Human Learning Process Research Articles

Integrating human learning and reinforcement learning: A novel approach to agent training

Off-policy reinforcement learning (RL) algorithms are known for improving sample efficiency by employing prior experiences in experience replay memory. However, most existing off-policy RL algorithms are bottlenecked by the slow convergence speed and demand of a large number of interaction samples. In contrast, on-policy RL algorithms converge fast and continuously generate new samples. However, their success heavily depends on the accuracy of the generated samples. To address these challenges, a novel RL framework called HI-FER (Human-learning Inspired Frequent Experience Replay) is proposed by mimicking the process of human learning. HI-FER employs a parallelized experience replay and repetitive training framework to expedite the convergence rate of off-policy algorithms, which imitates the function of the human brain’s parallel information processing and repetitive learning. Additionally, a periodic network reset strategy and dynamic memory updating are leveraged by imitating the forgetting mechanism of humans to prevent overfitting triggered by repetitive updating on limited experiences. Extensive comparison experiments and ablation studies are performed on benchmark environments to evaluate the proposed method. The empirical results demonstrate that HI-FER outperforms the baselines in terms of sample efficiency on state-based (14% improvements) and image-based (51% improvements) tasks from DMControl. Project website and code: https://github.com/Arya87/HI-FER.

Knowledge-Aware Learning Framework Based on Schema Theory to Complement Large Learning Models

ABSTRACT Despite tremendous recent progress, extant artificial intelligence (AI) still falls short of matching human learning in effectiveness and efficiency. One fundamental disparity is that humans possess a wealth of prior knowledge, while AI lacks the essential commonsense knowledge required for learning tasks. Guided by schema theory, we employ the design science research methodology to introduce a novel knowledge-aware learning framework to harness the knowledge-based processes in human learning. Unlike existing pre-trained large language models (LLMs) and knowledge-aware approaches that treat knowledge in considerably different ways from humans, our theoretically grounded framework closely mimics how humans acquire, represent, activate, and utilize knowledge. The extensive evaluations in the context of text analytics tasks demonstrate that our design achieves comparable performance to the state-of-the-art LLMs and enhances model generalizability and learning efficiency. This study takes a step forward by bringing cognitive science into building cognitively plausible AI and human-AI collaboration research.

FeatWalk: Enhancing Few-Shot Classification through Local View Leveraging

Few-shot learning is a challenging task due to the limited availability of training samples. Recent few-shot learning studies with meta-learning and simple transfer learning methods have achieved promising performance. However, the feature extractor pre-trained with the upstream dataset may neglect the extraction of certain features which could be crucial for downstream tasks. In this study, inspired by the process of human learning in few-shot tasks, where humans not only observe the whole image (`global view') but also attend to various local image regions (`local view') for comprehensive understanding of detailed features, we propose a simple yet effective few-shot learning method called FeatWalk which can utilize the complementary nature of global and local views, therefore providing an intuitive and effective solution to the problem of insufficient local information extraction from the pre-trained feature extractor. Our method can be easily and flexibly combined with various existing methods, further enhancing few-shot learning performance. Extensive experiments on multiple benchmark datasets consistently demonstrate the effectiveness and versatility of our method.The source code is available at https://github.com/exceefind/FeatWalk.

Emergent Communication for Numerical Concepts Generalization

Research on emergent communication has recently gained significant traction as a promising avenue for the linguistic community to unravel human language's origins and explore artificial intelligence's generalization capabilities. Current research has predominantly concentrated on recognizing qualitative patterns of object attributes(e.g., shape and color) and paid little attention to the quantitative relationship among object quantities which is known as the part of numerical concepts. The ability to generalize numerical concepts, i.e., counting and calculations with unseen quantities, is essential, as it mirrors humans' foundational abstract reasoning abilities. In this work, we introduce the NumGame, leveraging the referential game framework, forcing agents to communicate and generalize the numerical concepts effectively. Inspired by the human learning process of numbers, we present a two-stage training approach that sequentially fosters a rudimentary numerical sense followed by the ability of arithmetic calculation, ultimately aiding agents in generating semantically stable and unambiguous language for numerical concepts. The experimental results indicate the impressive generalization capabilities to unseen quantities and regularity of the language emergence from communication.

ProCC: Progressive Cross-Primitive Compatibility for Open-World Compositional Zero-Shot Learning

Open-World Compositional Zero-shot Learning (OW-CZSL) aims to recognize novel compositions of state and object primitives in images with no priors on the compositional space, which induces a tremendously large output space containing all possible state-object compositions. Existing works either learn the joint compositional state-object embedding or predict simple primitives with separate classifiers. However, the former method heavily relies on external word embedding methods, and the latter ignores the interactions of interdependent primitives, respectively. In this paper, we revisit the primitive prediction approach and propose a novel method, termed Progressive Cross-primitive Compatibility (ProCC), to mimic the human learning process for OW-CZSL tasks. Specifically, the cross-primitive compatibility module explicitly learns to model the interactions of state and object features with the trainable memory units, which efficiently acquires cross-primitive visual attention to reason high-feasibility compositions, without the aid of external knowledge. Moreover, to alleviate the invalid cross-primitive interactions, especially for partial-supervision conditions (pCZSL), we design a progressive training paradigm to optimize the primitive classifiers conditioned on pre-trained features in an easy-to-hard manner. Extensive experiments on three widely used benchmark datasets demonstrate that our method outperforms other representative methods on both OW-CZSL and pCZSL settings by large margins.

Dissecting human anatomy learning process through anatomical education with augmented reality: AEducAR 2.0, an updated interdisciplinary study.

Anatomical education is pivotal for medical students, and innovative technologies like augmented reality (AR) are transforming the field. This study aimed to enhance the interactive features of the AEducAR prototype, an AR tool developed by the University of Bologna, and explore its impact on human anatomy learning process in 130 second-year medical students at the International School of Medicine and Surgery of the University of Bologna. An interdisciplinary team of anatomists, maxillofacial surgeons, biomedical engineers, and educational scientists collaborated to ensure a comprehensive understanding of the study's objectives. Students used the updated version of AEducAR, named AEducAR 2.0, to study three anatomical topics, specifically the orbit zone, facial bones, and mimic muscles. AEducAR 2.0 offered two learning activities: one explorative and one interactive. Following each activity, students took a test to assess learning outcomes. Students also completed an anonymous questionnaire to provide background information and offer their perceptions of the activity. Additionally, 10 students participated in interviews for further insights. The results demonstrated that AEducAR 2.0 effectively facilitated learning and students' engagement. Students totalized high scores in both quizzes and declared to have appreciated the interactive features that were implemented. Moreover, interviews shed light on the interesting topic of blended learning. In particular, the present study suggests that incorporating AR into medical education alongside traditional methods might prove advantageous for students' academic and future professional endeavors. In this light, this study contributes to the growing research emphasizing the potential role of AR in shaping the future of medical education.

Critical thinking: An analytical study on the impact of Artificial Intelligence (AI) usage on students at Al Washliyah Darussalam University, Banda Aceh

Artificial Intelligence (AI) is a technology that can mimic human thinking and learning processes through algorithms and computer systems. AI brings significant changes to the education sector by providing various conveniences. On the other hand, the use of AI technology requires adjustments in the learning process at higher education institutions. This research aims to analyze the impact of Artificial Intelligence (AI) usage on the critical thinking skills of students at Al Washliyah Darussalam University (UNADA) Banda Aceh. This study employs a Mixed Methods approach, utilizing quantitative methods with a descriptive approach to analyze the frequency of AI usage, while qualitative methods are used to analyze changes in thinking patterns and students' analytical abilities in the learning process. Data collection was conducted through surveys. The sampling technique used was purposive sampling, involving 287 students who use AI in their learning. The research findings indicate that the distribution of questionnaire data shows that the usage of Artificial Intelligence (AI) includes 31.3% (ChatGPT), 6.3% (Perplexity), 12.5% (Gemini), and 5% using other AI technologies. The impact of AI usage reveals that AI can enhance students' critical thinking skills, particularly in data analysis, problem-solving, and decision-making. However, negative impacts faced by students include dependency on technology and a lack of active engagement in classroom discussions. The results and findings of this study highlight the importance for university lecturers and policymakers to design effective learning strategies that utilize AI to develop critical thinking skills among students.

Domain-incremental learning without forgetting based on random vector functional link networks

Incremental learning is a paradigm that extends knowledge by learning from new data, often used to add new classes to an existing model or to learn a new domain. It imposes strict limitations on the model’s access to data from previous tasks, making it similar to the human learning process. The main challenge of incremental learning is catastrophic forgetting, where previous knowledge is severely forgotten while learning new tasks. In this work, we propose a novel approach for domain-incremental learning. Inspired by the Normal Equation, we accumulate the Gram Matrix from each task’s hidden layer output to update a simplified RVFL model. This algorithm achieves performance comparable to joint training while strictly adhering to privacy restrictions. With issues such as forgetting, storage requirements and privacy protection be addressed, this algorithm has the potential to play a crucial role in the field of edge computing and other related fields.

A deeper knowledge tracking model integrating cognitive theory and learning behavior

Knowledge tracing (KT), which aims to trace human knowledge learning process by using machines, has widely applied in online learning systems. It dynamically models student’s knowledge states in relation to different learning factors through their learning interactions. Recently, KT has attracted many researches attention due to its good performance to using deep learning. Although most of KT models have shown outstanding results, they have limitations: either ignore the human cognitive law and learning behavior, or lack the ability to go deeper modeling to trace knowledge state. In this paper, we propose a deeper knowledge tracking model integrating cognitive theory and learning behavior (CLDKT). It united the advantages of memory network and recurrent neural network of the existing deep learning KT models for modeling student learning. To better implement CLDKT, we add the residual network (ResNet) to realize the deep modeling of learning behaviors. Extensive experiments on three open benchmark datasets to evaluate our model. Experimental results demonstrate that (I) CLDKT outperforms the state-of-the-art KT models on students’ performance prediction. (II) CLDKT can deeper modeling to trace knowledge state owing to the ResNet import. (III) CLDKT has better interpretability and predictability, which proves the effectiveness of the knowledge tracing model integrating cognitive law and learning behavior.

Unifying Education as Problem Solving

Public education in the United States is breaking down on two dimensions: its bureaucratic organization and its conceptualization of the learning process. These breakdowns result from attempts to deliver public education using business models and have failed to support the needs of both students and teachers. This breakdown has been further exacerbated by an old misconception about intelligence based on inherent, immutable aspects of biology. This view medicalizes learning problems, enabling schools to categorize already marginalized students producing a stigmatization that negatively impacts their development. To address these problems it is proposed that public education be reframed as a developmental specialization of problem-solving skills. This reframing would ground the conceptualization of learning in the most natural human learning process: problem-solving. Such an approach could directly support students in addressing problems that undermine their individual performance. This problem-solving-centered approach could form both a supportive and successful core educational routine that would model and support stronger collaborative structures between students, teachers, and administrators, building stronger institutions.

Curriculum Consistency Learning and Multi-Scale Contrastive Constraint in Semi-Supervised Medical Image Segmentation.

Data scarcity poses a significant challenge in medical image segmentation, thereby highlighting the importance of leveraging sparse annotation data. In addressing this issue, semi-supervised learning has emerged as an effective approach for training neural networks using limited labeled data. In this study, we introduced a curriculum consistency constraint within the context of semi-supervised medical image segmentation, thus drawing inspiration from the human learning process. By dynamically comparing patch features with full image features, we enhanced the network's ability to learn. Unlike existing methods, our approach adapted the patch size to simulate the human curriculum process, thereby progressing from easy to hard tasks. This adjustment guided the model toward improved convergence optima and generalization. Furthermore, we employed multi-scale contrast learning to enhance the representation of features. Our method capitalizes on the features extracted from multiple layers to explore additional semantic information and point-wise representations. To evaluate the effectiveness of our proposed approach, we conducted experiments on the Kvasir-SEG polyp dataset and the ISIC 2018 skin lesion dataset. The experimental results demonstrated that our method surpassed state-of-the-art semi-supervised methods by achieving a 9.2% increase in the mean intersection over union (mIoU) for the Kvasir-SEG dataset. This improvement substantiated the efficacy of our proposed curriculum consistency constraint and multi-scale contrastive loss.

CL3: Generalization of Contrastive Loss for Lifelong Learning.

Lifelong learning portrays learning gradually in nonstationary environments and emulates the process of human learning, which is efficient, robust, and able to learn new concepts incrementally from sequential experience. To equip neural networks with such a capability, one needs to overcome the problem of catastrophic forgetting, the phenomenon of forgetting past knowledge while learning new concepts. In this work, we propose a novel knowledge distillation algorithm that makes use of contrastive learning to help a neural network to preserve its past knowledge while learning from a series of tasks. Our proposed generalized form of contrastive distillation strategy tackles catastrophic forgetting of old knowledge, and minimizes semantic drift by maintaining a similar embedding space, as well as ensures compactness in feature distribution to accommodate novel tasks in a current model. Our comprehensive study shows that our method achieves improved performances in the challenging class-incremental, task-incremental, and domain-incremental learning for supervised scenarios.

Reciprocal Human-Machine Learning: A Theory and an Instantiation for the Case of Message Classification

There is growing agreement among researchers and developers that in certain machine-learning (ML) tasks, it may be advantageous to keep a “human in the loop” rather than rely on fully autonomous systems. Continual human involvement can mitigate machine bias and performance deterioration while enabling humans to continue learning from insights derived by ML. Yet a microlevel theory that effectively facilitates joint and continual learning in both humans and machines is still lacking. To address this need, we adopt a design science approach and build on theories of human reciprocal learning to develop an abstract configuration for reciprocal human-ML (RHML) in the context of text message classification. This configuration supports learning cycles between humans and machines who repeatedly exchange feedback regarding a classification task and adjust their knowledge representations accordingly. Our configuration is instantiated in Fusion, a novel technology artifact. Fusion is developed iteratively in two case studies of cybersecurity forums (drug trafficking and hacker attacks), in which domain experts and ML models jointly learn to classify textual messages. In the final stage, we conducted two experiments of the RHML configuration to gauge both human and machine learning processes over eight learning cycles. Generalizing our insights, we provide formal design principles for the development of systems to support RHML. This paper was accepted by D. J. Wu, special issue on the human-algorithm connection. Funding: This work was supported by the Israel’s Ministry of Defence [Grant R4441197567] and the Israel’s Ministry of Science and Technology [Grant 207076]. Supplemental Material: The data files are available at https://doi.org/10.1287/mnsc.2022.03518 .

Integrating Curricula with Replays: Its Effects on Continual Learning

Humans engage in learning and reviewing processes with curricula when acquiring new skills or knowledge. This human learning behavior has inspired the integration of curricula with replay methods in continual learning agents. The goal is to emulate the human learning process, thereby improving knowledge retention and facilitating learning transfer. Existing replay methods in continual learning agents involve the random selection and ordering of data from previous tasks, which has shown to be effective. However, limited research has explored the integration of different curricula with replay methods to enhance continual learning. Our study takes initial steps in examining the impact of integrating curricula with replay methods on continual learning in three specific aspects: the interleaved frequency of replayed exemplars with training data, the sequence in which exemplars are replayed, and the strategy for selecting exemplars into the replay buffer. These aspects of curricula design align with cognitive psychology principles and leverage the benefits of interleaved practice during replays, easy-to-hard rehearsal, and exemplar selection strategy involving exemplars from a uniform distribution of difficulties. Based on our results, these three curricula effectively mitigated catastrophic forgetting and enhanced positive knowledge transfer, demonstrating the potential of curricula in advancing continual learning methodologies. Our code and data are available: https://github.com/ZhangLab-DeepNeuroCogLab/Integrating-Curricula-with-Replays

Debiased Contrastive Curriculum Learning for Progressive Generalizable Person Re-Identification

Domain generalization (DG) in person re-identification (ReID) is an extremely challenging but essential task, which aims to learn a generalizable model over multiple labeled source domains that can perform well on unseen target domains. Most existing DG strategies in ReID directly aggregate multiple source data together for training, incurring a large inter-domain bias and unstable model optimization that lead the model apt to overfitting domain bias and the model training more time-consuming respectively, thus hampering the generalization and convergence speed of the model. To tackle the aforementioned issues, inspired by Curriculum Learning that mimics the process of human lifelong DG learning (from easy to hard), we put forward a novel Debiased Contrastive Curriculum Learning (DCCL) strategy for DG ReID, which is designed to incrementally enhance generalization in an easy-to-hard training way that can continuously accumulate learning experience to make learning in unknown domains easier and effectively eliminate the domain bias to help the model learn rich domain-invariant discriminative features, thereby strengthening generalization and accelerating convergence for the model. In addition, to simultaneously learn class-level and instance-level discriminative representations, we raise a non-parametric hybrid contrastive loss to equip the DCCL model. We also particularly design an inter-domain mix module to variegate the features of the newly added source domain at each stage of DCCL, further establishing the advantages of DCCL. Extensive experimental results on four public ReID benchmarks fully demonstrate that our DCCL can effectively strengthen the generalization capacity of the model to unseen domains and outperform the state-of-the-art methods.

HAT-GAE: Self-supervised graph autoencoders with hierarchical adaptive masking and trainable corruption

In recent years, self-supervised autoencoders have developed into a powerful representation learning framework in computer vision and natural language processing. However, their application to graph data has resulted in limited performance due to the non-Euclidean and complex structure of graphs compared to images or text, as well as the limitations of conventional autoencoder architectures. In this paper, we investigate factors impacting the performance of autoencoders on graph data and propose a novel autoencoder model for graph representation learning. Our model incorporates a hierarchical adaptive masking mechanism to incrementally increase the difficulty of training to mimic the process of human cognitive learning and a trainable corruption scheme to enhance the robustness of learned representations. Through extensive experimentation on ten benchmark datasets, we demonstrate the superiority of our proposed method over state-of-the-art graph representation learning models.

That’s political! A Freirean perspective towards coaching as a social practice

Demands to re-embed coaching into its larger social context of operation have generated calls to better document the political aspects of this human learning and development process. To address this critical social turn, this empirical article explores the reframing by practitioners of their understanding of coaching practice, using a Freirean lens of oppression and emancipation. The study consists of a 9-month co-operative inquiry with a group of Egyptian practitioners engaged in a praxis comprising initial training and subsequent cycles of action and reflexive workshops about their coaching sessions. Our study found that a Freire inspired praxis led coaches to develop a dialectical understanding of oppression – moving from conflicting dichotomies to a dynamic view, which has implications for their attitude to and role in coaching. We discuss how the resulting politicization allowed the reframing of coaching as a social practice.

Deep Reinforcement Learning Based Optimization Algorithm for Permutation Flow-Shop Scheduling

As a new analogy paradigm of human learning process, reinforcement learning (RL) has become an emerging topic in computational intelligence (CI). The synergy between the RL and CI is an emerging way to develop efficient solution algorithms for solving complex combinatorial optimization (CO) problems like machine scheduling problem. In this paper, we proposed an efficient optimization algorithm based on Deep RL for solving permutation flow-shop scheduling problem (PFSP) to minimize the maximum completion time. Firstly, a new deep neural network (PFSPNet) is designed for the PFSP to achieve the end-to-end output without limitation of problem sizes. Secondly, an actor-critic method of RL is used to train the PFSPNet without depending on the collection of high-quality labelled data. Thirdly, an improvement strategy is designed to refine the solution provided by the PFSPNet. Simulation results and statistical comparison show that the proposed optimization algorithm based on deep RL can obtain better results than the existing heuristics in similar computational time for solving the PFSP.

Encoder–decoder cycle for visual question answering based on perception-action cycle

In this study, we propose a novel encoder–decoder cycle (EDC) framework inspired by the human learning process called the perception-action cycle to tackle challenging problems such as visual question answering (VQA) and visual relationship detection (VRD). EDC considers the understanding of the visual features of an image as perception and the act of answering the question regarding that image as an action. In the perception-action cycle, information is primarily collected from the environment and then passed to sensory structures in the brain to form an understanding of the environment. Acquired knowledge is then passed to motor structures to perform an action on the environment. Next, sensory structures perceive the altered environment and improve their understanding of the surrounding world. This process of understanding the environment, performing an action correspondingly, and then re-evaluating the initial understanding occurs cyclically in human life. EDC initially mimics this mechanism of introspection by comprehending and refining visual features to acquire the proper knowledge for answering the question. Subsequently, it decodes visual and language features into answer features, feeding them back cyclically to the encoder. In the VRD task, EDC decodes visual features to generate predicate features. We evaluate the proposed framework on the TDIUC, VQA 2.0, and VRD datasets, which outperforms the state-of-the-art models on the TDIUC and VRD datasets.

A Study on Variable Selections and Prediction for Crop Recommender System with Soil Nutrients Using Stochastic Model and Machine Learning Approaches

To develop a crop recommendation system using soil nutrient data, you'll need a dataset containing details on soil nutrients and the crops that thrive in particular soil conditions. While I can't supply a specific dataset, I can offer guidance on the types of data you should seek or gather for building such a system. Machine learning, a subset of artificial intelligence (AI) and computer science, centers on harnessing data and algorithms to replicate human learning processes, steadily enhancing its precision over time. This paper considers crop recommender dataset with soil nutrients-related dataset like N, P, K, ph, EC, S, Cu, Fe, Mn, Zn, B, label. The machine learning approaches are used to analyze and predict the dataset using Logistic, Multilayer Perceptron, Simple Logistic, Hoeffding Tree, random forest, random tree, and REP tree. Numerical illustrations are provided to prove the proposed results with test statistics or accuracy parameters.