Multi-dimensional Long Short-term Memory Research Articles

Deep learning-based recognition system for pashto handwritten text: benchmark on PHTI.

This article introduces a recognition system for handwritten text in the Pashto language, representing the first attempt to establish a baseline system using the Pashto Handwritten Text Imagebase (PHTI) dataset. Initially, the PHTI dataset underwent pre-processed to eliminate unwanted characters, subsequently, the dataset was divided into training 70%, validation 15%, and test sets 15%. The proposed recognition system is based on multi-dimensional long short-term memory (MD-LSTM) networks. A comprehensive empirical analysis was conducted to determine the optimal parameters for the proposed MD-LSTM architecture; Counter experiments were used to evaluate the performance of the proposed system comparing with the state-of-the-art models on the PHTI dataset. The novelty of our proposed model, compared to other state of the art models, lies in its hidden layer size (i.e., 10, 20, 80) and its Tanh layer size (i.e., 20, 40). The system achieves a Character Error Rate (CER) of 20.77% as a baseline on the test set. The top 20 confusions are reported to check the performance and limitations of the proposed model. The results highlight complications and future perspective of the Pashto language towards the digital transition.

Enhancing rainwater harvesting and groundwater recharge efficiency with multi-dimensional LSTM and clonal selection algorithm

Rainwater harvesting stands out as a promising solution to alleviate water scarcity and alleviate pressure on conventional water reservoirs. This work introduces a pioneering strategy to elevate the efficiency of rainwater harvesting systems through the fusion of Multi-Dimensional Long Short-Term Memory (LSTM) networks and the Clonal Selection Algorithm (CSA). The Multi-Dimensional LSTM networks serve to model intricate temporal and spatial rainfall patterns, enabling precise predictions regarding the optimal times and locations for rainwater abundance. This insight is pivotal in refining the design and operation of rainwater harvesting setups. Drawing inspiration from the immune system, the Clonal Selection Algorithm is employed to optimize site selection and resource allocation, ensuring the maximal utilization of harvested rainwater. The adaptability and robustness of CSA prove invaluable in tackling the dynamic nature of rainfall patterns. This research endeavor is dedicated to enhancing groundwater levels and optimizing its sources through the implementation of efficient harvesting techniques. By delving into innovative methodologies, it aims to contribute significantly to sustainable water management practices and ensure a reliable supply of groundwater for various societal needs. The experiments are conducted to study the effectiveness of rainwater harvesting systems, where the proposed method achieves increased efficiency, thereby reducing dependence on conventional water sources and contributing to sustainable water management practices. The proposed CSA-LSTM model demonstrates superior performance compared to ACO-ANN and PSO-BPNN, achieving higher training, testing, and validation accuracies while exhibiting lower training, testing, and validation losses. Additionally, CSA-LSTM showcases excellent site suitability, high resource utilization, and robustness to changes, with a fast response time, emphasizing its potential for efficient and effective applications.

A Hybrid Feature Selection-multidimensional LSTM Framework for Deformation Prediction of Super High Arch Dams

Developing a deformation prediction model with strong robustness and applicability based on monitoring data is an important task in establishing a safety monitoring system for super high arch dams. This study proposes a hybrid deep learning (DL) model for deformation prediction, named HFS/xs-MLSTM. The proposed multidimensional long short-term memory (MLSTM) DL achieves integrated modeling of two factors, environment and temporal correlation, by constructing a multivariate matrix, and achieves sequential prediction by rolling through the sliding windows. This operation solves the problem of insufficient prediction accuracy due to the traditional method considering only a single factor. To address the problems of computational time consumption and DL overfitting caused by variable redundancy, a hybrid feature selection method (HFS/xs) based on eXtreme gradient boosting and Spearman is proposed. Analysis shows that HFS/xs can comprehensively reflect the intrinsic laws of the data and effectively filter out the most concise set of variables, which greatly improves the operability and stability of MLSTM. Compared with traditional methods, HFS/xs-MLSTM has the best performance and applicability for deformation prediction in different zones and operating conditions. This study can provide reliable a priori knowledge for the construction of safety monitoring system for super high arch dams.

When Massive GPU Parallelism Ain’t Enough: A Novel Hardware Architecture of 2D-LSTM Neural Network

Multidimensional Long Short-Term Memory (MD-LSTM) neural network is an extension of one-dimensional LSTM for data with more than one dimension. MD-LSTM achieves state-of-the-art results in various applications, including handwritten text recognition, medical imaging, and many more. However, its implementation suffers from the inherently sequential execution that tremendously slows down both training and inference compared to other neural networks. The main goal of the current research is to provide acceleration for inference of MD-LSTM. We advocate that Field-Programmable Gate Array (FPGA) is an alternative platform for deep learning that can offer a solution when the massive parallelism of GPUs does not provide the necessary performance required by the application. In this article, we present the first hardware architecture for MD-LSTM. We conduct a systematic exploration to analyze a tradeoff between precision and accuracy. We use a challenging dataset for semantic segmentation, namely historical document image binarization from the DIBCO 2017 contest and a well-known MNIST dataset for handwritten digit recognition. Based on our new architecture, we implement FPGA-based accelerators that outperform Nvidia Geforce RTX 2080 Ti with respect to throughput by up to 9.9 and Nvidia Jetson AGX Xavier with respect to energy efficiency by up to 48 . Our accelerators achieve higher throughput, energy efficiency, and resource efficiency than FPGA-based implementations of convolutional neural networks (CNNs) for semantic segmentation tasks. For the handwritten digit recognition task, our FPGA implementations provide higher accuracy and can be considered as a solution when accuracy is a priority. Furthermore, they outperform earlier FPGA implementations of one-dimensional LSTMs with respect to throughput, energy efficiency, and resource efficiency.

The algorithm of current prediction based on multi-dimensional Long Short Term Memory networks

Precisely predict the current on the node circuit in advance, which can facilitate the manager to manage and control the risk in advance. The current, voltage, residual current, cable temperature, ambient temperature, and ambient humidity on the circuit constitute a multi-dimensional sequence characteristic sequence data. This paper proposes an improved Long Short-Term Memory (LSTM) cyclic neural network method, which combines multi-dimensional Train the neural network in a specific way to obtain the LSTM current prediction model. The experiment is validated with the perception data in an intelligent building, and compared with the classic neural network and Recurrent Neural Network (RNN) prediction method. The experimental results show that, compared with the prediction methods such as BP neural network and RNN network, the multi-dimensional LSTM has a higher accuracy in predicting the current in the circuit, and the MSE index is as low as 0.5. And this method can detect potential electrical fire hazards and give early warning 1–2 time intervals in advance to avoid electrical fires.

Predicting and Interpreting Spatial Accidents through MDLSTM.

Predicting and interpreting the spatial location and causes of traffic accidents is one of the current hot topics in traffic safety. This research purposed a multi-dimensional long-short term memory neural network model (MDLSTM) to fit the non-linear relationships between traffic accident characteristics and land use properties, which are further interpreted to form local and general rules. More variables are taken into account as the input land use properties and the output traffic accident characteristics. Five types of traffic accident characteristics are simultaneously predicted with higher accuracy, and three levels of interpretation, including the hidden factor-traffic potential, the potential-determine factors, which varies between grid cells, and the general rules across the whole study area are analyzed. Based on the model, some interesting insights were revealed including the division line in the potential traffic accidents in Shenyang (China). It is also purposed that the relationship between land use and accidents differ from previous researches in the neighboring and regional aspects. Neighboring grids have strong spatial connections so that the relationship of accidents in a continuous area is relatively similar. In a larger region, the spatial location is found to have a great influence on the traffic accident and has a strong directionality.

Trip Generation Prediction Based on the Convolutional Neural Network-Multidimensional Long-Short Term Memory Neural Network Model at Grid Cell Scale

The prediction of trip generation is an essential problem for effective traffic engineering and urban management. Traditional methods are on the large spatial scale (e.g. Traffic analysis Zone, TAZ), based on the single source and fewer types data. It is difficult to carry out refined research on smaller spatial units, due to the high aggregation of personal trip survey data. In addition, the experience-based models cannot easily capture complex non-linear relationship, which leads to lower accuracy. Multi-sources data provides the possibility to improve the prediction accuracy of trip generation. Based on the point of interest data (POI), more disaggregate spatial unit are subdivided, and grid-scale spatial correlations are taken into consideration. This paper proposes a Convolutional Neural Network-Multidimensional Long-short term memory neural network (CNN-MDLSTM) model to analyze the spatial correlation between trip generation and land use features, capture prominent features in a spatial range through the convolution structure and describe the spatial interaction using the sequence transfer structure of Long-short term memory neural network (LSTM). The deformed Multi-Dimensional Long-Short Term Memory neural network (MDLSTM) is used to adapt to the two-dimensional spatial relationship. Through case analysis and comparative analysis between models, it is shown that CNN-MDLSTM characterizes the quantitative law of trip generation and land use features better than the other neural network models. In addition, this study also discusses the output prediction accuracy of different land use grid cells and the impact of different land use characteristics on the prediction accuracy.

A Deep Learning based Arabic Script Recognition System: Benchmark on KHAT

This paper presents a deep learning benchmark on a complex dataset known as KFUPM Handwritten Arabic TexT (KHATT). The KHATT data-set consists of complex patterns of handwritten Arabic text-lines. This paper contributes mainly in three aspects i.e., (1) pre-processing, (2) deep learning based approach, and (3) data-augmentation. The pre-processing step includes pruning of white extra spaces plus de-skewing the skewed text-lines. We deploy a deep learning approach based on Multi-Dimensional Long Short-Term Memory (MDLSTM) networks and Connectionist Temporal Classification (CTC). The MDLSTM has the advantage of scanning the Arabic text-lines in all directions (horizontal and vertical) to cover dots, diacritics, strokes and fine inflammation. The data-augmentation with a deep learning approach proves to achieve better and promising improvement in results by gaining 80.02% Character Recognition (CR) over 75.08% as baseline.

Urdu Nasta'liq text recognition using implicit segmentation based on multi-dimensional long short term memory neural networks.

The recognition of Arabic script and its derivatives such as Urdu, Persian, Pashto etc. is a difficult task due to complexity of this script. Particularly, Urdu text recognition is more difficult due to its Nasta’liq writing style. Nasta’liq writing style inherits complex calligraphic nature, which presents major issues to recognition of Urdu text owing to diagonality in writing, high cursiveness, context sensitivity and overlapping of characters. Therefore, the work done for recognition of Arabic script cannot be directly applied to Urdu recognition. We present Multi-dimensional Long Short Term Memory (MDLSTM) Recurrent Neural Networks with an output layer designed for sequence labeling for recognition of printed Urdu text-lines written in the Nasta’liq writing style. Experiments show that MDLSTM attained a recognition accuracy of 98% for the unconstrained Urdu Nasta’liq printed text, which significantly outperforms the state-of-the-art techniques.