Knowledge Transfer Techniques Research Articles

Knowledge transfer in lifelong machine learning: a systematic literature review

Lifelong Machine Learning (LML) denotes a scenario involving multiple sequential tasks, each accompanied by its respective dataset, in order to solve specific learning problems. In this context, the focus of LML techniques is on utilizing already acquired knowledge to adapt to new tasks efficiently. Essentially, LML concerns about facing new tasks while exploiting the knowledge previously gathered from earlier tasks not only to help in adapting to new tasks but also to enrich the understanding of past ones. By understanding this concept, one can better grasp one of the major obstacles in LML, known as Knowledge Transfer (KT). This systematic literature review aims to explore state-of-the-art KT techniques within LML and assess the evaluation metrics and commonly utilized datasets in this field, thereby keeping the LML research community updated with the latest developments. From an initial pool of 417 articles from four distinguished databases, 30 were deemed highly pertinent for the information extraction phase. The analysis recognizes four primary KT techniques: Replay, Regularization, Parameter Isolation, and Hybrid. This study delves into the characteristics of these techniques across both neural network (NN) and non-neural network (non-NN) frameworks, highlighting their distinct advantages that have captured researchers’ interest. It was found that the majority of the studies focused on supervised learning within an NN modelling framework, particularly employing Parameter Isolation and Hybrid for KT. The paper concludes by pinpointing research opportunities, including investigating non-NN models for Replay and exploring applications outside of computer vision (CV).

Constrained multitasking optimization via co-evolution and domain adaptation

Constrained multitasking optimization (CMTO) obtains increasing attention recently. The goal of CMTO is to handle multiple constrained tasks simultaneously. There are two limitations of existing studies on CMTO: (i) existing knowledge transfer techniques for CMTO may not work due to non-intersecting feasible domains and the unconsidered relationship between constraints and objectives; and (ii) knowledge diversity is lacking in existing CMTO algorithms because the representation of knowledge is biased to feasible solutions. To address these limitations, this paper proposes a co-evolution and domain adaptation (CEDA) method for CMTO. First, a new constraint relaxation-based domain adaptation technique for knowledge transfer is devised. Domain adaptation can effectively address the limitations imposed by non-intersecting feasible domains. In addition, as the population evolves, the knowledge representation is biased to different kinds of solutions. Second, a co-evolutionary strategy is proposed to improve the knowledge diversity. The two proposed techniques are with generality and can be readily integrated into different multitasking frameworks. In this paper, the CEDA method is combined with two popular multitasking (i.e., multifactorial-based and multi-population-based) frameworks. The constructed CEDA-based algorithms are compared with fifteen state-of-the-art algorithms on a CMTO benchmark suite and a real-world application. Experimental results demonstrate the superiority of the proposed CEDA method.

Evaluating the stealth of reinforcement learning-based cyber attacks against unknown scenarios using knowledge transfer techniques

Organizations are vulnerable to cyber attacks as they rely on computer networks and the internet for communication and data storage. While Reinforcement Learning (RL) is a widely used strategy to simulate and learn from these attacks, RL-guided offensives against unknown scenarios often lead to early exposure due to low stealth resulting from mistakes during the training phase. To address this issue, this work evaluates if the use of Knowledge Transfer Techniques (KTT), such as Transfer Learning and Imitation Learning, reduces the probability of early exposure by smoothing mistakes during training. This study developed a laboratory platform and a method to compare RL-based cyber attacks using KTT for unknown scenarios. The experiments simulated 2 unknown scenarios using 4 traditional RL algorithms and 4 KTT. In the results, although some algorithms using KTT obtained superior results, they were not so significant for stealth during the initial epochs of training. Nevertheless, experiments also revealed that throughout the entire learning cycle, Trust Region Policy Optimization (TRPO) is a promising algorithm for conducting cyber offensives based on Reinforcement Learning.

X4D-SceneFormer: Enhanced Scene Understanding on 4D Point Cloud Videos through Cross-Modal Knowledge Transfer

The field of 4D point cloud understanding is rapidly developing with the goal of analyzing dynamic 3D point cloud sequences. However, it remains a challenging task due to the sparsity and lack of texture in point clouds. Moreover, the irregularity of point cloud poses a difficulty in aligning temporal information within video sequences. To address these issues, we propose a novel cross-modal knowledge transfer framework, called X4D-SceneFormer. This framework enhances 4D-Scene understanding by transferring texture priors from RGB sequences using a Transformer architecture with temporal relationship mining. Specifically, the framework is designed with a dual-branch architecture, consisting of an 4D point cloud transformer and a Gradient-aware Image Transformer (GIT). The GIT combines visual texture and temporal correlation features to offer rich semantics and dynamics for better point cloud representation. During training, we employ multiple knowledge transfer techniques, including temporal consistency losses and masked self-attention, to strengthen the knowledge transfer between modalities. This leads to enhanced performance during inference using single-modal 4D point cloud inputs. Extensive experiments demonstrate the superior performance of our framework on various 4D point cloud video understanding tasks, including action recognition, action segmentation and semantic segmentation. The results achieve 1st places, i.e., 85.3% (+7.9%) accuracy and 47.3% (+5.0%) mIoU for 4D action segmentation and semantic segmentation, on the HOI4D challenge, outperforming previous state-of-the-art by a large margin. We release the code at https://github.com/jinglinglingling/X4D.

Soil Moisture Forecast Using Transfer Learning: An Application in the High Tropical Andes

Soil moisture is a critical variable in the hydrological cycle and the climate system, significantly impacting water resources, ecosystem functioning, and the occurrence of extreme events. However, soil moisture data are often scarce, and soil water dynamics are not fully understood in mountainous regions such as the tropical Andes of Ecuador. This study aims to model and predict soil moisture dynamics using in situ-collected hydrometeorological data for training and data-driven machine-learning techniques. Our results highlight the fundamental role of vegetation in controlling soil moisture dynamics and significant differences in soil water balance related to vegetation types and topography. A baseline model was developed to predict soil moisture dynamics using neural network techniques. Subsequently, by employing transfer-learning techniques, this model was effectively applied to different soil horizons and profiles, demonstrating its generalization capacity and adaptability. The use of neural network schemes and knowledge transfer techniques allowed us to develop predictive models for soil moisture trained on in situ-collected hydrometeorological data. The transfer-learning technique, which leveraged the knowledge from a pre-trained model to a model with a similar domain, yielded results with errors on the order of 1×10−6<ϵ<1×10−3. For the training data, the forecast of the base network demonstrated excellent results, with the lowest magnitude error metric RMSE equal to 4.77×10−6, and NSE and KGE both equal to 0.97. These models show promising potential to accurately predict short-term soil moisture dynamics with potential applications for natural hazard monitoring in mountainous regions.

Enhancing frame-level student engagement classification through knowledge transfer techniques

Enhancing frame-level student engagement classification through knowledge transfer techniques

Tailored continuous risk management in nanosatellite space project VZLUSAT-1 using FMECA

The purpose of the research is to demonstrate an application of the existing methodology for risk assessment Failure Mode and Effect Criticality Analysis (FMECA) in nanosatellite projects together with a risk register proposal elaborated based on a literature review and expert interviews. The paper also elaborates on the integration of FMECA and knowledge capture and transfer technique. Literature evidence and the use case of the VZLUSAT-1 CubeSat show that it is challenging to implement proper project management and especially risk management practices into nanosatellite (or CubeSat) projects. The risks identified are mostly connected with the technical part, design, and operation of the satellite, and personal, financial, stakeholder, or managerial risks are omitted. The mission success rate can be significantly improved when risks are adequately considered at stages preceding the launch and operation of the satellite.

Improving predictions and understanding of primary and ultimate biodegradation rates with machine learning models

This study aimed to develop machine learning based quantitative structure biodegradability relationship (QSBR) models for predicting primary and ultimate biodegradation rates of organic chemicals, which are essential parameters for environmental risk assessment. For this purpose, experimental primary and ultimate biodegradation rates of high consistency were compiled for 173 organic compounds. A significant number of descriptors were calculated with a collection of quantum/computational chemistry software and tools to achieve comprehensive representation and interpretability. Following a pre-screening process, multiple QSBR models were developed for both primary and ultimate endpoints using three algorithms: extreme gradient boosting (XGBoost), support vector machine (SVM), and multiple linear regression (MLR). Furthermore, a unified QSBR model was constructed using the knowledge transfer technique and XGBoost. Results demonstrated that all QSBR models developed in this study had good performance. Particularly, SVM models exhibited high level of goodness of fit (coefficient of determination on the training set of 0.973 for primary and 0.980 for ultimate), robustness (leave-one-out cross-validated coefficient of 0.953 for primary and 0.967 for ultimate), and external predictive ability (external explained variance of 0.947 for primary and 0.958 for ultimate). The knowledge transfer technique enhanced model performance by learning from properties of two biodegradation endpoints. Williams plots were used to visualize the application domains of the models. Through SHapley Additive exPlanations (SHAP) analysis, this study identified key features affecting biodegradation rates. Notably, MDEO-12, APC2D1_C_O, and other features contributed to primary biodegradation, while AATS0v, AATS2v, and others inhibited it. For ultimate biodegradation, features like No. of Rotatable Bonds, APC2D1_C_O, and minHBa were contributors, while C1SP3, Halogen Ratio, GGI4, and others hindered the process. Also, the study quantified the contributions of each feature in predictions for individual chemicals. This research provides valuable tools for predicting both primary and ultimate biodegradation rates while offering insights into the mechanisms.

Automated Customization of On-Device Inference for Quality-of-Experience Enhancement

The rapid uptake of intelligent applications is pushing deep learning (DL) capabilities to mobile devices. However, the heterogeneities in device capacity, DNN performances, and user preferences make it challenging to provide satisfactory Quality of Experience (QoE) to mobile users. This paper studies automated customization for DL inference on mobile devices (termed as on-device inference), and our goal is to enhance user QoE by configuring the on-device inference with an appropriate DNN for users under different usage scenarios. The core of our method is a DNN selection module that learns user QoE patterns on-the-fly and identifies the best-fit DNN for on-device inference with the learned knowledge. It leverages an online learning algorithm, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">NeuralUCB</i> , that has excellent generalization ability for handling various user QoE patterns. We also embed the knowledge transfer technique in NeuralUCB to expedite the learning process. However, NeuralUCB frequently solicits QoE ratings from users, which incurs non-negligible inconvenience. To address this problem, we design feedback solicitation schemes to reduce the number of QoE solicitations while maintaining the learning efficiency of NeuralUCB. A pragmatic problem, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">aggregated QoE</i> , is further investigated to improve the practicality of our framework. We conduct experiments on both synthetic and real-world data. The results indicate that our method efficiently learns the user QoE pattern with few solicitations and provides drastic QoE enhancement for mobile devices.

On the Use of Knowledge Transfer Techniques for Biomedical Named Entity Recognition

Biomedical named entity recognition (BioNER) is a preliminary task for many other tasks, e.g., relation extraction and semantic search. Extracting the text of interest from biomedical documents becomes more demanding as the availability of online data is increasing. Deep learning models have been adopted for biomedical named entity recognition (BioNER) as deep learning has been found very successful in many other tasks. Nevertheless, the complex structure of biomedical text data is still a challenging aspect for deep learning models. Limited annotated biomedical text data make it more difficult to train deep learning models with millions of trainable parameters. The single-task model, which focuses on learning a specific task, has issues in learning complex feature representations from a limited quantity of annotated data. Moreover, manually constructing annotated data is a time-consuming job. It is, therefore, vital to exploit other efficient ways to train deep learning models on the available annotated data. This work enhances the performance of the BioNER task by taking advantage of various knowledge transfer techniques: multitask learning and transfer learning. This work presents two multitask models (MTMs), which learn shared features and task-specific features by implementing the shared and task-specific layers. In addition, the presented trained MTM is also fine-tuned for each specific dataset to tailor it from a general features representation to a specialized features representation. The presented empirical results and statistical analysis from this work illustrate that the proposed techniques enhance significantly the performance of the corresponding single-task model (STM).

Total-body low-dose CT image denoising using a prior knowledge transfer technique with a contrastive regularization mechanism.

Reducing the radiation exposure experienced by patients in total-body computed tomography (CT) imaging has attracted extensive attention in the medical imaging community. A low radiation dose may result in increased noise and artifacts that greatly affect the subsequent clinical diagnosis. To obtain high-quality total-body low-dose CT (LDCT) images, previous deep learning-based research works developed various network architectures. However, most of these methods only employ normal-dose CT (NDCT) images as ground truths to guide the training process of the constructed denoising network. As a result of this simple restriction, the reconstructed images tend to lose favorable image details and easily generate oversmoothed textures. This study explores how to better utilize the information contained in the feature spaces of NDCT images to guide the LDCT image reconstruction process and achieve high-qualityresults. We propose a novel intratask knowledge transfer (KT) method that leverages the knowledge distilled from NDCT images as an auxiliary component of the LDCT image reconstruction process. Our proposed architecture is named the teacher-student consistency network (TSC-Net), which consists of teacher and student networks with identical architectures. By employing the designed KT loss, the student network is encouraged to emulate the teacher network in the representation space and gain robust prior content. In addition, to further exploit the information contained in CT scans, a contrastive regularization mechanism (CRM) built upon contrastive learning is introduced. The CRM aims to minimize and maximize the L2 distances from the predicted CT images to the NDCT samples and to the LDCT samples in the latent space, respectively. Moreover, based on attention and the deformable convolution approach, we design a dynamic enhancement module (DEM) to improve the network capability to transform input informationflows. By conducting ablation studies, we prove the effectiveness of the proposed KT loss, CRM, and DEM. Extensive experimental results demonstrate that the TSC-Net outperforms the state-of-the-art methods in both quantitative and qualitative evaluations. Additionally, the excellent results obtained for clinical readings also prove that our proposed method can reconstruct high-quality CT images for clinicalapplications. Based on the experimental results and clinical readings, the TSC-Net has better performance than other approaches. In our future work, we may explore the reconstruction of LDCT images by fusing the positron emission tomography (PET) and CT modalities to further improve the visual quality of the reconstructed CTimages.

Međuuniverzitetska saradnja Francuske i Bosne i Hercegovine - studentska mobilnost i uloga francuskog obrazovnog sistema u osnaživanju visokog obrazovanja u Bosni i Hercegovini

Higher education plays a key role in the development of a country and society. Cooperation between different universities and educational institutions at the international level is important for enriching the educational experience of students, sharing knowledge and innovation, and promoting cultural diversity. This scientific paper examines the scope of the influence of the French Republic and its educational model on the development of higher education in the Federation of Bosnia and Herzegovina. The specificity of Bosnia and Herzegovina is reflected in the multiculturalism that dominates, but also in the fact that this country has been constituting itself for the last few decades. In order to represent the interests and needs of different groups in the most successful ways, Bosnia and Herzegovina applied the existing examples of good practice from France in various social spheres, among others in the field of education. The French education system is one of the world's most successful higher education systems. In addition, the French state, through various incentives such as donations or scholarships to the best students, influenced the development and institutionalization of universities around the world. The basic hypothesis on which this work is based is that inter-university cooperation between France and Bosnia and Herzegovina contributes to the improvement of the quality of higher education, especially in terms of academic excellence and research capacities of Bosnia and Herzegovina. The method on the basis of which the hypothesis is tested is the method of technology and knowledge transfer techniques, and the comparative method will be used as an additional method.

Knowledge transfer for tuning microwave filters with unknown individual differences

Improving the production efficiency of Microwave Filters (MFs) is of great practical significance for constructing modern communication systems. The characteristics of MFs are with various degrees of individual differences caused by materials. The most tuning techniques have to start from scratch for tuning different MFs, which is time-consuming. The knowledge from previous tuning processes can be used in the current tuning due to the relevance of MFs. Motivated by this, a knowledge transfer method for tuning MFs with unknown individual differences is proposed. The main contributions are threefold: 1) An Knowledge Transfer technique for Optimization Tuning (KTOT) is created to tune various MFs efficiently; 2) an Evaluation mechanism of Difference Degree (EDD) is proposed to guide transfer; 3) a Guiding strategy of Knowledge Transfer Intensity (KTIG) in accordance with the difference degree is presented to reuse knowledge reasonably. The high-efficiency of KTOT and the effectiveness of KTIG based on EDD are demonstrated.

Automatic knowledge exchange between ontologies and semantic graphs

This article presents an innovative knowledge management approach for the validation and transfer of knowledge between semantic networks or ontologies and a target ontology represented in web ontology language (OWL) format. This process has been designed for addressing the quality improvement of ontologies automatically created by learning techniques. The knowledge transfer process is a semi-automatic computer aided method to assist the domain expert to improve the target ontology. To validate our proposal, we have used an automatically generated target ontology. We used knowledge transfer from the well-known Babelnet semantic graph and a manually generated ontology to improve the quality of the target ontology. Finally, to show the suitability of our proposal in the ontology-fixing process, we compare the improved target ontology resulting from the application of the proposed validation and knowledge transfer techniques with its original version. We developed an example of our proposal in our OntologyFixer tool, which is available on a GitHub repository ( https://github.com/gabyluna/OntologyFixer/tree/ontofixer_v2 ).

Dexterous robotic manipulation using deep reinforcement learning and knowledge transfer for complex sparse reward‐based tasks

AbstractThis paper describes a deep reinforcement learning (DRL) approach that won Phase 1 of the Real Robot Challenge (RRC) 2021, and then extends this method to a more difficult manipulation task. The RRC consisted of using a TriFinger robot to manipulate a cube along a specified positional trajectory, but with no requirement for the cube to have any specific orientation. We used a relatively simple reward function, a combination of a goal‐based sparse reward and a distance reward, in conjunction with Hindsight Experience Replay (HER) to guide the learning of the DRL agent (Deep Deterministic Policy Gradient [DDPG]). Our approach allowed our agents to acquire dexterous robotic manipulation strategies in simulation. These strategies were then deployed on the real robot and outperformed all other competition submissions, including those using more traditional robotic control techniques, in the final evaluation stage of the RRC. Here we extend this method, by modifying the task of Phase 1 of the RRC to require the robot to maintain the cube in a particular orientation, while the cube is moved along the required positional trajectory. The requirement to also orient the cube makes the agent less able to learn the task through blind exploration due to increased problem complexity. To circumvent this issue, we make novel use of a Knowledge Transfer (KT) technique that allows the strategies learned by the agent in the original task (which was agnostic to cube orientation) to be transferred to this task (where orientation matters). KT allowed the agent to learn and perform the extended task in the simulator, which improved the average positional deviation from 0.134 to 0.02 m, and average orientation deviation from 142° to 76° during evaluation. This KT concept shows good generalization properties and could be applied to any actor‐critic learning algorithm.

Effect of Use of Tacit Knowledge Transfer Techniques on Organizational Performance of Kenya Agricultural and Livestock Research Organization Researchers

Tacit knowledge is key in managing the performance of agricultural research organizations. This study analyzed the adequacy of tacit knowledge transfer techniques, how tacit knowledge transfer enables achievement of performance indicators of Kenya Agricultural and Livestock Research Organization (KALRO), and the nature of association between tacit knowledge transfer and researchers' performance. The study adopted a descriptive survey research design, used cluster sampling and a semi-structured questionnaire to collect data from 191 respondents in KALRO research centers. Data were analyzed using the Likert scale and Chi-square in Statistical Package for Social Sciences (SPSS) version 20. Collaborative research, workshops and seminars are the most adequate techniques for transferring tacit knowledge among agricultural researchers and enhancing their performance. Cognitive Self-Motivation, collective and local tacit knowledge are the most useful types of tacit knowledge in enhancing agricultural researchers' performance. Management of agricultural research projects and writing of research fund-winning proposals were the performance indicators that researchers were most enabled to attain by tacit knowledge. Chi-square showed that there was significant association between the types of tacit knowledge and performance of researchers. In conclusion, use of tacit knowledge transfer techniques enables researchers to achieve their organization’s performance indicators and a significant positive association exists between tacit knowledge and researchers’ performance. KALRO needs to encourage more use of the most employed techniques in management of agricultural research projects and writing of research fund winning proposals through workshops, seminars and its knowledge management policy.

Evaluation of transfer learning models for predicting the lateral strength of reinforced concrete columns

Transfer learning aims to extract knowledge from one or more source tasks and apply the knowledge to a different task for more accurate predictions. The main purposes of this study are to investigate different knowledge transfer techniques, apply them to accurately predict the lateral strength of reinforced concrete columns with only a small amount of training data, and compare the transferability of each method. According to the various source and target domains, three different experiments are designed to directly compare the performance across section type, shear reinforcement area, and concrete compressive strength. In all cases in this study, knowledge transfer techniques show better prediction performance than the models trained without any knowledge transfer techniques. Therefore, we can conclude that transferring pre-trained knowledge from the source domain enables a model to better explain the response variable in the target domain. The performance improvement is particularly emphasized when the available data for the target domain is small. Thus, transfer learning can be one way to address the data scarcity problem in structural engineering. Furthermore, transferring the pre-trained knowledge is more associated with the underlying physical relationship between the source and the target domains and less associated with the discrepancy between the source and the target domain distributions.

A Computer Vision Approach to Identifying Ticks Related to Lyme Disease.

Background: Lyme disease (caused by Borrelia burgdorferi) is an infectious disease transmitted to humans by a bite from infected blacklegged ticks (Ixodes scapularis) in eastern North America. Lyme disease can be prevented if antibiotic prophylaxis is given to a patient within 72 hours of a blacklegged tick bite. Therefore, recognizing a blacklegged tick could facilitate the management of Lyme disease. Methods: In this work, we build an automated detection tool that can differentiate blacklegged ticks from other tick species using advanced computer vision approaches in real-time. Specially, we use convolution neural network models, trained end-to-end, to classify tick species. Also, advanced knowledge transfer techniques are adopted to improve the performance of convolution neural network models. Results: Our best convolution neural network model achieves 92% accuracy on unseen tick species. Conclusion: Our proposed vision-based approach simplifies tick identification and contributes to the emerging work on public health surveillance of ticks and tick-borne diseases. In addition, it can be integrated with the geography of exposure and potentially be leveraged to inform the risk of Lyme disease infection. This is the first report of using deep learning technologies to classify ticks, providing the basis for automation of tick surveillance, and advancing tick-borne disease ecology and risk management.

Model compression and simplification pipelines for fast deep neural network inference in FPGAs in HEP

Resource utilization plays a crucial role for successful implementation of fast real-time inference for deep neural networks (DNNs) and convolutional neural networks (CNNs) on latest generation of hardware accelerators (FPGAs, SoCs, ACAPs, GPUs). To fulfil the needs of the triggers that are in development for the upgraded LHC detectors, we have developed a multi-stage compression approach based on conventional compression strategies (pruning and quantization) to reduce the memory footprint of the model and knowledge transfer techniques, crucial to streamline the DNNs simplifying the synthesis phase in the FPGA firmware and improving explainability. We present the developed methodologies and the results of the implementation in a working engineering pipeline used as pre-processing stage to high level synthesis tools (HLS4ML, Xilinx Vivado HLS, etc.). We show how it is possible to build ultra-light deep neural networks in practice, by applying the method to a realistic HEP use-case: a toy simulation of one of the triggers planned for the HL-LHC.

Probabilistic Knowledge Transfer for Lightweight Deep Representation Learning.

Knowledge-transfer (KT) methods allow for transferring the knowledge contained in a large deep learning model into a more lightweight and faster model. However, the vast majority of existing KT approaches are designed to handle mainly classification and detection tasks. This limits their performance on other tasks, such as representation/metric learning. To overcome this limitation, a novel probabilistic KT (PKT) method is proposed in this article. PKT is capable of transferring the knowledge into a smaller student model by keeping as much information as possible, as expressed through the teacher model. The ability of the proposed method to use different kernels for estimating the probability distribution of the teacher and student models, along with the different divergence metrics that can be used for transferring the knowledge, allows for easily adapting the proposed method to different applications. PKT outperforms several existing state-of-the-art KT techniques, while it is capable of providing new insights into KT by enabling several novel applications, as it is demonstrated through extensive experiments on several challenging data sets.