Interactive Learning Framework Research Articles

Unsupervised cross domain semantic segmentation with mutual refinement and information distillation

Unsupervised cross domain semantic segmentation recently has gained much attention, due to its powerful ability of solving the segmentation problem on unlabeled domains. Traditional methods often employ an adversarial network to confuse the source and target inputs, so as to align them in a new feature space. However, these methods cannot well fuse the source and target domain information because the information of the two domains does not really interact with each other and are passed through separate network branches. To tackle this problem, we propose a real interactive learning framework, named Mutual Refinement and Information Distillation (MURID), to align the two domains. Concretely, MURID introduces a Mutual Refinement module in shallow network layers to enhance information sharing and integration between the source and target domains, which can effectively transfer knowledge from source domain to target domain. In addition, in order to avoid using the same structure for testing as for training, which would result in huge computational requirements, we exploit an Information Distillation module to simplify the testing network while maintaining the powerful inference capability of the training. Moreover, we incorporate Curriculum Learning, a self-training mechanism that iteratively trains the network using pseudo-labels obtained from the target domain, to further improve performance. Extensive experiments were conducted on three popular datasets, i.e., GTA5→Cityscapes and Synthia→Cityscapes, and the results demonstrate the state-of-the-art performance of our method. Additionally, detailed analysis and ablation studies are also carried out to validate the effectiveness of each designed module.

Semi-Path: An interactive semi-supervised learning framework for gigapixel pathology image analysis

The efficacy of supervised deep learning in medical image analyses, particularly in pathology, is hindered by the necessity for extensive manual annotations. Annotating images at the gigapixel level manually proves to be a highly labor-intensive and time-consuming task. Semi-supervised learning (SSL) has emerged as a promising approach that leverages unlabeled data to reduce labeling efforts. In this work, we introduce Semi-Path, a practical SSL framework enhanced with active learning (AL) for gigapixel pathology tasks. Unlike existing methods that treat SSL and AL as independent components where AL incurs significant computational complexity to SSL, we propose a deep fusion of SSL and AL into a unified framework. Our framework introduces Informative Active Annotation (IAA) that employs a SSL-AL iterative structure to effectively extract knowledge from unlabeled pathology data. This structure significantly minimizes labeling efforts and computational complexity. Then, we propose Adaptive Pseudo-Labeling (APL) to address heterogeneity in class distribution, and prediction difficulty that are often observed in real-world pathology tasks. We evaluate Semi-Path on pathology image classification and segmentation tasks over three datasets that include WSIs from breast, colorectal, and brain tissues. The experimental results demonstrate the consistent superiority of Semi-Path over state-of-the-art methods.

An Interactive Framework of Cross-Lingual NLU for In-Vehicle Dialogue.

As globalization accelerates, the linguistic diversity and semantic complexity of in-vehicle communication is increasing. In order to meet the needs of different language speakers, this paper proposes an interactive attention-based contrastive learning framework (IABCL) for the field of in-vehicle dialogue, aiming to effectively enhance cross-lingual natural language understanding (NLU). The proposed framework aims to address the challenges of cross-lingual interaction in in-vehicle dialogue systems and provide an effective solution. IABCL is based on a contrastive learning and attention mechanism. First, contrastive learning is applied in the encoder stage. Positive and negative samples are used to allow the model to learn different linguistic expressions of similar meanings. Its main role is to improve the cross-lingual learning ability of the model. Second, the attention mechanism is applied in the decoder stage. By articulating slots and intents with each other, it allows the model to learn the relationship between the two, thus improving the ability of natural language understanding in languages of the same language family. In addition, this paper constructed a multilingual in-vehicle dialogue (MIvD) dataset for experimental evaluation to demonstrate the effectiveness and accuracy of the IABCL framework in cross-lingual dialogue. With the framework studied in this paper, IABCL improves by 2.42% in intent, 1.43% in slot, and 2.67% in overall when compared with the latest model.

Dynamic spatiotemporal interactive graph neural network for multivariate time series forecasting

Multivariate time series (MTS) forecasting holds significant importance in decision-making for complex real-world phenomena. However, the presence of nonlinear temporal correlations within variables and dynamic spatial correlations between variables makes accurate MTS prediction challenging. Currently, there are many researchers who build various spatiotemporal graph neural networks (STGNNs) and apply them to this field. However, most existing methods construct the graph structure using a single type of information and separately capture temporal and spatial features. These factors can result in models that fail to extract complete spatiotemporal features, thereby limiting their performance. To overcome these limitations, we propose the dynamic spatiotemporal interactive graph neural network (DSTIGNN), a novel STGNN for MTS forecasting. The proposed dynamic graph inference module models the dynamic spatial association between variables by fusing two types of heterogeneous information and is combined with the dynamic graph convolution module to propagate information in spatial dimensions. Meanwhile, downsampling operations and multiple sample convolution modules are used to jointly capture multiresolution temporal correlations. Subsequently, these modules are integrated into a spatiotemporal interactive learning framework, enabling the synchronous capture of temporal and spatial features. We have performed numerous experiments on six benchmark datasets, and the experimental results demonstrate that DSTIGNN achieves state-of-the-art performance.

I Enjoy Writing and Playing, Do You?: A Personalized and Emotion Grounded Dialogue Agent Using Generative Adversarial Network

Social chatbots have gained immense popularity, and their appeal lies in their capacity to respond to diverse requests, but also in their ability to develop an emotional connection with users. To develop and promote social chatbots, we need to concentrate on increasing user interaction and consider both the intellectual and emotional quotient in conversational agents. In this work, we propose the task of empathetic, personalized dialogue generation giving the machine the capability to respond emotionally and in accordance with the persona of the user. We design a generative adversarial framework, named EP-GAN (Empathy and Persona aware Generative Adversarial Network) that generates responses that are sensitive to the emotion of the user and corresponds to the persona. The persona information is encoded as memory vectors that, along with the dialogue history, are fed to the decoder for generation. An interactive adversarial learning framework is implemented to verify whether the generated responses elicit the emotion and persona in dialogues. Experimental results show that the EP-GAN framework significantly outperforms the existing baselines for both automatic and manual evaluation. We achieve an improvement of 1% in BLEU-4 and 2% in the emotion accuracy metric compared to the best performing baseline.

Online POI Recommendation: Learning Dynamic Geo-Human Interactions in Streams

Online POI recommendation strives to recommend future visitable places to users over time. It is crucial to improve the user experience of location-based social networking applications (e.g., Google Maps, Yelp). While numerous studies focus on capturing user visit preferences, geo-human interactions are ignored. However, users make visit decisions based on the status of geospatial contexts. In the meantime, such visits will change the status of geospatial contexts. Therefore, disregarding such geo-human interactions in streams will result in inferior recommendation performance. To fill this gap, in this paper, we propose a novel deep interactive reinforcement learning framework to model geo-human interactions. Specifically, this framework has two main parts: the representation module and the imitation module. The purpose of the representation module is to capture geo-human interactions and convert them into embedding vectors (state). The imitation module is a reinforced agent whose job is to imitate user visit behavior by recommending next-visit POI (action) based on the state. Imitation performance is regarded as a reward signal to optimize the whole interactive framework. When the model converges, the imitation module can precisely perceive users and geospatial contexts to provide accurate POI recommendations. Finally, we conduct extensive experiments to validate the superiority of our framework.

Fraud detection on multi-relation graphs via imbalanced and interactive learning

In recent years, the application of Graph Neural Networks (GNNs) in fraud detection has gained considerable attention. GNNs have demonstrated their efficacy in leveraging the abundant relational information inherent in graph-structured data for such tasks. However, despite the remarkable progress achieved thus far, several challenges still need to be addressed in the current GNN-based algorithms. On the one hand, GNNs exhibit limited performance when confronted with imbalanced label distribution between fraudsters and benign users. On the other hand, the embedding representations learned for nodes and their neighbors in the last layer are frequently processed in a simplistic manner, such as concatenation or averaging, which may compromise the performance of downstream tasks. To alleviate the above problems, we propose a novel imbalanced and interactive learning framework for fraud detection on multi-relation graphs (IMINF for short). Concretely, firstly, we design a novel neighbor sampler that can be trained using supervised contrastive learning. This sampler selects the most similar and consistent neighbors to update the target node's features, taking into account label information from multiple perspectives. Secondly, we introduce the learnable contrastive embedding and an improved relational aggregator to enhance the representation of the target node. Thirdly, we propose a new interactive learning module, which consists of explicit and implicit transformation layers, with the aim of capturing deep interactive relations between a target node and its neighbors. Additionally, we devise a feature mapping module prior to the relation aggregation step, transferring the raw input features to a latent feature subspace. Extensive experiments on two real-world fraud detection datasets demonstrate that the proposed model outperforms existing baselines across all evaluation metrics, especially for the Amazon fraud detection dataset.

A novel GRU-TCN network based Interactive Behavior Learning of multi-energy Microgrid under incomplete information

The interactive operation of multi-energy Microgrid (MEMG) in energy market is facing more and more incomplete information decision-making requirements due to the enhanced privacy of users, which brings great challenges to the participation flexibility of MEMG in demand response programs. Meanwhile, considering the uncertainty of the external MEMG unit combination, the randomness of the environmental meteorology, the interactive characteristics of the external MEMGs are complex and time-varying. To solve these problems, a novel GRU-TCN network based Interactive Behavior Learning method for MEMG interactive operation is proposed. The proposed Interactive Behavior Learning framework needs merely externally available input information and historical interaction power data to efficiently predict the interaction power, which protects the data privacy of users. Besides, the GRU-TCN network also leverages the strengths of Recurrent Neural Network and Convolutional Neural Networks while incorporating the self-attention mechanism. This combination enables the GRU-TCN network to effectively capture the relationship between interactive power and the available resources data beyond the MEMG. The example analysis and test are carried out on a typical MEMG system, and the case study turns out that the proposed GRU-TCN network has the capability to capture coupling mechanisms even in the presence of incomplete information and present more accurate prediction of multi-energy interactive data.

Identifying Humans Based on Biometric Iris Recognition using an Interactive Transfer Learning Framework

Identifying Humans Based on Biometric Iris Recognition using an Interactive Transfer Learning Framework

From center to surrounding: An interactive learning framework for hyperspectral image classification

Owing to rich spectral and spatial information, hyperspectral image (HSI) can be utilized for finely classifying different land covers. With the emergence of deep learning techniques, convolutional neural network (CNN), fully convolutional network (FCN), and recurrent neural network (RNN) have been widely applied in the field of HSI classification. Recently, transformer-based approaches represented by Vision Transformer (ViT) have yielded promising performance on numerous tasks and have been introduced to classify HSI. However, existing methods based on the above architectures still face three crucial issues that limit the classification performance: 1) geometric constraints caused by input data, 2) contribution fuzziness of central pixels with details, and 3) interaction gap between local areas and further environments. To tackle the above problems, an interactive learning framework inspired by ViT is proposed from a center to surrounding perspective, namely the center-to-surrounding interactive learning (CSIL) framework. Different from existing works, the CSIL framework enables to achieve multi-scale, detail-aware, and space-interactive classification based on a well-designed hierarchical region sampling strategy, center transformer, and surrounding transformer. Specifically, a hierarchical region sampling strategy is first proposed to flexibly generate the center region, neighbor region, and surrounding region, respectively. Thus, multi-scale input data breaks the geometric constraints. Second, a center transformer is presented to obtain core characteristics in detail based on the center region. In this way, central pixels are remarkably highlighted and the details are easily perceived. Third, a surrounding transformer including interactive self-attention learning is formulated for interacting both locally fine-grained distributions in the neighbor region and further coarse-grained environments in the surrounding region. With this structure, short- and long-term dependencies can be modeled, emphasized, and exchanged to bridge the interaction gap. Finally, the features from center transformer and surrounding transformer are integrated, then fed into a multi-layer perceptron for the optimization of semantic representation. Extensive experiments on six HSI datasets including small-, medium-, and large-scale scenes demonstrate the superiority over state-of-the-art CNN–, FCN-, RNN- and transformer-based approaches, even with very few training samples (for example 0.19% in complex HanChuan city scene). The source code will be available soon at https://github.com/jqyang22/CSIL.

Interactive Machine Learning Framework for Predicting Non-Performing Bank Loans

Interactive Machine Learning Framework for Predicting Non-Performing Bank Loans

Polyphony: an Interactive Transfer Learning Framework for Single-Cell Data Analysis.

Reference-based cell-type annotation can significantly reduce time and effort in single-cell analysis by transferring labels from a previously-annotated dataset to a new dataset. However, label transfer by end-to-end computational methods is challenging due to the entanglement of technical (e.g., from different sequencing batches or techniques) and biological (e.g., from different cellular microenvironments) variations, only the first of which must be removed. To address this issue, we propose Polyphony, an interactive transfer learning (ITL) framework, to complement biologists' knowledge with advanced computational methods. Polyphony is motivated and guided by domain experts' needs for a controllable, interactive, and algorithm-assisted annotation process, identified through interviews with seven biologists. We introduce anchors, i.e., analogous cell populations across datasets, as a paradigm to explain the computational process and collect user feedback for model improvement. We further design a set of visualizations and interactions to empower users to add, delete, or modify anchors, resulting in refined cell type annotations. The effectiveness of this approach is demonstrated through quantitative experiments, two hypothetical use cases, and interviews with two biologists. The results show that our anchor-based ITL method takes advantage of both human and machine intelligence in annotating massive single-cell datasets.

Interactive active learning for fairness with partial group label

The rapid development of AI technologies has found numerous applications across various domains in human society. Ensuring fairness and preventing discrimination are critical considerations in the development of AI models. However, incomplete information often hinders the complete collection of sensitive attributes in real-world applications, primarily due to the high cost and potential privacy violations associated with such data collection. Label reconstruction through building another learner on sensitive attributes is a common approach to address this issue. However, existing methods focus solely on improving the prediction accuracy of the sensitive learner as a separate model, while ignoring the disparity between its accuracy and the fairness of the base model. To bridge this gap, this paper proposes an interactive learning framework that aims to optimize the sensitive learner while considering the fairness of the base learner. Furthermore, a new active sampling strategy is developed to select the most valuable data for the sensitive learner regarding the fairness of the base model. The effectiveness of our proposed method in improving model fairness is demonstrated through comprehensive evaluations conducted on various datasets and fairness criteria.

Deadline-Aware Deep-Recurrent-Q-Network Governor for Smart Energy Saving

Complex cyber-physical-social systems (CPSS) consist of battery-supplied devices with low energy consumption requirements. It is essential to maintain the timing performance of computing or communication tasks while saving the device energy. Linux OS provides built-in frequency governors for power management. However, these governors are not able to incorporate the timing requirements of the application for decision-making and cannot adapt the power management decision to the specific application on target devices. This paper presents an intelligent Linux frequency Deep-Recurrent-Q-Network (DRQN) governor for dedicated applications with deadline requirements running on CPSS devices through machine learning. Although machine learning algorithms have made considerable breakthroughs in recent years, deploying them to real small devices is challenging because of the computational overhead. To tackle the computation overhead problem, an interactive learning framework is designed where the DRQN model performs only the lightest inference in the kernel while using the online data (time-series) to learn and update itself in real-time at the user level. The governor is tested on both standalone devices and networked devices. The experiment shows that DRQN can self-develop tradeoff policy to meet the user's need with low overhead. The energy saved by DRQN ranges from 7% to 33% for various deadlines.

Factors Influencing Analysis for Level of Engineering English Education Based on Artificial Intelligence Technology

The essence of English for engineering is English for professional purposes (ESP). The assessment of the level of education in engineering English classrooms is one of the key issues currently being discussed in all schools. While the traditional English classroom teaching model has been criticized due to many problems, the changes in the new English curriculum and the changes in the assessment methods have also created new demands on the English teaching methods. The artificial intelligence technology brings a new direction for the optimization of English classroom, and it also provides new support for the realization of intelligent and collaborative English classroom teaching. In view of the current situation that the college English teaching evaluation mode is still dominated by summative evaluation, this paper summarizes the current problems of poor teaching effect, unbalanced ability cultivation, and mismatch between evaluation and teaching in engineering English. On this basis, it adopts the BPNN network combined with the interactive mechanism of English teaching to establish a multi-dimensional interactive English learning framework of teacher, student, corpus, and AI resource base. And the BPNN algorithm process was optimized using the gray wolf algorithm to improve the engineering English teaching model. Finally, an experiment on teaching engineering English was conducted within a university, and the experimental results showed that teaching objectives, teaching contents, teaching methods, and teaching effects had important effects on teaching effectiveness. In addition, the teaching framework constructed using the improved BPNN algorithm was better in terms of the learning effect at the same time, especially in writing, during the teaching process. Finally, the experimental results show that BPNN optimized by gray wolf algorithm can achieve better teaching effect than BPNN.

Using data about classroom practices to stimulate significant conversations and aid reflection

ABSTRACT This study explored the way in which detailed data about how time is spent on classroom activities, generated by the FILL+ tool (Framework for Interactive Learning in Lectures), can stimulate professional conversations about teaching practices and aid reflection for STEMM lecturers. The lecturers felt that personalised data provided an unbiased view of the lecture, overcoming the difficulty of relying on their memory alone. They indicated that this approach would help them to reflect on their teaching, particularly when the data was surprising, and many felt that it would encourage them to make changes to their teaching practice.

Correct Me If I am Wrong: Interactive Learning for Robotic Manipulation

Learning to solve complex manipulation tasks from visual observations is a dominant challenge for real-world robot learning. Although deep reinforcement learning algorithms have recently demonstrated impressive results in this context, they still require an impractical amount of time-consuming trial-and-error iterations. In this work, we consider the promising alternative paradigm of interactive learning in which a human teacher provides feedback to the policy during execution, as opposed to imitation learning where a pre-collected dataset of perfect demonstrations is used. Our proposed CEILing (Corrective and Evaluative Interactive Learning) framework combines both corrective and evaluative feedback from the teacher to train a stochastic policy in an asynchronous manner, and employs a dedicated mechanism to trade off human corrections with the robot’s own experience. We present results obtained with our framework in extensive simulation and real-world experiments to demonstrate that CEILing can effectively solve complex robot manipulation tasks directly from raw images in less than one hour of real-world training.

Design and Application of Automatic Feedback Scaffolding in Forums to Promote Learning

Forums are essential components facilitating interactions in online courses. However, in large-scale courses, many posts generated, which results in learners’ difficulties. First, the posts are poorly organized and some deviate from the topic, making it difficult for learners’ knowledge acquisition. Second, learners cannot receive timely feedback and guidance, making the learning progress unclear for them. Well-designed scaffoldings should be built based on challenges of forums to improve learners’ learning outcomes, knowledge construction, and completion rate. While targeting the problems in online forums, this article proposed principles for the design of online scaffolding after analyzing the requirements of online learning scaffolding or scripts. Subsequently, in this article, we designed an automatic feedback scaffolding based on the principles and a knowledge construction model. The scaffolding provided learners with timely feedback and related learning guidance. Tags were used to assist learners in acquiring relevant information more easily. The scaffolding was then integrated into the Learning Cell Knowledge Community and used in an online course for 955 learners. The results showed that automatic feedback scaffolding positively affected learners’ learning and promoted positive knowledge transformation. Furthermore, we found that the scaffolding could help learners induce more constructive behaviors defined in the Interactive, Constructive, Active, and Passive deep learning framework that demonstrated the reason for learners’ knowledge transformation. At last, learners’ course completion rate also increased with the help of the scaffolding, which provided evidence that well-designed scaffolding can result in positive educational outcomes. In addition, the principles proposed could also contribute to further scaffolding design and practices.

Stakeholder Power Analysis of the Facilitators and Barriers for Telehealth Solution Implementation in China: A Qualitative Study of Individual Users in Beijing and Interviews With Institutional Stakeholders.

BackgroundFacing COVID 19, the use of telehealth solutions grows exponentially. However, despite the large investments made into telehealth solutions, the implementation process remains slow and sluggish. Moreover, during COVID-19, older people experienced difficulties and had the highest mortality rates, and those lucky enough to survive faced tremendous pressure to use QR code-based health monitoring systems.ObjectiveThis paper aims to determine the barriers and incentives for the implementation of telehealth solutions via a case study about telehealth implementation in China.MethodsWe conducted 8 semi-structured interviews following the design of the interactive learning framework (research question defining, participant recruitment, exploratory stage, consultation stage, integration stage, and follow-up interview). One interview with a government official from the National Health Commission and another interview with a government official from the China Center for Disease Control and Prevention was conducted in the exploratory stage. The consultation stage comprised one interview with a business manager from the Huawei Wearable Unit, one interview with a business manager from Alibaba Health Brain Unit, and one interview with a business manager from Xiaomi. Two interviews with doctors from Fudan University-affiliated Huashan Hospital and Fudan University-affiliated Zhongshan hospital were conducted in the integration stage. In addition, 8 focus group studies with 64 participants from rural and urban Beijing were conducted. Finally, another telephone interview with a business manager of the Xiaomi Wearable Unit was conducted in the follow-up stage.ResultsTelehealth solutions are designed to assist health care providers in realizing the quadruple aim of better health outcomes, lowering health care costs, improved health care quality, and improved doctor and patient experiences. Governments have high incentives to improve health care efficiency via telehealth solutions. However, they have limited resources to make the necessary infrastructure transformation.ConclusionsTo fully realize the potential of telehealth devices, heavy infrastructure investment in the telecommunication network is required beforehand to resolve the interoperability issue occurring during the data collection process for telehealth solutions. The industry also demands a mature business model incorporating collaboration between various stakeholders and industrial partners to invest in infrastructure. Governments have high interest and significant influence on building the necessary infrastructure for telehealth solution implementation in China. Industrial actors have a high interest and a medium level of power for telehealth solution implementation. Users have high interest but a lower level of power for the usage of telehealth solutions, and doctors have low interest and a medium level of power for telehealth solutions implementation.

Dynamic Interactive Learning for Lightweight Detectors in Remote Sensing Imagery

The lightweight model has played an important role in the remote sensing (RS) realm. The existing researchers have proposed many models with lightweight structures, but their performance still has a gap compared with the deep model. A promising approach to optimize the lightweight model is knowledge distillation (KD), which can be viewed as knowledge transfer from the teacher model. However, the existing KD approaches have some issues. On one hand, offline distillation methods usually ignore the interactive learning between the student model and the teacher model. On the other hand, knowledge transfer does not consider instance property. This offline distillation strategy without property perception may not suitable for multiscale, diverse, and complex RS instances and results in a suboptimum training status. In this article, we propose a dynamic interactive learning (DIL) framework for optimizing RS lightweight detectors. First, we propose an instance interaction learning module. It calculates the value of every instance in the batch of the teacher and student prediction by each model’s real-time state and instance property. Then according to the DIL thought, we facilitate the low-quality instance to learn from the high-quality one whether it is from the teacher or student model. Moreover, we also propose the instance property perception (IPP) strategy that weighs the distillation knowledge of instances according to their feature, category, and location property. In the proposed DIL framework, both the teacher and student models are trained together and it is cost-free in the testing phase. Extensive experiments on three RS datasets demonstrate the effectiveness of the DIL.