Learning Of New Tasks Research Articles

The effects of task novelty for age cohort and cognition level on memory for everyday virtual meal preparation tasks

Nonsensical information increases task novelty, which makes it difficult to rely on previous learning and provides insight into the learning of new tasks. This study investigated procedural-based action scripts in everyday memory for meal preparation tasks in virtual reality. The sample (N = 171) consisted of 3 groups determined by age and cognitive function: young adults (YA; n = 61), older adults with normal cognition (OA; n = 82), and older adults with impaired cognition (IC; n = 28). The three groups completed the Virtual Kitchen Protocol, a virtual reality-based measure of learning and memory for cooking both familiar and nonsensical meals. Results showed that YAs had a greater recall for both familiar and nonsensical meals than OAs or ICs. Additionally, novel stimuli (i.e., nonsensical meal tasks) appear to impact older adults more than young adults. Among older adults, impaired cognition was associated with lower performance on both the sensical and nonsensical meals compared to normal cognition. All three groups performed better on familiar tasks than nonsensical tasks. These results were consistent with the notion that familiarity may be of greater use than novelty in the context of procedural-based action scripts.

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.

Neural Manifold Modulated Continual Reinforcement Learning for Musculoskeletal Robots

The continual learning and development are significant for robots to learn multiple tasks sequentially. The difficulty lies in balancing the efficient learning of new tasks and overcoming catastrophic forgetting of old tasks. Although many continual learning methods have been proposed for pattern recognition, continual reinforcement learning methods for redundant musculoskeletal and robotic systems are few and have limitations. Therefore, inspired by the developmental mechanisms in motor cortex, this article proposes a neural manifold modulated continual reinforcement learning method for musculoskeletal and robotic systems. First, a recurrent neural network (RNN) with expected neural manifold is designed and conditions of weights are derived. Second, the ability of projectors for characterizing the neural manifold within RNN is analyzed. Furthermore, a continual reinforcement learning method of RNN is proposed with the modulation of neural manifold. The method is validated by redundant musculoskeletal and robotic systems in simulation. The results suggest that it can realize continual reinforcement learning of multiple tasks in different movements and environments. Furthermore, compared with related works, the proposed method achieves better performance.

Continual Learning With Knowledge Distillation: A Survey.

The foremost challenge in continual learning is to mitigate catastrophic forgetting, allowing a model to retain knowledge of previous tasks while learning new tasks. Knowledge distillation (KD), a form of regularization, has gained significant attention for its ability to maintain a model's performance on previous tasks by mimicking the outputs of earlier models during the learning of new tasks, thus reducing forgetting. This article offers a comprehensive survey of continual learning methods employing KD within the realm of image classification. We provide a detailed analysis of how KD is utilized in continual learning methods, categorizing its application into three distinct paradigms. Besides, we classify these methods based on the type of knowledge source used and thoroughly examine how KD consolidates memory in continual learning from the perspective of loss functions. In addition, we have conducted extensive experiments on CIFAR-100, TinyImageNet, and ImageNet-100 across ten KD-integrated continual learning methods to analyze the role of KD in continual learning, and we have further discussed its effectiveness in other continual learning tasks. Our extensive experimental evidence demonstrates that KD plays a crucial role in mitigating forgetting in continual learning and substantiates that, when used with data replay, classification bias adversely affects the effectiveness of KD, whereas employing a separated softmax loss can significantly enhance its efficacy.

Few-shot classification via efficient meta-learning with hybrid optimization

Meta-learning is one of the important methods to solve the challenging few-shot learning setting by using previous knowledge and experience to guide the learning of new tasks. Model-agnostic meta-learning (MAML) is one of the most popular meta-learning algorithms, and many variants of MAML have appeared in recent years. However, the performance of this algorithm for few-shot classification falls behind some other algorithms working on this problem. Therefore, its generalization performance needs to be further explored and improved. In view of the generalization problem, we found that MAML always shares an initialization in the process of parameter update, ignoring the bias between different tasks, resulting in limited generalization performance. On the other hand, the sample diversity of meta-learning model is low, and shallow network training is generally used, so it is difficult to obtain good performance based on deep neural network models. Based on these problems, we propose a hybrid optimization meta-learning method based on data augmentation, initialization attenuation, and resolution increase, called Mix-MAML. Experimental results show that our method reaches 76.93% classification accuracy on mini-ImageNet with 100 × 100 resolution, and 83.62% classification accuracy on CIFAR-FS with 80 × 80 resolution in the 5-way 5-shot settings under ResNet12, which achieves comparable or even better performance than other algorithms in some standard few-shot learning benchmarks without changing MAML simplicity and model-agnostic.

Clustered Task-Aware Meta-Learning by Learning From Learning Paths.

To enable effective learning of new tasks with only a few examples, meta-learning acquires common knowledge from the existing tasks with a globally shared meta-learner. To further address the problem of task heterogeneity, recent developments balance between customization and generalization by incorporating task clustering to generate task-aware modulation to be applied to the global meta-learner. However, these methods learn task representation mostly from the features ofinput data, while the task-specific optimization process with respect to the base-learner is often neglected. In this work, we propose a Clustered Task-Aware Meta-Learning (CTML) framework with task representation learned from both features and learning paths. We first conduct rehearsed task learning from the common initialization, and collect a set of geometric quantities that adequately describes this learning path. By inputting this set of values into a meta path learner, we automatically abstract path representation optimized for downstream clustering and modulation. Aggregating the path and feature representations results in an improved task representation. To further improve inference efficiency, we devise a shortcut tunnel to bypass the rehearsed learning process at a meta-testing time. Extensive experiments on two real-world application domains: few-shot image classification and cold-start recommendation demonstrate the superiority of CTML compared to state-of-the-art methods. We provide our code at https://github.com/didiya0825.

Few-shot and meta-learning methods for image understanding: a survey

State-of-the-art deep learning systems (e.g., ImageNet image classification) typically require very large training sets to achieve high accuracies. Therefore, one of the grand challenges is called few-shot learning where only a few training samples are required for good performance. In this survey, we illuminate one of the key paradigms in few-shot learning called meta-learning. These meta-learning methods, by simulating the tasks which will be presented at inference through episodic training, can effectively employ previous prior knowledge to guide the learning of new tasks. In this paper, we provide a comprehensive overview and key insights into the meta-learning approaches and categorize them into three branches according to their technical characteristics, namely metric-based, model-based and optimization-based meta-learning. Due to the major importance of the evaluation process, we also present an overview of current widely used benchmarks, as well as performances of recent meta-learning methods on these datasets. Based on over 200 papers in this survey, we conclude with the major challenges and future directions of few-shot learning and meta-learning.

Combining multitask and transfer learning with deep Gaussian processes for autotuning-based performance engineering

We combine deep Gaussian processes (DGPs) with multitask and transfer learning for the performance modeling and optimization of HPC applications. Deep Gaussian processes merge the uncertainty quantification advantage of Gaussian processes (GPs) with the predictive power of deep learning. Multitask and transfer learning allow for improved learning efficiency when several similar tasks are to be learned simultaneously and when previous learned models are sought to help in the learning of new tasks, respectively. A comparison with state-of-the-art autotuners shows the advantage of our approach on two application problems. In this article, we combine DGPs with multitask and transfer learning to allow for both an improved tuning of an application parameters on problems of interest but also the prediction of parameters on any potential problem the application might encounter.

Prototypical Quadruplet for Few-Shot Class Incremental Learning

The scarcity of data and incremental learning of new tasks pose two major bottlenecks for many modern computer vision algorithms. The phenomenon of catastrophic forgetting, i.e., the model's inability to classify previously learned data after training with new batches of data, is a major challenge. Conventional methods address catastrophic forgetting while compromising the current session's training. Generative replay-based approaches, such as generative adversarial networks (GANs), have been proposed to mitigate catastrophic forgetting, but training GANs with few samples may lead to instability. To address these challenges, we propose a novel method that improves classification robustness by identifying a better embedding space using an improved contrasting loss. Our approach retains previously acquired knowledge in the embedding space, even when trained with new classes, by updating previous session class prototypes to represent the true class mean, which is crucial for our nearest class mean classification strategy. We demonstrate the effectiveness of our method by showing that the embedding space remains intact after training the model with new classes and outperforms existing state-of-the-art algorithms in terms of accuracy across different sessions.

Experts Collaboration Learning for Continual Multi-Modal Reasoning.

Multi-modal reasoning, which aims to capture logical and causal structures in visual content and associate them with cues from other modality inputs (e.g., texts) to perform various types of reasoning, is an important research topic in artificial intelligence (AI). Existing works for multi-modal reasoning mainly exploit offline learning, where the training samples of all types of reasoning tasks are assumed to be available at once. Here we focus on continual learning for multi-modal reasoning (i.e., continual multi-modal reasoning), where the model is required to continuously learn to solve novel types of multi-modal reasoning tasks in a lifelong fashion. Continual multi-modal reasoning is challenging since the model needs to be able to effectively learn various types of new reasoning tasks, meanwhile avoiding forgetting. Here we propose a novel brain-inspired exp erts [Formula: see text]llaboration network (Expo), which incorporates multiple learning blocks (experts). When encountering a new task, our network dynamically assembles and updates a set of task-specific experts that are most relevant to learning the current task, by either utilizing learned experts or exploring new experts. This thus enables effective learning of new tasks, and meanwhile consolidates previously learned reasoning skills. Moreover, to automatically find optimal task-specific experts, an effective experts selection strategy is designed. Extensive experiments demonstrate the efficacy of our model for continual multi-modal reasoning.

Few-Shot Image Classification: Current Status and Research Trends

Conventional image classification methods usually require a large number of training samples for the training model. However, in practical scenarios, the amount of available sample data is often insufficient, which easily leads to overfitting in network construction. Few-shot learning provides an effective solution to this problem and has been a hot research topic. This paper provides an intensive survey on the state-of-the-art techniques in image classification based on few-shot learning. According to the different deep learning mechanisms, the existing algorithms are divided into four categories: transfer learning based, meta-learning based, data augmentation based, and multimodal based methods. Transfer learning based methods transfer useful prior knowledge from the source domain to the target domain. Meta-learning based methods employ past prior knowledge to guide the learning of new tasks. Data augmentation based methods expand the amount of sample data with auxiliary information. Multimodal based methods use the information of the auxiliary modal to facilitate the implementation of image classification tasks. This paper also summarizes the few-shot image datasets available in the literature, and experimental results tested by some representative algorithms are provided to compare their performance and analyze their pros and cons. In addition, the application of existing research outcomes on few-shot image classification in different practical fields are discussed. Finally, a few future research directions are identified.

Meta-learning approaches for learning-to-learn in deep learning: A survey

Compared to traditional machine learning, deep learning can learn deeper abstract data representation and understand scattered data properties. It has gained considerable attention for its extraordinary performances. However, existing deep learning algorithms perform poorly on new tasks. Meta-learning, known as learning to learn, is one of the effective techniques to overcome this issue. Meta-learning’s generalization ability to unknown tasks is improved by employing prior knowledge to assist the learning of new tasks. There are mainly three types of meta-learning methods: metric-based, model-based, and optimization-based meta-learning. We investigate classical algorithms and recent meta-learning advances. Second, we survey meta-learning application in real world scenarios. Finally, we discuss present challenges and future research directions of meta-learning.

Neuroadaptive Training via fNIRS in Flight Simulators.

Training to master a new skill often takes a lot of time, effort, and financial resources, particularly when the desired skill is complex, time sensitive, or high pressure where lives may be at risk. Professions such as aircraft pilots, surgeons, and other mission-critical operators that fall under this umbrella require extensive domain-specific dedicated training to enable learners to meet real-world demands. In this study, we describe a novel neuroadaptive training protocol to enhance learning speed and efficiency using a neuroimaging-based cognitive workload measurement system in a flight simulator. We used functional near-infrared spectroscopy (fNIRS), which is a wearable, mobile, non-invasive neuroimaging modality that can capture localized hemodynamic response and has been used extensively to monitor the anterior prefrontal cortex to estimate cognitive workload. The training protocol included four sessions over 2 weeks and utilized realistic piloting tasks with up to nine levels of difficulty. Learners started at the lowest level and their progress adapted based on either behavioral performance and fNIRS measures combined (neuroadaptive) or performance measures alone (control). Participants in the neuroadaptive group were found to have significantly more efficient training, reaching higher levels of difficulty or significantly improved performance depending on the task, and showing consistent patterns of hemodynamic-derived workload in the dorsolateral prefrontal cortex. The results of this study suggest that a neuroadaptive personalized training protocol using non-invasive neuroimaging is able to enhance learning of new tasks. Finally, we outline here potential avenues for further optimization of this fNIRS based neuroadaptive training approach. As fNIRS mobile neuroimaging is becoming more practical and accessible, the approaches developed here can be applied in the real world in scale.

Role of serotonin in modulating the development and function of adult-born neurons in the olfactory bulb.

Role of serotonin in modulating the development and function of adult-born neurons in the olfactory bulb.

Application of Meta Learning in Face Recognition

Now is the information age, artificial intelligence rises under the background of this era, but in the field of artificial intelligence, deep learning stands out in machine learning with its excellent learning ability and is rapidly applied to face recognition engineering. However, the success of mainstream deep learning often depends on a lot of training data and training time. This paper proposes to use meta-learning technology to realize face recognition. metalearning is to use the prior knowledge and experience to guide the learning of new tasks and has the ability to learn so that it can avoid the phenomenon of overfitting in the case of only a small number of samples. We regard the data set as a task, that is, meta-train task and meta-test task, using the MAML algorithm [1] fine-tuning gradient model based on prior knowledge to obtain optimization parameters to realize face recognition. Experiments show that deep learning neural networks using meta-learning technology can achieve higher accuracy and rate than ordinary face recognition neural networks. In the experiment, the recognition rate of the face can reach more than 92.6%. It has more intelligence and has made a contribution to the development of face recognition technology.

Tacit knowledge transfer and spillover learning in ramp-ups

Purpose With shortening product life cycles and an increasing number of product variants, manufacturing firms perform more production ramp-ups. In this context, learning is crucially important to quickly achieve high production process quality and stability. The paper aims to discuss this issue. Design/methodology/approach Through a laboratory experiment, this study analyzes spillover learning between consecutive ramp-ups and how this phenomenon is influenced by tacit knowledge transfer through observation and imitation. Findings The results prove the existence of spillover learning between consecutive ramp-ups. Moreover, they provide evidence how tacit knowledge transfer through observation and imitation enhances learning of new tasks in consecutive production ramp-ups. Research limitations/implications Future research could focus on the specific psychological processes driving tacit knowledge transfer and spillover learning, a topic which is only touched upon in this paper. Practical implications The findings show that manufacturing firms should not only aim at reaching a steep learning curve during a single production ramp-up, but should also take into account the effects of spillover learning with regard to future production ramp-ups. Furthermore, the paper provides novel insights concerning the allocation of workers to production tasks with regard to previous experience when introducing new personnel and during ramp-up phases. Originality/value Previous evidence on the existence and characteristics of spillover learning in production ramp-up situations is not conclusive. This paper provides new and unambiguous insights by considering different organizational settings.

Uncertainty-based modulation for lifelong learning

The creation of machine learning algorithms for intelligent agents capable of continuous, lifelong learning is a critical objective for algorithms being deployed on real-life systems in dynamic environments. Here we present an algorithm inspired by neuromodulatory mechanisms in the human brain that integrates and expands upon Stephen Grossberg’s ground-breaking Adaptive Resonance Theory proposals. Specifically, it builds on the concept of uncertainty, and employs a series of “neuromodulatory” mechanisms to enable continuous learning, including self-supervised and one-shot learning. Algorithm components were evaluated in a series of benchmark experiments that demonstrate stable learning without catastrophic forgetting. We also demonstrate the critical role of developing these systems in a closed-loop manner where the environment and the agent’s behaviors constrain and guide the learning process. To this end, we integrated the algorithm into an embodied simulated drone agent. The experiments show that the algorithm is capable of continuous learning of new tasks and under changed conditions with high classification accuracy (>94%) in a virtual environment, without catastrophic forgetting. The algorithm accepts high dimensional inputs from any state-of-the-art detection and feature extraction algorithms, making it a flexible addition to existing systems. We also describe future development efforts focused on imbuing the algorithm with mechanisms to seek out new knowledge as well as employ a broader range of neuromodulatory processes.

A Cognitive Load Theory Simulation Design to Assess and Manage Deteriorating Patients.

Background Cognitive load theory (CLT) uses an understanding of brain architecture for educational design, with implications for simulation. Since working memory is limited, minimising extraneous cognitive load improves learning of new tasks (intrinsic load) and use of existing knowledge (germane load). This study evaluates the effectiveness of low-fidelity simulation (LFS) utilising CLT principles in the assessment and management of the deteriorating patient (AMDP). Method CLT design principles informed the choice of LFS and simulation design. The self-rated ability of 13 undergraduate nurses across seven aspects of AMDP was measured in a pre-post design. Results Self-rated ability increased from 2.98 (SD = 0.19) to 4.47 (SD = 0.12) (p < 0.001), with improvements across all AMDP aspects. Conclusion This study indicates that CLT informed design has benefits for simulation. LFS may be preferred to high fidelity simulation for AMDP teaching and medical simulation for novice learners.

Self-organizing maps for storage and transfer of knowledge in reinforcement learning

The idea of reusing or transferring information from previously learned tasks (source tasks) for the learning of new tasks (target tasks) has the potential to significantly improve the sample efficiency of a reinforcement learning agent. In this work, we describe a novel approach for reusing previously acquired knowledge by using it to guide the exploration of an agent while it learns new tasks. In order to do so, we employ a variant of the growing self-organizing map algorithm, which is trained using a measure of similarity that is defined directly in the space of the vectorized representations of the value functions. In addition to enabling transfer across tasks, the resulting map is simultaneously used to enable the efficient storage of previously acquired task knowledge in an adaptive and scalable manner. We empirically validate our approach in a simulated navigation environment and also demonstrate its utility through simple experiments using a mobile micro-robotics platform. In addition, we demonstrate the scalability of this approach and analytically examine its relation to the proposed network growth mechanism. Furthermore, we briefly discuss some of the possible improvements and extensions to this approach, as well as its relevance to real-world scenarios in the context of continual learning.

Contrasting effects of pre-training on acquisition of operant and radial armmazetasks in rats.

Performing multiple tasks either simultaneously, in rapid alternation or in succession, is routine in daily life. Further, testing rodents in a battery of tests is common both in drug discovery and behavioral phenotyping research. However, learning of new tasks can be influenced by prior experience(s). There has been some research on 'switching cost' involved in the transition from one behavior to another. However, there has been no specific assessment of the effect of learning an operant paradigm on performance in a spatial memory task and vice versa. Accordingly, we evaluated task switching between two forms of learning paradigms, operant conditioning and radial arm maze (RAM) tasks. In experiment 1, rats were trained for operant conditioning with food reward followed by a partially baited RAM task. In experiment 2, rats were trained first on a RAM task followed by operant learning. Pre-training on the operant task, impaired the acquisition of the RAM. On the contrary, pre-training on the RAM enhanced operant performance. Our study reveals significant effects of the test order on task-switching in rats. This knowledge can be useful when framing test sequences in test batteries for drug discovery research and screening genetically modified mice.