Search results for: memory network

Authors: Tet H. Yeap

Abstract:

An analog restricted Hopfield Network is presented in this paper. It consists of two layers of nodes, visible and hidden nodes, connected by directional weighted paths forming a bipartite graph with no intralayer connection. An energy or Lyapunov function was derived to show that the proposed network will converge to stable states. By introducing hidden nodes, the proposed network can be trained to store patterns and has increased memory capacity. Training to be an associative memory, simulation results show that the associative memory performs better than a classical Hopfield network by being able to perform better memory recall when the input is noisy.

Keywords: restricted Hopfield network, Lyapunov function, simultaneous perturbation stochastic approximation

Procedia PDF Downloads 134

Authors: Motonobu Hattori

Abstract:

In neural networks, when new patters are learned by a network, they radically interfere with previously stored patterns. This drawback is called catastrophic forgetting. We have already proposed a biologically inspired dual-network memory model which can much reduce this forgetting for static patterns. In this model, information is first stored in the hippocampal network, and thereafter, it is transferred to the neocortical network using pseudo patterns. Because, temporal sequence learning is more important than static pattern learning in the real world, in this study, we improve our conventional dual-network memory model so that it can deal with temporal sequences without catastrophic forgetting. The computer simulation results show the effectiveness of the proposed dual-network memory model.

Keywords: catastrophic forgetting, dual-network, temporal sequences, hippocampal

Procedia PDF Downloads 273

Authors: Danilo López, Nelson Vera, Luis Pedraza

Abstract:

This paper analyzes fundamental ideas and concepts related to neural networks, which provide the reader a theoretical explanation of Long Short-Term Memory (LSTM) networks operation classified as Deep Learning Systems, and to explicitly present the mathematical development of Backward Pass equations of the LSTM network model. This mathematical modeling associated with software development will provide the necessary tools to develop an intelligent system capable of predicting the behavior of licensed users in wireless cognitive radio networks.

Keywords: neural networks, multilayer perceptron, long short-term memory, recurrent neuronal network, mathematical analysis

Procedia PDF Downloads 421

Authors: Mathew Wakefield, Matthew Mitchell, Lisa Wise, Christopher McCarthy

Abstract:

The properties of memory representations in artificial neural networks have cognitive implications. Distributed representations that encode instances as a pattern of activity across layers of nodes afford memory compression and enforce the selection of a single point in instance space. These encoding schemes also appear to distort the representational space, as well as trading off the ability to validate that input information is within the bounds of past experience. In contrast, a localist representation which encodes some meaningful information into individual nodes in a network layer affords less memory compression while retaining the integrity of the representational space. This allows the validity of an input to be determined. The validity (or familiarity) of input along with the capacity of localist representation for multiple instance selections affords a memory sampling approach that dynamically balances the bias-variance trade-off. When the input is familiar, bias may be high by referring only to the most similar instances in memory. When the input is less familiar, variance can be increased by referring to more instances that capture a broader range of features. Using this approach in a localist instance memory network, an experiment demonstrates a relationship between representational conflict, generalization performance, and memorization demand. Relatively small sampling ranges produce the best performance on a classic machine learning dataset of visual objects. Combining memory validity with conflict detection produces a reliable confidence judgement that can separate responses with high and low error rates. Confidence can also be used to signal the need for supervisory input. Using this judgement, the need for supervised learning as well as memory encoding can be substantially reduced with only a trivial detriment to classification performance.

Keywords: artificial neural networks, representation, memory, conflict monitoring, confidence

Procedia PDF Downloads 129

Authors: Yong Chen

Abstract:

To understand the detailed molecular mechanisms of memory formation in engram cells is one of the most fundamental questions in neuroscience. Recent single-cell RNA-seq (scRNA-seq) and single-nucleus RNA-seq (snRNA-seq) techniques have allowed us to explore the sparsely activated engram ensembles, enabling access to the molecular mechanisms that underlie experience-dependent memory formation and consolidation. However, the absence of specific and powerful computational methods to detect memory-related genes (modules) and their regulatory relationships in the sc/snRNA-seq datasets has strictly limited the analysis of underlying mechanisms and memory coding principles in mammalian brains. Here, we present a deep-learning method named SCENTBOX, to detect memory-related gene modules and causal regulatory relationships among themfromsc/snRNA-seq datasets. SCENTBOX first constructs codifferential expression gene network (CEGN) from case versus control sc/snRNA-seq datasets. It then detects the highly correlated modules of differential expression genes (DEGs) in CEGN. The deep network embedding and attention-based convolutional neural network strategies are employed to precisely detect regulatory relationships among DEG genes in a module. We applied them on scRNA-seq datasets of TRAP; Ai14 mouse neurons with fear memory and detected not only known memory-related genes, but also the modules and potential causal regulations. Our results provided novel regulations within an interesting module, including Arc, Bdnf, Creb, Dusp1, Rgs4, and Btg2. Overall, our methods provide a general computational tool for processing sc/snRNA-seq data from case versus control studie and a systematic investigation of fear-memory-related gene modules.

Keywords: sc/snRNA-seq, memory formation, deep learning, gene module, causal inference

Procedia PDF Downloads 120

Authors: Yin Yuanling

Abstract:

A Deep Feedback Recurrent Neural Network (DFRNN) and Bidirectional Long Short-Term Memory (BiLSTM) are designed to address the problem of low accuracy of traditional relationship extraction models. This method combines a deep feedback-based recurrent neural network (DFRNN) with a bi-directional long short-term memory (BiLSTM) approach. The method combines DFRNN, which extracts local features of text based on deep feedback recurrent mechanism, BiLSTM, which better extracts global features of text, and Self-Attention, which extracts semantic information. Experiments show that the method achieves an F1 value of 76.69% on the CEC dataset, which is 0.0652 better than the BiLSTM+Self-ATT model, thus optimizing the performance of the deep learning method in the event relationship extraction task.

Keywords: event relations, deep learning, DFRNN models, bi-directional long and short-term memory networks

Procedia PDF Downloads 146

Authors: Nagaratna Rajur, J. D. Mallapur, Y. B. Kirankumar

Abstract:

The security in mobile communication is very different from the internet or telecommunication, because of its poor user interface and limited processing capacity, as well as combination of complex network protocols. Hence, it poses a challenge for less memory usage and low computation speed based security system. Security involves all the activities that are undertaken to protect the value and on-going usability of assets and the integrity and continuity of operations. An effective network security strategies requires identifying threats and then choosing the most effective set of tools to combat them. Cryptography is a simple and efficient way to provide security in communication. RSA is an asymmetric key approach that is highly reliable and widely used in internet communication. However, it has not been efficiently implemented in mobile communication due its computational complexity and large memory utilization. The proposed algorithm modifies the current RSA to be useful in mobile communication by reducing its computational complexity and memory utilization.

Keywords: M-RSA, sensor networks, sensor applications, security

Procedia PDF Downloads 343

Authors: Ian O'Loughlin

Abstract:

Memory networks aim to integrate some of the recent successes in machine learning with a dynamic memory base that can be updated and deployed in artificial reasoning tasks. These models involve training networks to identify, update, and operate over stored elements in a large memory array in order, for example, to ably perform question and answer tasks parsing real-world and simulated discourses. This family of approaches still faces numerous challenges: the performance of these network models in simulated domains remains considerably better than in open, real-world domains, wide-context cues remain elusive in parsing words and sentences, and even moderately complex sentence structures remain problematic. This innovation, employing an array of stored and updatable ‘memory’ elements over which the system operates as it parses text input and develops responses to questions, is a compelling one for at least two reasons: first, it addresses one of the difficulties that standard machine learning techniques face, by providing a way to store a large bank of facts, offering a way forward for the kinds of long-term reasoning that, for example, recurrent neural networks trained on a corpus have difficulty performing. Second, the addition of a stored long-term memory component in artificial reasoning seems psychologically plausible; human reasoning appears replete with invocations of long-term memory, and the stored but dynamic elements in the arrays of memory networks are deeply reminiscent of the way that human memory is readily and often characterized. However, this apparent psychological plausibility is belied by a recent turn in the study of human memory in cognitive science. In recent years, the very notion that there is a stored element which enables remembering, however dynamic or reconstructive it may be, has come under deep suspicion. In the wake of constructive memory studies, amnesia and impairment studies, and studies of implicit memory—as well as following considerations from the cognitive neuroscience of memory and conceptual analyses from the philosophy of mind and cognitive science—researchers are now rejecting storage and retrieval, even in principle, and instead seeking and developing models of human memory wherein plasticity and dynamics are the rule rather than the exception. In these models, storage is entirely avoided by modeling memory using a recurrent neural network designed to fit a preconceived energy function that attains zero values only for desired memory patterns, so that these patterns are the sole stable equilibrium points in the attractor network. So although the array of long-term memory elements in memory networks seem psychologically appropriate for reasoning systems, they may actually be incurring difficulties that are theoretically analogous to those that older, storage-based models of human memory have demonstrated. The kind of emergent stability found in the attractor network models more closely fits our best understanding of human long-term memory than do the memory network arrays, despite appearances to the contrary.

Keywords: artificial reasoning, human memory, machine learning, neural networks

Procedia PDF Downloads 272

Authors: Pingping Lin, Xudong Luo, Yifan Fan

Abstract:

Sentiment analysis is a very active research topic. Every day, Facebook, Twitter, Weibo, and other social media, as well as significant e-commerce websites, generate a massive amount of comments, which can be used to analyse peoples opinions or emotions. The existing methods for sentiment analysis are based mainly on sentiment dictionaries, machine learning, and deep learning. The first two kinds of methods rely on heavily sentiment dictionaries or large amounts of labelled data. The third one overcomes these two problems. So, in this paper, we focus on the third one. Specifically, we survey various sentiment analysis methods based on convolutional neural network, recurrent neural network, long short-term memory, deep neural network, deep belief network, and memory network. We compare their futures, advantages, and disadvantages. Also, we point out the main problems of these methods, which may be worthy of careful studies in the future. Finally, we also examine the application of deep learning in multimodal sentiment analysis and aspect-level sentiment analysis.

Keywords: document analysis, deep learning, multimodal sentiment analysis, natural language processing

Procedia PDF Downloads 164

Authors: Ci Lin, Tet Yeap, Iluju Kiringa

Abstract:

This paper introduces the subspace rotation algorithm (SRA) to train the Restricted Hopfield Network (RHN) as an auto-associative memory. Subspace rotation algorithm is a gradient-free subspace tracking approach based on the singular value decomposition (SVD). In comparison with Backpropagation Through Time (BPTT) on training RHN, it is observed that SRA could always converge to the optimal solution and BPTT could not achieve the same performance when the model becomes complex, and the number of patterns is large. The AUTS case study showed that the RHN model trained by SRA could achieve a better structure of attraction basin with larger radius(in general) than the Hopfield Network(HNN) model trained by Hebbian learning rule. Through learning 10000 patterns from MNIST dataset with RHN models with different number of hidden nodes, it is observed that an several components could be adjusted to achieve a balance between recovery accuracy and noise resistance.

Keywords: hopfield neural network, restricted hopfield network, subspace rotation algorithm, hebbian learning rule

Procedia PDF Downloads 120

Authors: Deon de Jager, Yahya Zweiri, Dimitrios Makris

Abstract:

The three most important components in the cognitive architecture for cognitive robotics is memory representation, memory recall, and action-selection performed by the executive. In this paper, action selection, performed by the executive, is defined as a memory quantification and optimization process. The methodology describes the real-time construction of episodic memory through semantic memory optimization. The optimization is performed by set-based particle swarm optimization, using an adaptive entropy memory quantification approach for fitness evaluation. The performance of the approach is experimentally evaluated by simulation, where a UAV is tasked with the collection and delivery of a medical package. The experiments show that the UAV dynamically uses the episodic memory to autonomously control its velocity, while successfully completing its mission.

Keywords: cognitive robotics, semantic memory, episodic memory, maximum entropy principle, particle swarm optimization

Procedia PDF Downloads 158

Authors: Gagnesh Kumar, Nitin Gupta

Abstract:

Ternary Content Addressable Memory cells are mainly popular in network routers for packet forwarding and packet classification, but they are also useful in a variety of other applications that require high-speed table look-up. The main TCAM-design challenge is to decrease the power consumption associated with the large amount of parallel active circuitry, without compromising with speed or memory density. Furthermore, when the channel length decreases, leakage power becomes more significant, and it can even dominate dynamic power at lower technologies. In this paper, we propose a TCAM-design technique, called Virtual Power Supply technique that reduces the leakage by a substantial amount.

Keywords: match line (ML), search line (SL), ternary content addressable memory (TCAM), Leakage power (LP)

Procedia PDF Downloads 300

Authors: Swati Tyagi, Syed Abbas

Abstract:

Fractional-order Hopfield neural networks are generally used to model the information processing among the interacting neurons. To show the constancy of the processed information, it is required to analyze the stability of these systems. In this work, we perform Mittag-Leffler stability for the corresponding Caputo fractional-order bidirectional associative memory (BAM) neural networks with various time-delays. We derive sufficient conditions to ensure the existence and uniqueness of the equilibrium point by using the theory of topological degree theory. By applying the fractional Lyapunov method and Mittag-Leffler functions, we derive sufficient conditions for the global Mittag-Leffler stability, which further imply the global asymptotic stability of the network equilibrium. Finally, we present two suitable examples to show the effectiveness of the obtained results.

Keywords: bidirectional associative memory neural network, existence and uniqueness, fractional-order, Lyapunov function, Mittag-Leffler stability

Procedia PDF Downloads 366

Authors: Merve Akca

Abstract:

Retrieval-induced forgetting (RIF) refers to the phenomenon that selective retrieval of some information impairs memory for related, but not previously retrieved information. Despite age differences in retrieval-induced forgetting regarding retrospective memory being documented, this research aimed to highlight age differences in RIF of the prospective memory tasks for the first time. By using retrieval-practice paradigm, this study comparatively examined RIF effects in retrospective memory and event-based prospective memory in young and old adults. In this experimental study, a mixed factorial design with age group (Young, Old) as a between-subject variable, and memory type (Prospective, Retrospective) and item type (Practiced, Non-practiced) as within-subject variables was employed. Retrieval-induced forgetting was observed in the retrospective but not in the prospective memory task. Therefore, the results indicated that selective retrieval of past events led to suppression of other related past events in both age groups but not the suppression of memory for future intentions.

Keywords: prospective memory, retrieval-induced forgetting, retrieval inhibition, retrospective memory

Procedia PDF Downloads 317

Authors: Sorcha Bennett, Joe Sullivan

Abstract:

Triple-Level Cell (TLC) NAND Flash memory at, and below, 20nm (nanometer) is still largely unexplored by researchers, and with the ever more commonplace existence of Flash in consumer and enterprise applications there is a need for such gaps in knowledge to be filled. At the time of writing, there was little published data or literature on TLC, and more specifically reliability testing, with a further emphasis on both endurance and retention. This paper will give an introduction to NAND Flash memory, followed by an overview of the relevant current research on the reliability of Flash memory, along with the planned future work which will provide results to help characterise the reliability of TLC memory.

Keywords: endurance, patterns, raw flash, reliability, retention, TLC NAND flash memory, trade-off

Procedia PDF Downloads 361

Authors: Lizhi Ma, Dan Liu

Abstract:

Pipeline parallelism has been widely used to accelerate distributed deep learning to alleviate GPU memory bottlenecks and to ensure that models can be trained and deployed smoothly under limited graphics memory conditions. However, in highly heterogeneous distributed clusters, traditional model partitioning methods are not able to achieve load balancing. The overlap of communication and computation is also a big challenge. In this paper, HePipe is proposed, an efficient pipeline parallel training method for highly heterogeneous clusters. According to the characteristics of the neural network model pipeline training task, oriented to the 2-level heterogeneous cluster computing topology, a training method based on the 2-level stage division of neural network modeling and partitioning is designed to improve the parallelism. Additionally, a multi-forward 1F1B scheduling strategy is designed to accelerate the training time of each stage by executing the computation units in advance to maximize the overlap between the forward propagation communication and backward propagation computation. Finally, a dynamic recomputation strategy based on task memory requirement prediction is proposed to improve the fitness ratio of task and memory, which improves the throughput of the cluster and solves the memory shortfall problem caused by memory differences in heterogeneous clusters. The empirical results show that HePipe improves the training speed by 1.6×−2.2× over the existing asynchronous pipeline baselines.

Keywords: pipeline parallelism, heterogeneous cluster, model training, 2-level stage partitioning

Procedia PDF Downloads 19

Authors: Debneil Saha Roy

Abstract:

Weather forecasting is a complex topic and is well suited for analysis by deep learning approaches. With the wide availability of weather observation data nowadays, these approaches can be utilized to identify immediate comparisons between historical weather forecasts and current observations. This work explores the application of deep learning techniques to weather forecasting in order to accurately predict the weather over a given forecast hori­zon. Three deep neural networks are used in this study, namely, Multi-Layer Perceptron (MLP), Long Short Tunn Memory Network (LSTM) and a combination of Convolutional Neural Network (CNN) and LSTM. The predictive performance of these models is compared using two evaluation metrics. The results show that forecasting accuracy increases with an increase in the complexity of deep neural networks.

Keywords: convolutional neural network, deep learning, long short term memory, multi-layer perceptron

Procedia PDF Downloads 178

Authors: Dengpan Wu, Dan Liu

Abstract:

In view of the present cloud security problem has increasingly become one of the major obstacles hindering the development of the cloud computing, put forward a kind of memory based on Xen cloud environment security isolation technology implementation. And based on Xen virtual machine monitor system, analysis of the model of memory virtualization is implemented, using Xen memory virtualization system mechanism of super calls and grant table, based on the virtual machine manager internal implementation of access control module (ACM) to design the security isolation system memory. Experiments show that, the system can effectively isolate different customer domain OS between illegal access to memory data.

Keywords: cloud security, memory isolation, xen, virtual machine

Procedia PDF Downloads 409

Authors: Douglas Rossi Ramos

Abstract:

The current digital conjuncture, called by some authors as 'Internet of Things' (IoT), 'Web 2.0' or even 'Web 3.0', consists of a network that encompasses any communication of objects and entities, such as data, information, technologies, and people. At this juncture, especially characterized by an "object socialization," communication can no longer be represented as a simple informational flow of messages from a sender, crossing a channel or medium, reaching a receiver. The idea of communication must, therefore, be thought of more broadly in which it is possible to analyze the process communicative from interactions between humans and nonhumans. To think about this complexity, a communicative process that encompasses both humans and other beings or entities communicating (objects and things), it is necessary to constitute a new epistemology of communication to rethink concepts and notions commonly attributed to humans such as 'memory.' This research aims to contribute to this epistemological constitution from the discussion about the notion of memory according to the complex ontology of Henri Bergson. Among the results (the notion of memory in Bergson presents itself as a conceptual tool for the analysis of posthumanism and the anthropomorphic conjuncture of the new advent of digital), there was the need to think about an ontological memory, analyzed as a being itself (being itself of memory), as a strategy for understanding the forms of interaction and communication that constitute the new digital conjuncture, in which communicating beings or entities tend to interact with each other. Rethinking the idea of communication beyond the dimension of transmission in informative sequences paves the way for an ecological perspective of the digital dwelling condition.

Keywords: communication, digital, Henri Bergson, memory

Procedia PDF Downloads 165

Authors: Yiwei Li, Boyu Tian, Mingyu Gao

Abstract:

Hybrid main memory systems combine both performance and capacity advantages from heterogeneous memory technologies. With larger capacities, higher associativities, and finer granularities, hybrid memory systems currently exhibit significant metadata storage and lookup overheads for flexibly remapping data blocks between the two memory tiers. To alleviate the inefficiencies of existing designs, we propose Trimma, the combination of a multi-level metadata structure and an efficient metadata cache design. Trimma uses a multilevel metadata table to only track truly necessary address remap entries. The saved memory space is effectively utilized as extra DRAM cache capacity to improve performance. Trimma also uses separate formats to store the entries with non-identity and identity mappings. This improves the overall remap cache hit rate, further boosting the performance. Trimma is transparent to software and compatible with various types of hybrid memory systems. When evaluated on a representative DDR4 + NVM hybrid memory system, Trimma achieves up to 2.4× and on average 58.1% speedup benefits, compared with a state-of-the-art design that only leverages the unallocated fast memory space for caching. Trimma addresses metadata management overheads and targets future scalable large-scale hybrid memory architectures.

Keywords: memory system, data cache, hybrid memory, non-volatile memory

Procedia PDF Downloads 80

Authors: Farzaneh Badinloo, Niloufar Jalali-Moghadam, Reza Kormi-Nouri

Abstract:

The aim of this study was to explore the short-term and working memory performances across age and gender in school aged children. Most of the studies have been interested in looking into memory changes in adult subjects. This study was instead focused on exploring both short-term and working memories of children over time. Totally 410 school child participants belonging to four age groups (approximately 8, 10, 12 and 14 years old) among which were 201 girls and 208 boys were employed in the study. digits forward and backward tests of the Wechsler children intelligence scale-revised were conducted respectively as short-term and working memory measures. According to results, there was found a general increment in both short-term and working memory scores across age (p ˂ .05) by which whereas short-term memory performance was shown to increase up to 12 years old, working memory scores showed no significant increase after 10 years old of age. No difference was observed in terms of gender (p ˃ .05). In conclusion, this study suggested that both short-term and working memories improve across age in children where 12 and 10 years of old are likely the crucial age periods in terms of short-term and working memories development.

Keywords: age, gender, short-term memory, working memory

Procedia PDF Downloads 479

Authors: Lu Zhao, Nadir Farhi, Yeltsin Valero, Zoi Christoforou, Nadia Haddadou

Abstract:

In this paper, a long short-term memory (LSTM) neural network model is proposed to replicate simultaneously car-following and lane-changing behaviors in road networks. By combining two kinds of LSTM layers and three input designs of the neural network, six variants of the LSTM model have been created. These models were trained and tested on the NGSIM 101 dataset, and the results were evaluated in terms of longitudinal speed and lateral position, respectively. Then, we compared the LSTM model with a classical car-following model (the intelligent driving model (IDM)) in the part of speed decision. In addition, the LSTM model is compared with a model using classical neural networks. After the comparison, the LSTM model demonstrates higher accuracy than the physical model IDM in terms of car-following behavior and displays better performance with regard to both car-following and lane-changing behavior compared to the classical neural network model.

Keywords: traffic modeling, neural networks, LSTM, car-following, lane-change

Procedia PDF Downloads 263

Authors: E. U. Dampage, Ovindi D. Bandara, Vinushi S. Waraketiya, Samitha S. R. De Silva, Yasiru S. Gunarathne

Abstract:

The knowledge of water inflow figures is paramount in decision making on the allocation for consumption for numerous purposes; irrigation, hydropower, domestic and industrial usage, and flood control. The understanding of how reservoir inflows are affected by different climatic and hydrological conditions is crucial to enable effective water management and downstream flood control. In this research, we propose a method using a Long Short Term Memory (LSTM) Artificial Neural Network (ANN) to assist the aforesaid decision-making process. The Kotmale reservoir, which is the uppermost reservoir in the Mahaweli reservoir complex in Sri Lanka, was used as the test bed for this research. The ANN uses the runoff in the Kotmale reservoir catchment area and the effect of Sea Surface Temperatures (SST) to make a forecast for seven days ahead. Three types of ANN are tested; Multi-Layer Perceptron (MLP), Convolutional Neural Network (CNN), and LSTM. The extensive field trials and validation endeavors found that the LSTM ANN provides superior performance in the aspects of accuracy and latency.

Keywords: convolutional neural network, CNN, inflow, long short-term memory, LSTM, multi-layer perceptron, MLP, neural network

Procedia PDF Downloads 152

Authors: Laura Jenkins, Tim Eschle, Joanne Ciafone, Colin Hamilton

Abstract:

An original working memory model suggested the separation of visual and verbal systems in working memory architecture, in which only visual working memory components were used during visual working memory tasks. It was later suggested that the visuo spatial sketch pad was the only memory component at use during visual working memory tasks, and components such as the phonological loop were not considered. In more recent years, a contrasting approach has been developed with the use of an executive resource to incorporate both visual and verbal representations in visual working memory paradigms. This was supported using research demonstrating the use of verbal representations and an executive resource in a visual matrix patterns task. The aim of the current research is to investigate the working memory architecture during both a quantitative and a qualitative visual working memory task. A dual task method will be used. Three secondary tasks will be used which are designed to hit specific components within the working memory architecture – Dynamic Visual Noise (visual components), Visual Attention (spatial components) and Verbal Attention (verbal components). A comparison of the visual working memory tasks will be made to discover if verbal representations are at use, as the previous literature suggested. This direct comparison has not been made so far in the literature. Considerations will be made as to whether a domain specific approach should be employed when discussing visual working memory tasks, or whether a more domain general approach could be used instead.

Keywords: semantic organisation, visual memory, change detection

Procedia PDF Downloads 596

Authors: Yagya Raj Pandeya, Joonwhoan Lee

Abstract:

The dynamic behavior of music and video makes it difficult to evaluate musical instrument playing in a video by computer system. Any television or film video clip with music information are rich sources for analyzing musical instruments using modern machine learning technologies. In this research, we integrate the audio and video information sources using convolutional neural network (CNN) and pass network learned features through recurrent neural network (RNN) to preserve the dynamic behaviors of audio and video. We use different pre-trained CNN for music and video feature extraction and then fine tune each model. The music network use 2D convolutional network and video network use 3D convolution (C3D). Finally, we concatenate each music and video feature by preserving the time varying features. The long short term memory (LSTM) network is used for long-term dynamic feature characterization and then use late fusion with generalized mean. The proposed network performs better performance to recognize the musical instrument using audio-video multimodal neural network.

Keywords: multimodal, 3D convolution, music-video feature extraction, generalized mean

Procedia PDF Downloads 215

Authors: Yiwei Li, Mingyu Gao

Abstract:

Integrating hybrid memories with heterogeneous processors could leverage heterogeneity in both compute and memory domains for better system efficiency. To ensure performance isolation, we introduce Hydrogen, a hardware architecture to optimize the allocation of hybrid memory resources to heterogeneous CPU-GPU systems. Hydrogen supports efficient capacity and bandwidth partitioning between CPUs and GPUs in both memory tiers. We propose decoupled memory channel mapping and token-based data migration throttling to enable flexible partitioning. We also support epoch-based online search for optimized configurations and lightweight reconfiguration with reduced data movements. Hydrogen significantly outperforms existing designs by 1.21x on average and up to 1.31x.

Keywords: hybrid memory, heterogeneous systems, dram cache, graphics processing units

Procedia PDF Downloads 99

Authors: Xiang-Yao Zheng, Jen-Cheng Wang, Joe-Air Jiang

Abstract:

Improving the line ampacity while using existing power grids is an important issue that electricity dispatchers are now facing. Using the information provided by the dynamic thermal rating (DTR) of transmission lines, an overhead power grid can operate safely. However, dispatchers usually lack real-time DTR information. Thus, this study proposes a long-short-term memory (LSTM)-based method, which is one of the neural network models. The LSTM-based method predicts the DTR of lines using the weather data provided by Central Weather Bureau (CWB) of Taiwan. The possible thermal bottlenecks at different locations along the line and the margin of line ampacity can be real-time determined by the proposed LSTM-based prediction method. A case study that targets the 345 kV power grid of TaiPower in Taiwan is utilized to examine the performance of the proposed method. The simulation results show that the proposed method is useful to provide the information for the smart grid application in the future.

Keywords: electricity dispatch, line ampacity prediction, dynamic thermal rating, long-short-term memory neural network, smart grid

Procedia PDF Downloads 284

Authors: Anchana Julsiri, Seree Chadcham

Abstract:

The present study determined the effects of listening to pleasant Thai classical music on increasing working memory in elderly. Thai classical music without lyrics that made participants feel fun and aroused was used in the experiment for 3.19-5.40 minutes. The accuracy scores of Counting Span Task (CST), upper alpha ERD%, and theta ERS% were used to assess working memory of participants both before and after listening to pleasant Thai classical music. The results showed that the accuracy scores of CST and upper alpha ERD% in the frontal area of participants after listening to Thai classical music were significantly higher than before listening to Thai classical music (p < .05). Theta ERS% in the fronto-parietal network of participants after listening to Thai classical music was significantly lower than before listening to Thai classical music (p < .05).

Keywords: brain wave, elderly, pleasant Thai classical music, working memory

Procedia PDF Downloads 459

Authors: Muhammad Fajar Suardi, Fathimatufzzahra, Dela Isnaini Sendra

Abstract:

Learning is an activity that has been used to do, especially for a student or academics. But a handful of people have not been using and maximizing their brains work and some also do not know a good brain work time in capturing the lessons, so that knowledge is absorbed is also less than the maximum. Indonesia Memory Sports Council (IMSC) is an institution which is engaged in the performance of the brain and the development of effective learning methods by using several techniques that can be used in considering the lessons and knowledge to grasp well, including: loci method, substitution method, and chain method. This study aims to determine the techniques and benefits of using the method given in learning and memorization by applying memory techniques taught by Indonesia Memory Sports Council (IMSC) to students and the difference if not using this method. This research uses quantitative research with survey method addressed to students of Indonesian Memory Sports Council (IMSC). The results of this study indicate that learn, understand and remember the lesson using the techniques of memory which is taught in Indonesia Memory Sport Council is very effective and faster to absorb the lesson than learning without using the techniques of memory, and this affects the academic achievement of students in each educational institution.

Keywords: chain method, Indonesia memory sports council, loci method, substitution method

Procedia PDF Downloads 290

Authors: Rubaiyea Uddin

Abstract:

The hippocampus, located in the limbic system, is most commonly known for its role in memory and spatial navigation (as cited in Brain Reward and Pathways). It maintains an especially important role in specifically episodic and declarative memory. The hippocampus has also recently been linked to dopamine, the reward pathway’s primary neurotransmitter. Since research has found that dopamine also contributes to memory consolidation and hippocampal plasticity, this neurotransmitter is potentially responsible for contributing to the hippocampus’s role in memory formation. In this experiment we tested to see the effect of tactile cues on spatial navigation for eight different mice. We used a radial arm that had one designated 'reward' arm containing sucrose. The presence or absence of bedding was our tactile cue. We attempted to see if the memory of that cue would enhance the mice’s memory of having received the reward in that arm. The results from our study showed there was no significant response from the use of tactile cues on spatial navigation on our 129 mice. Tactile cues therefore do not influence spatial navigation.

Keywords: mice, radial arm maze, memory, spatial navigation, tactile cues, hippocampus, reward, sensory skills, Alzheimer’s, neurodegnerative disease

Procedia PDF Downloads 649