Unsupervised Incremental Learning Research Articles

An enhanced structural developmental neural network with information saturation for continual unsupervised learning

In this paper, an enhanced structural developmental neural network with information saturation (ESDNNIS) is proposed for continual unsupervised learning. It improves SDNNIS in the following respects: (1) it makes a distinction between the developmental stages of neurons to better regulate the updating of neurons; (2) it adopts the forgetting mechanism to control the network size; (3) it changes the way coverage domain parameter is initialized and updated to divide clusters more accurately; and (4) it adds a parent–child relationship update mechanism to provide better robustness to sample input order. ESDNNIS inherits the unsupervised incremental learning and class segmentation capabilities of SDNNIS, and achieves better class incremental clustering for more complex datasets while using fewer neurons than SDNNIS. The experimental results show that the accuracy of clustering results of ESDNNIS is improved by 10.94% on average compared with SDNNIS, while the number of generated neurons is reduced by 71.93% on average.

Unsupervised incremental transfer learning with knowledge distillation for online remaining useful life prediction of rotating machinery

Online remaining useful life (RUL) prediction has been solved by deep transfer learning, but is still with challenges as follows: (1) Incomplete and unlabeled data under actual operation; (2) Condition monitoring data is streaming with unknown distribution; and (3) The distribution of degradation data is variable. To solve them, an unsupervised incremental transfer learning approach with knowledge distillation (KD) is proposed. First, a time series recursive prediction model is built to generate pseudo labels. Second, an online KD network is constructed to realize unsupervised domain adaptation. Finally, an incremental updating mechanism is designed in the online KD network for online RUL prediction with the pseudo labels. Comparative experiments on the IEEE PHM Challenge 2012 rolling bearing dataset show that the proposed method, being computationally inexpensive, can effectively achieve dynamic prediction only with sequentially-collected data, which can provide an effective RUL prediction solution for rotating machinery under an open environment.

Hardware implementation of an active learning self-organizing neural network to predict the power fluctuation events of a photovoltaic grid-tied system

Power fluctuations happens due to highly stochastic solar irradiance nature and this could cause unwanted power quality issues in photovoltaic (PV) system. First predicting the PV output then mitigating the fluctuation by using energy storage system is one of the most effective power fluctuation mitigation method. Incremental learning algorithm such as incremental self-organizing map can be used to actively learn and predict power output. Although significant improvement in terms of mitigated events can be achieved in this kind of system, it still suffers from prediction accuracy due to the highly fluctuating PV profile. In this paper, a time-series unsupervised learning algorithm namely the Time-Series Self-Organizing Incremental Neural Network (TS-SOINN) is proposed to better predict the highly stochastic PV output. It incorporates a novel weighted memory layer to give higher emphasis to recent observation and improves data overlap issues in the conventional self-organizing map algorithm. In the simulation results, the TS-SOINN achieves prediction rate of 93.81% which outperforms the latest unsupervised incremental learning algorithm, M-SOINN by 33.44% and the TS-SOINN mitigated 89.13% power fluctuation events whereas M-SOINN only mitigated 79.62% events. In addition, the TS-SOINN is implemented in Altera Stratix V GS Field Programmable Gate Array (FPGA) board to run real-time prediction. This hardware architecture is able to increase new node and remove inactive nodes throughout the real-time prediction. The hardware TS-SOINN is integrated with mitigation engine to smoothen out power fluctuation events in PV grid-tied system, the experimental results show that the proposed system mitigated 83.33% of power fluctuation events at the grid and the battery state-of-charge maintains within 30%-100%.

The Ensemble of Unsupervised Incremental Learning Algorithm for Time Series Data

Data mining is one of the key concepts to discover hidden knowledge from available data. Along with the data mining, data analytics is a field to analyze and process data in a scientific and cognitive angle. It is more helpful to convert knowledge to actionable knowledge for accurate decision making. Data Stream Mining is another challenging area than normal Data Mining due to its dynamics. Dynamics of data in a stream includes changes in data frequency, volume and nature. This paper focuses on the behavior of data mining of machines in process/manufacturing industries. In general, such data is continuous numerical and time series data captured by various industrial sensors. By nature, equipment or machinery behaviour can change over time. It requires calibration/replacement before failure of machinery. By analyzing data, one can find the behavior or state change. To identify that, dynamic models are required to be built using data mining and data stream mining. Thus, we are following a semi-novel approach for building such models using “Ensemble of Unsupervised Incremental Learning" method. Results show how the existing methods are different from the proposed method. This method can be applied for any other domain like image/audio/video or text mining.

Towards a self-sufficient face verification system

The absence of a previous collaborative manual enrolment represents a significant handicap towards designing a face verification system for face re-identification purposes. In this scenario, the system must learn the target identity incrementally, using data from the video stream during the operational authentication phase. So, manual labelling cannot be assumed apart from the first few frames. On the other hand, even the most advanced methods trained on large-scale and unconstrained datasets suffer performance degradation when no adaptation to specific contexts is performed. This work proposes an adaptive face verification system, for the continuous re-identification of target identity, within the framework of incremental unsupervised learning. Our Dynamic Ensemble of SVM is capable of incorporating non-labelled information to improve the performance of any model, even when its initial performance is modest. The proposal uses the self-training approach and is compared against other classification techniques within this same approach. Results show promising behaviour in terms of both knowledge acquisition and impostor robustness.

Learning Representations in Model-Free Hierarchical Reinforcement Learning

Common approaches to Reinforcement Learning (RL) are seriously challenged by large-scale applications involving huge state spaces and sparse delayed reward feedback. Hierarchical Reinforcement Learning (HRL) methods attempt to address this scalability issue by learning action selection policies at multiple levels of temporal abstraction. Abstraction can be had by identifying a relatively small set of states that are likely to be useful as subgoals, in concert with the learning of corresponding skill policies to achieve those subgoals. Many approaches to subgoal discovery in HRL depend on the analysis of a model of the environment, but the need to learn such a model introduces its own problems of scale. Once subgoals are identified, skills may be learned through intrinsic motivation, introducing an internal reward signal marking subgoal attainment. We present a novel model-free method for subgoal discovery using incremental unsupervised learning over a small memory of the most recent experiences of the agent. When combined with an intrinsic motivation learning mechanism, this method learns subgoals and skills together, based on experiences in the environment. Thus, we offer an original approach to HRL that does not require the acquisition of a model of the environment, suitable for large-scale applications. We demonstrate the efficiency of our method on a variant of the rooms environment.

An incremental unsupervised learning based trajectory controller for a 4 wheeled skid steer mobile robot

This paper proposes a trajectory controller for a 4-wheel skid steering mobile robot designed for use in an oil palm plantation. The nature of the working environment requires adaptive control to eliminate noise and to learn necessary variations on-the-go. The proposed control system is based on the Enhanced Self Organizing Incremental Neural Network (ESOINN), and is able to produce exceptional trajectory control without the use of a kinematic / dynamic model of the mobile robot by training the network with measured trajectory data as well as simulated data by incremental learning. Our simulation results show that the ESOINN is able to adapt to new training samples and errors have been reduced after only a few iterations of incremental learning. The RMSE error of the output of the initial network was reduced by almost 50% after 3 stages of incremental learning. When comparing training times, ESOINN had a much faster computation time with each consecutive incremental learning instance as compared to other non-incremental methods such as self-organizing maps (SOM), K-means clustering and an adaptive Neural Network. In addition, ESOINN produced improved performance after each consecutive stage of learning, proving its reliability, unlike the other mentioned methods, which gave varied performance during each stage.

An adaptive algorithm for anomaly and novelty detection in evolving data streams

In the era of big data, considerable research focus is being put on designing efficient algorithms capable of learning and extracting high-level knowledge from ubiquitous data streams in an online fashion. While, most existing algorithms assume that data samples are drawn from a stationary distribution, several complex environments deal with data streams that are subject to change over time. Taking this aspect into consideration is an important step towards building truly aware and intelligent systems. In this paper, we propose GNG-A, an adaptive method for incremental unsupervised learning from evolving data streams experiencing various types of change. The proposed method maintains a continuously updated network (graph) of neurons by extending the Growing Neural Gas algorithm with three complementary mechanisms, allowing it to closely track both gradual and sudden changes in the data distribution. First, an adaptation mechanism handles local changes where the distribution is only non-stationary in some regions of the feature space. Second, an adaptive forgetting mechanism identifies and removes neurons that become irrelevant due to the evolving nature of the stream. Finally, a probabilistic evolution mechanism creates new neurons when there is a need to represent data in new regions of the feature space. The proposed method is demonstrated for anomaly and novelty detection in non-stationary environments. Results show that the method handles different data distributions and efficiently reacts to various types of change.

Automatic event detection in microblogs using incremental machine learning

The global popularity of microblogs has led to an increasing accumulation of large volumes of text data on microblogging platforms such as Twitter. These corpora are untapped resources to understand social expressions on diverse subjects. Microblog analysis aims to unlock the value of such expressions by discovering insights and events of significance hidden among swathes of text. Besides velocity; diversity of content, brevity, absence of structure and time‐sensitivity are key challenges in microblog analysis. In this paper, we propose an unsupervised incremental machine learning and event detection technique to address these challenges. The proposed technique separates a microblog discussion into topics to address the key problem of diversity. It maintains a record of the evolution of each topic over time. Brevity, time‐sensitivity and unstructured nature are addressed by these individual topic pathways which contribute to generate a temporal, topic‐driven structure of a microblog discussion. The proposed event detection method continuously monitors these topic pathways using multiple domain‐independent event indicators for events of significance. The autonomous nature of topic separation, topic pathway generation, new topic identification and event detection, appropriates the proposed technique for extensive applications in microblog analysis. We demonstrate these capabilities on tweets containing #microsoft and tweets containing #obama.

A Self-Organizing Incremental Neural Network based on local distribution learning

In this paper, we propose an unsupervised incremental learning neural network based on local distribution learning, which is called Local Distribution Self-Organizing Incremental Neural Network (LD-SOINN). The LD-SOINN combines the advantages of incremental learning and matrix learning. It can automatically discover suitable nodes to fit the learning data in an incremental way without a priori knowledge such as the structure of the network. The nodes of the network store rich local information regarding the learning data. The adaptive vigilance parameter guarantees that LD-SOINN is able to add new nodes for new knowledge automatically and the number of nodes will not grow unlimitedly. While the learning process continues, nodes that are close to each other and have similar principal components are merged to obtain a concise local representation, which we call a relaxation data representation. A denoising process based on density is designed to reduce the influence of noise. Experiments show that the LD-SOINN performs well on both artificial and real-word data.

Разработка алгоритмов технического зрения для обеспечения инкрементного неконтролируемого обучения в задачах детектирования образов движущихся сложноструктурированных объектов

Разработка алгоритмов технического зрения для обеспечения инкрементного неконтролируемого обучения в задачах детектирования образов движущихся сложноструктурированных объектов

Unsupervised Incremental Learning of Associative Cubes with Orthogonal Kernels

An 'associative cube', a class of auto-associative memories, is revisited here, in which training data and hidden orthogonal basis functions such as wavelet packets or Fourier kernels, are combined in the weight cube. This weight cube has hidden units in its depth, represented by a three dimensional cubic structure. We develop an unsupervised incremental learning mechanism based upon the adaptive least squares method. Training data are mapped into orthogonal basis vectors in a least-squares sense by updating the weights which minimize an energy function. Therefore, a prescribed orthogonal kernel is incrementally assigned to an incoming data. Next, we show how a decoding procedure finds the closest one with a competitive network in the hidden layer. As noisy test data are applied to an associative cube, the nearest one among the original training data are restored in an optimal sense. The simulation results confirm robustness of associative cubes even if test data are heavily distorted by various types of noise.

Incrementally learning objects by touch: online discriminative and generative models for tactile-based recognition.

Human beings not only possess the remarkable ability to distinguish objects through tactile feedback but are further able to improve upon recognition competence through experience. In this work, we explore tactile-based object recognition with learners capable of incremental learning. Using the sparse online infinite Echo-State Gaussian process (OIESGP), we propose and compare two novel discriminative and generative tactile learners that produce probability distributions over objects during object grasping/palpation.To enable iterative improvement, our online methods incorporate training samples as they become available. We also describe incremental unsupervised learning mechanisms, based on novelty scores and extreme value theory, when teacher labels are not available. We present experimental results for both supervised and unsupervised learning tasks using the iCub humanoid, with tactile sensors on its five-fingered anthropomorphic hand, and 10 different object classes. Our classifiers perform comparably to state-of-the-art methods (C4.5 and SVM classifiers) and findings indicate that tactile signals are highly relevant for making accurate object classifications. We also show that accurate "early" classifications are possible using only 20-30 percent of the grasp sequence.For unsupervised learning, our methods generate high quality clusterings relative to the widely-used sequential k-means and self-organising map (SOM), and we present analyses into the differences between the approaches.

A Load-Balancing Self-Organizing Incremental Neural Network

Clustering is widely used in machine learning, feature extraction, pattern recognition, image analysis, information retrieval, and bioinformatics. Online unsupervised incremental learning is an important branch of data clustering. However, accurately separating high-density overlapped areas in a network has a direct impact on the performance of the clustering algorithm. In this paper, we propose a load-balancing self-organizing incremental neural network (LB-SOINN) to achieve good clustering results and demonstrate that it is more stable than an enhanced SOINN (E-SOINN). LB-SOINN has all the advantages of E-SOINN, such as robustness to noise and online unsupervised incremental learning. It overcomes the shortcomings of the topology structure generated by E-SOINN, such as dependence on the sequence of the input data, and avoids the turbulence that occurs when separating a composite class into subclasses. Furthermore, we also introduce a distance combination framework to obtain good performance for high-dimensional space-clustering tasks. Experiments involving both artificial and real world data sets indicate that LB-SOINN has superior performance in comparison with E-SOINN and other methods.

Using unsupervised incremental learning to cope with gradual concept drift

Current computational approach to incremental learning requires a constant stream of labelled data to cope with gradual environmental changes known as concept drift. This paper examines a case where labelled data are unavailable. Inspired by the performance of the human visual system, capable of adjusting its concepts using unlabelled stimuli, we introduce a variant to an unsupervised competitive learning algorithm known as the Leader Follower (LF). This variant can adjust pre-learned concepts to environmental changes using unlabelled data samples. We motivate the needed change in the existing LF algorithm and compare between two variants to enable the accumulation of environmental changes when facing unbalanced sample ratio.

An automatic patient-specific seizure onset detection method in intracranial EEG based on incremental nonlinear dimensionalityreduction

Epileptic seizure features always include the morphology and spatial distribution of nonlinear waveforms in the electroencephalographic (EEG) signals. In this study, we propose a novel incremental learning scheme based on nonlinear dimensionality reduction for automatic patient-specific seizure onset detection. The method allows for identification of seizure onset times in long-term EEG signals acquired from epileptic patients. Firstly, a nonlinear dimensionality reduction (NDR) method called local tangent space alignment (LTSA) is used to reduce the dimensionality of available initial feature sets extracted with continuous wavelet transform (CWT). One-dimensional manifold which reflects the intrinsic dynamics of seizure onset is obtained. For each patient, IEEG recordings containing one seizure onset is sufficient to train the initial one-dimensional manifold. Secondly, an unsupervised incremental learning scheme is proposed to update the initial manifold when the unlabelled EEG segments flow in sequentially. The incremental learning scheme can cluster the new coming samples into the trained patterns (containing or not containing seizure onsets). Intracranial EEG recordings from 21 patients with duration of 193.8 h and 82 seizures are used for the evaluation of the method. Average sensitivity of 98.8%, average uninteresting false positive rate of 0.24/h, average interesting false positives rate of 0.25/h, and average detection delay of 10.8 s are obtained. Our method offers simple, accurate training with less human intervening and can be well used in off-line seizure detection. The unsupervised incremental learning scheme has the potential in identifying novel IEEG classes (different onset patterns) within thedata.

遺伝的アルゴリズムを適用したＦｕｚｚｙ ＡＲＴによる２リンクロボットアームの行動獲得の効率化

An efficient learning method using Fuzzy ART with Genetic Algorithm is proposed. The proposed method reduces the number of trials by using a policy acquired in other tasks because a reinforcement learning needs a lot of the number of trials until an agent acquires appropriate actions. Fuzzy ART is an incremental unsupervised learning algorithm in responce to arbitrary sequences of analog or binary input vectors. Our proposed method gives a policy by crossover or mutation when an agent observes unknown states. Selection controls the category proliferation problem of Fuzzy ART. The effectiveness of the proposed method was verified with the simulation of the reaching problem for the two-link robot arm. The proposed method achieves a reduction of both the number of trials and the number of states.

Experience-driven formation of parts-based representations in a model of layered visual memory

Growing neuropsychological and neurophysiological evidence suggests that the visual cortex uses parts-based representations to encode, store and retrieve relevant objects. In such a scheme, objects are represented as a set of spatially distributed local features, or parts, arranged in stereotypical fashion. To encode the local appearance and to represent the relations between the constituent parts, there has to be an appropriate memory structure formed by previous experience with visual objects. Here, we propose a model how a hierarchical memory structure supporting efficient storage and rapid recall of parts-based representations can be established by an experience-driven process of self-organization. The process is based on the collaboration of slow bidirectional synaptic plasticity and homeostatic unit activity regulation, both running at the top of fast activity dynamics with winner-take-all character modulated by an oscillatory rhythm. These neural mechanisms lay down the basis for cooperation and competition between the distributed units and their synaptic connections. Choosing human face recognition as a test task, we show that, under the condition of open-ended, unsupervised incremental learning, the system is able to form memory traces for individual faces in a parts-based fashion. On a lower memory layer the synaptic structure is developed to represent local facial features and their interrelations, while the identities of different persons are captured explicitly on a higher layer. An additional property of the resulting representations is the sparseness of both the activity during the recall and the synaptic patterns comprising the memory traces.

Identification of evolving fuzzy rule-based models

An approach to identification of evolving fuzzy rule-based (eR) models is proposed. eR models implement a method for the noniterative update of both the rule-base structure and parameters by incremental unsupervised learning. The rule-base evolves by adding more informative rules than those that previously formed the model. In addition, existing rules can be replaced with new rules based on ranking using the informative potential of the data. In this way, the rule-base structure is inherited and updated when new informative data become available, rather than being completely retrained. The adaptive nature of these evolving rule-based models, in combination with the highly transparent and compact form of fuzzy rules, makes them a promising candidate for modeling and control of complex processes, competitive to neural networks. The approach has been tested on a benchmark problem and on an air-conditioning component modeling application using data from an installation serving a real building. The results illustrate the viability and efficiency of the approach.

A statistical model for word discovery in transcribed speech

A statistical model for segmentation and word discovery in continuous speech is presented. An incremental unsupervised learning algorithm to infer word boundaries based on this model is described. ...