvement of catecholaminergic transmission on learning flexibility and highlights that baseline performance modulates the effect of the drug.

Humans forage for reward in reinforcement learning tasks.

How do we make good decisions in uncertain environments? In psychology and neuroscience, the classic answer is that we calculate the value of each option and then compare the values to choose the most rewarding, modulo some exploratory noise. An ethologist, conversely, would argue that we commit to one option until its value drops below a threshold, at which point we start exploring other options. In order to determine which view better describes human decision-making, we developed a novel, foraging-inspired sequential decision-making model and used it to ask whether humans compare to threshold ("Forage") or compare alternatives ("Reinforcement-Learn" [RL]). We found that the foraging model was a better fit for participant behavior, better predicted the participants' tendency to repeat choices, and predicted the existence of held-out participants with a pattern of choice that was almost impossible under RL. Together, these results suggest that humans use foraging computations, rather than RL, even in classic reinforcement learning tasks.

Training kindergarten children on learning from their mistakes.

This study investigated whether feedback on their errors and speed improves kindergarten children's performance in an executive function (EF) task. Children from Switzerland (N = 213, 49% female, Mage = 6.4 years) were tested in the Hearts and Flowers task pre- and post-training and trained either on a variant of this task with (n = 71) or without feedback (n = 72), or on a control learning task (n = 70). The feedback group performed more efficiently than the no-feedback group during the intervention and partially also in the post-test. Both EF training groups performed more efficiently than the control group in the post-test. These results suggest that kindergarten children detect and monitor their errors and even get better at it given the opportunity to practice. Moreover, they benefit additionally from external feedback. Integrating feedback into computerized cognitive training (and learning apps) could be a potential avenue for interventions in school settings.

MICCF: A Mutual Information Constrained Clustering Framework for Learning Clustering-Oriented Feature Representations

Deep clustering is a crucial task in machine learning and data mining that focuses on acquiring feature representations conducive to clustering. Previous research relies on self-supervised representation learning for general feature representations, such features may not be optimally suited for downstream clustering tasks. In this paper, we introduce MICCF, a framework designed to bridge this gap and enhance clustering performance. MICCF enhances feature representations by combining mutual information at different levels and employs an auxiliary alignment mutual information module for learning clustering-oriented features. To be specific, we propose a dual mutual information constraints module, incorporating minimal mutual information constraints at the feature level and maximal mutual information constraints at the instance level. This reduction in feature redundancy encourages the neural network to extract more discriminative features, while maximization ensures more unbiased and robust representations. To obtain clustering-oriented representations, the auxiliary alignment mutual information module utilizes pseudo-labels to maximize mutual information through a multi-classifier network, aligning features with the clustering task. The main network and the auxiliary one work in synergy to jointly optimize feature representations that are well-suited for the clustering task. We validate the effectiveness of our method through extensive experiments on six benchmark datasets. The results indicate that our method performs well in most scenarios, particularly on fine-grained datasets, where our approach effectively distinguishes subtle differences between closely related categories. Notably, our approach achieved a remarkable accuracy of 96.4% on the ImageNet-10 dataset, surpassing other comparison methods. The code is available at https://github.com/Li-Hyn/MICCF.git .

Modulating social learning-induced evaluation updating during human sleep

People often change their evaluations upon learning about their peers’ evaluations, i.e., social learning. Given sleep’s vital role in consolidating daytime experiences, sleep may facilitate social learning, thereby further changing people’s evaluations. Combining a social learning task and the sleep-based targeted memory reactivation technique, we asked whether social learning-induced evaluation updating can be modulated during sleep. After participants had indicated their initial evaluation of snacks, they learned about their peers’ evaluations while hearing the snacks’ spoken names. During the post-learning non-rapid-eye-movement sleep, we re-played half of the snack names (i.e., cued snack) to reactivate the associated peers’ evaluations. Upon waking up, we found that the social learning-induced evaluation updating further enlarged for both cued and uncued snacks. Examining sleep electroencephalogram (EEG) activity revealed that cue-elicited delta-theta EEG power and the overnight N2 sleep spindle density predicted post-sleep evaluation updating for cued but not for uncued snacks. These findings underscore the role of sleep-mediated memory reactivation and the associated neural activity in supporting social learning-induced evaluation updating.

Modulating social learning-induced evaluation updating during human sleep

People often change their evaluations upon learning about their peers’ evaluations, i.e., social learning. Given sleep’s vital role in consolidating daytime experiences, sleep may facilitate social learning, thereby further changing people’s evaluations. Combining a social learning task and the sleep-based targeted memory reactivation technique, we asked whether social learning-induced evaluation updating can be modulated during sleep. After participants had indicated their initial evaluation of snacks, they learned about their peers’ evaluations while hearing the snacks’ spoken names. During the post-learning non-rapid-eye-movement sleep, we re-played half of the snack names (i.e., cued snack) to reactivate the associated peers’ evaluations. Upon waking up, we found that the social learning-induced evaluation updating further enlarged for both cued and uncued snacks. Examining sleep electroencephalogram (EEG) activity revealed that cue-elicited delta-theta EEG power and the overnight N2 sleep spindle density predicted post-sleep evaluation updating for cued but not for uncued snacks. These findings underscore the role of sleep-mediated memory reactivation and the associated neural activity in supporting social learning-induced evaluation updating.

Effects of psychedelic, DOI, on nucleus accumbens dopamine signaling to predictable rewards and cues in rats.

Psychedelics produce lasting therapeutic responses in neuropsychiatric diseases suggesting they may disrupt entrenched associations and catalyze learning. Here, we examine psychedelic 5-HT2A/2C agonist, DOI, effects on dopamine signaling in the nucleus accumbens (NAc) core, a region extensively linked to reward learning, motivation, and drug-seeking. We measure phasic dopamine transients following acute DOI administration in rats during well learned Pavlovian tasks in which sequential cues predict rewards. We find that DOI (0.0-1.2 mg/kg, i.p.) increases dopamine signals, photometrically measured using GRABDA optical sensor, to rewards and proximal reward cues, but not to the distal cues that predict these events. We determine that the elevated dopamine produced by DOI to reward cues occurs independently of DOI-induced changes in reward value. The increased dopamine associated with predictable reward cues and rewards supports DOI-induced increases in prediction error signaling. These findings lay a foundation for developing psychedelic strategies aimed at engaging error-driven learning mechanisms to disrupt entrenched associations or produce new associations.

DyHANE: dynamic heterogeneous attributed network embedding through experience node replay

With real-world network systems typically comprising a large number of interactive components and inherently dynamic, Graph Continual Learning (GCL) has gained increasing popularity in recent years. Furthermore, most applications involve multiple entities and relationships with associated attributes, which has led to widely adopting Heterogeneous Information Networks (HINs) for capturing such rich structural and semantic meaning. In this context, we deal with the problem of learning multi-type node representations in a time evolving graph setting, harnessing the expressive power of Graph Neural Networks (GNNs). To this purpose, we propose a novel framework, named DyHANE—Dynamic Heterogeneous Attributed Network Embedding, which dynamically identifies a representative sample of multi-typed nodes as training set and updates the parameters of a GNN module, enabling the generation of up-to-date representations for all nodes in the network. We show the advantage of employing HINs on a data-incremental classification task. We compare the results obtained by DyHANE on a multi-step, incremental heterogeneous GAT model trained on a sample of changed and unchanged nodes, with the results obtained by either the same model trained from scratch or the same model trained solely on changed nodes. We demonstrate the effectiveness of the proposed approach in facing two major related challenges: (i) to avoid model re-train from scratch if only a subset of the network has been changed and (ii) to mitigate the risk of losing established patterns if the new nodes exhibit unseen properties. To the best of our knowledge, this is the first work that deals with the task of (deep) graph continual learning on HINs.

A Lightweight Task-Agreement Meta Learning for Low-Resource Speech Recognition

Meta-learning has proven to be a powerful paradigm for transferring knowledge from prior tasks to facilitate the quick learning of new tasks in automatic speech recognition. However, the differences between languages (tasks) lead to variations in task learning directions, causing the harmful competition for model’s limited resources. To address this challenge, we introduce the task-agreement multilingual meta-learning (TAMML), which adopts the gradient agreement algorithm to guide the model parameters towards a direction where tasks exhibit greater consistency. However, the computation and storage cost of TAMML grows dramatically with model’s depth increases. To address this, we further propose a simplification called TAMML-Light which only uses the output layer for gradient calculation. Experiments on three datasets demonstrate that TAMML and TAMML-Light achieve outperform meta-learning approaches, yielding superior results.Furthermore, TAMML-Light can reduce at least 80 %\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\%$$\\end{document} of the relative increased computation expenses compared to TAMML.

Resampling the optical transmission matrix to accelerate the training of the diffractive deep neural network

The diffractive deep neural network (D2NN) enables all-optical implementation of machine learning tasks. During the training, the Rayleigh–Sommerfeld (RS) diffraction integral is employed for connecting neurons between neighboring hidden layers. The RS formula can be rewritten as a transmission matrix (TM), which allows for the parallel computation of multiple vectorized light fields. However, the TM has a large size, demanding substantial computational resources, and resulting in long training time. In this paper, we propose to resample the TM in free space based on the propagation invariant modes (PIMs), thereby reducing the size of the matrix, and accelerating the propagation simulations. This method enables the training of the large-scale D2NN with reduced memory requirements and fast speed.

One-Shot Any-Scene Crowd Counting With Local-to-Global Guidance.

Due to different installation angles, heights, and positions of the camera installation in real-world scenes, it is difficult for crowd counting models to work in unseen surveillance scenes. In this paper, we are interested in accurate crowd counting based on the data collected by any surveillance camera, that is to count the crowd from any scene given only one annotated image from that scene. To this end, we firstly pose crowd counting as a one-shot learning task. Through the metric-learning, we propose a simple yet effective method that firstly estimates crowd characteristics and then transfers them to guide the model to count the crowd. Specifically, to fully capture these crowd characteristics of the target scene, we devise the Multi-Prototype Learner to learn the prototypes of foreground and density from the limited support image using the Expectation-Maximization algorithm. To learn the adaptation capability for any unseen scene, estimated multi prototypes are proposed to guide the crowd counting of query images in a local-to-global way. CNN is utilized to activate the local features. And transformer is introduced to correlate global features. Extensive experiments on three surveillance datasets suggest that our method outperforms the SOTA methods in the few-shot crowd counting.

Jumping to attributions during social evaluation

Social learning is crucial for human relationships and well-being. Self- and other- evaluations are universal experiences, playing key roles in many psychiatric disorders, particularly anxiety and depression. We aimed to deepen our understanding of the computational mechanisms behind social learning, which have been implicated in internalizing conditions like anxiety and depression. We built on prior work based on the Social Evaluation Learning Task (SELT) and introduced a new computational model to better explain rapid initial inferences and progressive refinement during serial social evaluations. The Social Evaluation Learning Task-Revised (SELT-R) was improved by stakeholder input, making it more engaging and suitable for adolescents. A sample of 130 adults from the UK completed the SELT-R and questionnaires assessing symptoms of depression and anxiety. ‘Classify-refine’ computational models were compared with previously successful Bayesian models. The ‘classify-refine’ models performed better, providing insight into how people infer the attributes and motives of others. Parameters of the best fitting model from the SELT-R were correlated with Anxiety factor scores, with higher symptoms associated with greater decision noise and higher (less flexible) policy certainty. Our results replicate findings regarding the classify-refine process and set the stage for future investigations into the cognitive mechanisms of self and other evaluations in internalizing disorders.

On the practical usefulness of the Hardware Efficient Ansatz

Variational Quantum Algorithms (VQAs) and Quantum Machine Learning (QML) models train a parametrized quantum circuit to solve a given learning task. The success of these algorithms greatly hinges on appropriately choosing an ansatz for the quantum circuit. Perhaps one of the most famous ansatzes is the one-dimensional layered Hardware Efficient Ansatz (HEA), which seeks to minimize the effect of hardware noise by using native gates and connectives. The use of this HEA has generated a certain ambivalence arising from the fact that while it suffers from barren plateaus at long depths, it can also avoid them at shallow ones. In this work, we attempt to determine whether one should, or should not, use a HEA. We rigorously identify scenarios where shallow HEAs should likely be avoided (e.g., VQA or QML tasks with data satisfying a volume law of entanglement). More importantly, we identify a Goldilocks scenario where shallow HEAs could achieve a quantum speedup: QML tasks with data satisfying an area law of entanglement. We provide examples for such scenario (such as Gaussian diagonal ensemble random Hamiltonian discrimination), and we show that in these cases a shallow HEA is always trainable and that there exists an anti-concentration of loss function values. Our work highlights the crucial role that input states play in the trainability of a parametrized quantum circuit, a phenomenon that is verified in our numerics.

Serotonergic input from the dorsal raphe nucleus shapes learning-associated odor responses in the olfactory bulb.

Neural activity in the olfactory bulb (OB) can represent odor information during different brain and behavioral states. For example, the odor responses of mitral/tufted (M/T) cells in the OB change during learning of odor-discrimination tasks and, at the network level, beta power increases and the high gamma (HG) power decreases during odor presentation in such tasks. However, the neural mechanisms underlying these observations remain poorly understood. Here, we investigate whether serotonergic modulation from the dorsal raphe nucleus (DRN) to the OB is involved in shaping activity during the learning process in a go/no-go task in mice. Fiber photometry was used to record the population activity of DRN serotonergic neurons during a go/no-go task. Invivo electrophysiology was used to record neural activity (single units and local field potentials) in the OB during the go/no-go task. Real-time place preference (RTPP) and intracranial light administration in a specific subarea (iClass) tests were used to assess the ability of mice to encoding reward information. Odor-evoked population activity in serotonergic neurons in the DRN was shaped during the learning process in a go/no-go task. In the OB, neural activity from oscillations to single cells showed complex, learning-associated changes and ability to encode information during an odor discrimination task. However, these properties were not observed after ablation of DRN serotonergic neurons. The activity of neural networks and single cells in the OB, and their ability to encode information about odor value, are shaped by serotonergic projections from the DRN.

Hyperspectral in-memory computing with optical frequency combs and programmable optical memories

The rapid rise of machine learning drives demand for extensive matrix-vector multiplication operations, thereby challenging the capacities of traditional von Neumann computing systems. Researchers explore alternatives, such as in-memory computing architecture, to find energy-efficient solutions. In particular, there is renewed interest in optical computing systems, which could potentially handle matrix-vector multiplication in a more energy-efficient way. Despite promising initial results, developing high-throughput optical computing systems to rival electronic hardware remains a challenge. Here, we propose and demonstrate a hyperspectral in-memory computing architecture, which simultaneously utilizes space and frequency multiplexing, using optical frequency combs and programmable optical memories. Our carefully designed three-dimensional opto-electronic computing system offers remarkable parallelism, programmability, and scalability, overcoming typical limitations of optical computing. We have experimentally demonstrated highly parallel, single-shot multiply-accumulate operations with precision exceeding 4 bits in both matrix-vector and matrix-matrix multiplications, suggesting the system’s potential for a wide variety of deep learning and optimization tasks. Our approach presents a realistic pathway to scale beyond peta operations per second, a major stride towards high-throughput, energy-efficient optical computing.

Functional states of prelimbic and related circuits during the acquisition of a GO/noGO task in rats.

GO/noGO tasks enable assessing decision-making processes and the ability to suppress a specific action according to the context. Here, rats had to discriminate between 2 visual stimuli (GO or noGO) shown on an iPad screen. The execution (for GO) or nonexecution (for noGO) of the selected action (to touch or not the visual display) were reinforced with food. The main goal was to record and to analyze local field potentials collected from cortical and subcortical structures when the visual stimuli were shown on the touch screen and during the subsequent activities. Rats were implanted with recording electrodes in the prelimbic cortex, primary motor cortex, nucleus accumbens septi, basolateral amygdala, dorsolateral and dorsomedial striatum, hippocampal CA1, and mediodorsal thalamic nucleus. Spectral analyses of the collected data demonstrate that the prelimbic cortex was selectively involved in the cognitive and motivational processing of the learning task but not in the execution of reward-directed behaviors. In addition, the other recorded structures presented specific tendencies to be involved in these 2 types of brain activity in response to the presentation of GO or noGO stimuli. Spectral analyses, spectrograms, and coherence between the recorded brain areas indicate their specific involvement in GO vs. noGO tasks.

Fusing crops representation into snippet via mutual learning for weakly supervised surveillance anomaly detection

AbstractIn recent years, the challenge of detecting anomalies in real‐world surveillance videos using weakly supervised data has emerged. Traditional methods, utilising multi‐instance learning (MIL) with video snippets, struggle with background noise and tend to overlook subtle anomalies. To tackle this, the authors propose a novel approach that crops snippets to create multiple instances with less noise, separately evaluates them and then fuses these evaluations for more precise anomaly detection. This method, however, leads to higher computational demands, especially during inference. Addressing this, our solution employs mutual learning to guide snippet feature training using these low‐noise crops. The authors integrate multiple instance learning (MIL) for the primary task with snippets as inputs and multiple‐multiple instance learning (MMIL) for an auxiliary task with crops during training. The authors’ approach ensures consistent multi‐instance results in both tasks and incorporates a temporal activation mutual learning module (TAML) for aligning temporal anomaly activations between snippets and crops, improving the overall quality of snippet representations. Additionally, a snippet feature discrimination enhancement module (SFDE) refines the snippet features further. Tested across various datasets, the authors’ method shows remarkable performance, notably achieving a frame‐level AUC of 85.78% on the UCF‐Crime dataset, while reducing computational costs.

Machine learning without a processor: Emergent learning in a nonlinear analog network

Standard deep learning algorithms require differentiating large nonlinear networks, a process that is slow and power-hungry. Electronic contrastive local learning networks (CLLNs) offer potentially fast, efficient, and fault-tolerant hardware for analog machine learning, but existing implementations are linear, severely limiting their capabilities. These systems differ significantly from artificial neural networks as well as the brain, so the feasibility and utility of incorporating nonlinear elements have not been explored. Here, we introduce a nonlinear CLLN-an analog electronic network made of self-adjusting nonlinear resistive elements based on transistors. We demonstrate that the system learns tasks unachievable in linear systems, including XOR (exclusive or) and nonlinear regression, without a computer. We find our decentralized system reduces modes of training error in order (mean, slope, curvature), similar to spectral bias in artificial neural networks. The circuitry is robust to damage, retrainable in seconds, and performs learned tasks in microseconds while dissipating only picojoules of energy across each transistor. This suggests enormous potential for fast, low-power computing in edge systems like sensors, robotic controllers, and medical devices, as well as manufacturability at scale for performing and studying emergent learning.