Abstract: This paper aims to provide an interpretation of artificial neural networks (ANNs) and explore some of its implications. The interpretation views ANNs as a memory which encodes instances of experience. An experiment explores the behavior of encoding and retrieval of instances from memory. A localised representation ANN is created that allows control over encoding and retrieved memory sample size and is experimented with using the MNIST digits dataset. The relationship between input familiarity, conflict within retrieved samples, and error rates is described and demonstrated to be an effective driver for memory encoding. Results indicate that selective encoding and retrieval samples that allow detection of memory conflicts produce optimal performance, and that error rates are normally distributed with input familiarity and conflict. By using input familiarity and sample consistency to guide memory encoding, the number of encoding trials on the dataset were reduced to 18.33% of the training data while maintaining good recognition performance on the test data.
Abstract: This article is based on observation of a cross-disciplinary, cross-institutional team that worked on an intervention called ‘Memory Mate’ for use in a UK Cancer Centre. This aimed to improve treatment outcomes for patients who had comorbid dementia or other memory impairment. Comorbid patients present ambiguous, spoiled identities, problematising the boundaries of health specialisms and frames of understanding. Memory Mate is theorised as a boundary object facilitating service transformation by changing relations between oncology and mental health care practice. It crosses the boundaries between oncology and mental health. Its introduction signifies an important step in reconfiguring relations between the specialisms. As a boundary object, it contains parallel, even contesting worlds, with potential to enable an eventual synthesis of the double stigma of cancer and dementia. Memory Mate comprises physical things, such as an animation, but its principal value is in the interaction it initiates across disciplines and services. It supports evolution of practices to address a newly emergent challenge for health service provision, namely the cancer patient with comorbid dementia/cognitive impairment. Getting clinicians from different disciplines working together on a practical solution generates a dialogue that can shift professional identity and change the culture of practice.
Abstract: Social media applications are increasingly being used in our everyday communications. These applications utilise end-to-end encryption mechanisms which make them suitable tools for criminals to exchange messages. These messages are preserved in the volatile memory until the device is restarted. Therefore, volatile forensics has become an important branch of digital forensics. In this study, the WormHex tool was developed to inspect the memory dump files for Windows and Mac based workstations. The tool supports digital investigators by enabling them to extract valuable data written in Arabic and English through web-based WhatsApp and Twitter applications. The results confirm that social media applications write their data into the memory, regardless of the operating system running the application, with there being no major differences between Windows and Mac.
Abstract: Time Series Forecasting (TSF) is used to predict the target variables at a future time point based on the learning from previous time points. To keep the problem tractable, learning methods use data from a fixed length window in the past as an explicit input. In this paper, we study how the performance of predictive models change as a function of different look-back window sizes and different amounts of time to predict into the future. We also consider the performance of the recent attention-based transformer models, which had good success in the image processing and natural language processing domains. In all, we compare four different deep learning methods (Recurrent Neural Network (RNN), Long Short-term Memory (LSTM), Gated Recurrent Units (GRU), and Transformer) along with a baseline method. The dataset (hourly) we used is the Beijing Air Quality Dataset from the website of University of California, Irvine (UCI), which includes a multivariate time series of many factors measured on an hourly basis for a period of 5 years (2010-14). For each model, we also report on the relationship between the performance and the look-back window sizes and the number of predicted time points into the future. Our experiments suggest that Transformer models have the best performance with the lowest Mean Absolute Errors (MAE = 14.599, 23.273) and Root Mean Square Errors (RSME = 23.573, 38.131) for most of our single-step and multi-steps predictions. The best size for the look-back window to predict 1 hour into the future appears to be one day, while 2 or 4 days perform the best to predict 3 hours into the future.
Abstract: Obfuscation is one of the most useful tools to prevent network compromise. Previous research focused on the obfuscation of the network communications between external-facing edge devices. This work proposes the use of two edge devices, external and internal facing, which communicates via private IPv4 addresses in a software-defined pseudo-random IP hopping. This methodology does not require additional IP addresses and/or resources to implement. Statistical analyses demonstrate that the hopping surface must be at least 1e3 IP addresses in size with a broad standard deviation to minimize the possibility of coincidence of monitored and communication IPs. The probability of breaking the hopping algorithm requires a collection of at least 1e6 samples, which for large hopping surfaces will take years to collect. The probability of dropped packets is controlled via memory buffers and the frequency of hops and can be reduced to levels acceptable for video streaming. This methodology provides an impenetrable layer of security ideal for information and supervisory control and data acquisition systems.
Abstract: Early detection of anomalies in data centers is important to reduce downtimes and the costs of periodic maintenance. However, there is little research on this topic and even fewer on the fusion of sensor data for the detection of abnormal events. The goal of this paper is to propose a method for anomaly detection in data centers by combining sensor data (temperature, humidity, power) and deep learning models. The model described in the paper uses one autoencoder per sensor to reconstruct the inputs. The auto-encoders contain Long-Short Term Memory (LSTM) layers and are trained using the normal samples of the relevant sensors selected by correlation analysis. The difference signal between the input and its reconstruction is then used to classify the samples using feature extraction and a random forest classifier. The data measured by the sensors of a data center between January 2019 and May 2020 are used to train the model, while the data between June 2020 and May 2021 are used to assess it. Performances of the model are assessed a posteriori through F1-score by comparing detected anomalies with the data center’s history. The proposed model outperforms the state-of-the-art reconstruction method, which uses only one autoencoder taking multivariate sequences and detects an anomaly with a threshold on the reconstruction error, with an F1-score of 83.60% compared to 24.16%.
Abstract: Prevent abuse and illegal activities in a given area of the sea is a very difficult and expensive task. Artificial intelligence offers the possibility to implement new methods to identify the vessel class type from the kinematic features of the vessel itself. The task strictly depends on the quality of the data. This paper explores the application of a deep Long Short-Term Memory model by using AIS flow only with a relatively low quality. The proposed model reaches high accuracy on detecting nine vessel classes representing the most common vessel types in the Ionian-Adriatic Sea. The model has been applied during the Adriatic-Ionian trial period of the international EU ANDROMEDA H2020 project to identify vessels performing behaviours far from the expected one, depending on the declared type.
Abstract: One of the most significant issues as attended a lot in recent years is that of recognizing the sentiments and emotions in social media texts. The analysis of sentiments and emotions is intended to recognize the conceptual information such as the opinions, feelings, attitudes and emotions of people towards the products, services, organizations, people, topics, events and features in the written text. These indicate the greatness of the problem space. In the real world, businesses and organizations are always looking for tools to gather ideas, emotions, and directions of people about their products, services, or events related to their own. This article uses the Twitter social network, one of the most popular social networks with about 420 million active users, to extract data. Using this social network, users can share their information and opinions about personal issues, policies, products, events, etc. It can be used with appropriate classification of emotional states due to the availability of its data. In this study, supervised learning and deep neural network algorithms are used to classify the emotional states of Twitter users. The use of deep learning methods to increase the learning capacity of the model is an advantage due to the large amount of available data. Tweets collected on various topics are classified into four classes using a combination of two Bidirectional Long Short Term Memory network and a Convolutional network. The results obtained from this study with an average accuracy of 93%, show good results extracted from the proposed framework and improved accuracy compared to previous work.
Abstract: Wheelchair users spend long hours in a sitting position, and selecting the right cushion is highly critical in preventing pressure ulcers in that demographic. Pressure Mapping Systems (PMS) are typically used in clinical settings by therapists to identify the sitting profile and pressure points in the sitting area to select the cushion that fits the best for the users. A PMS is a flexible mat composed of arrays of distributed networks of pressure sensors. The output of the PMS systems is a color-coded image that shows the intensity of the pressure concentration. Therapists use the PMS images to compare different cushions fit for each user. This process is highly subjective and requires good visual memory for the best outcome. This paper aims to develop an image processing technique to analyze the images of PMS and provide an objective measure to assess the cushions based on their pressure distribution mappings. In this paper, we first reviewed the skeletal anatomy of the human sitting area and its relation to the PMS image. This knowledge is then used to identify the important features that must be considered in image processing. We then developed an algorithm based on those features to analyze the images and rank them according to their fit to the user's needs.
Abstract: t-SNE is an embedding method that the data science community has widely used. It helps two main tasks: to display results by coloring items according to the item class or feature value; and for forensic, giving a first overview of the dataset distribution. Two interesting characteristics of t-SNE are the structure preservation property and the answer to the crowding problem, where all neighbors in high dimensional space cannot be represented correctly in low dimensional space. t-SNE preserves the local neighborhood, and similar items are nicely spaced by adjusting to the local density. These two characteristics produce a meaningful representation, where the cluster area is proportional to its size in number, and relationships between clusters are materialized by closeness on the embedding. This algorithm is non-parametric. The transformation from a high to low dimensional space is described but not learned. Two initializations of the algorithm would lead to two different embedding. In a forensic approach, analysts would like to compare two or more datasets using their embedding. A naive approach would be to embed all datasets together. However, this process is costly as the complexity of t-SNE is quadratic, and would be infeasible for too many datasets. Another approach would be to learn a parametric model over an embedding built with a subset of data. While this approach is highly scalable, points could be mapped at the same exact position, making them indistinguishable. This type of model would be unable to adapt to new outliers nor concept drift. This paper presents a methodology to reuse an embedding to create a new one, where cluster positions are preserved. The optimization process minimizes two costs, one relative to the embedding shape and the second relative to the support embedding’ match. The embedding with the support process can be repeated more than once, with the newly obtained embedding. The successive embedding can be used to study the impact of one variable over the dataset distribution or monitor changes over time. This method has the same complexity as t-SNE per embedding, and memory requirements are only doubled. For a dataset of n elements sorted and split into k subsets, the total embedding complexity would be reduced from O(n2) to O(n2/k), and the memory requirement from n2 to 2(n/k)2 which enables computation on recent laptops. The method showed promising results on a real-world dataset, allowing to observe the birth, evolution and death of clusters. The proposed approach facilitates identifying significant trends and changes, which empowers the monitoring high dimensional datasets’ dynamics.
Abstract: The ancestral relationship between humans and geographical environment has long been at the center of an interdisciplinary dialogue, which sees one of its main research nodes in the relationship between memory and places. Given its deep complexity, this symbiotic connection continues to look for a proper definition that appears increasingly negotiated by different disciplines. Numerous fields of knowledge are involved, from anthropology to semiotics of space, from photography to architecture, up to subjects traditionally far from these reasonings. This is the case of Design of Communication, a young discipline, now confident in itself and its objectives, aimed at finding and investigating original forms of visualization and representation, between sedimented knowledge and new technologies. In particular, Design of Communication for the Territory offers an alternative perspective to the debate, encouraging the reactivation and reconstruction of the memory of places. Recognizing mnemotopes as a cultural object of vertical interpretation of the memory-place relationship, design can become a real mediator of the territorial fixation of memories, making them increasingly accessible and perceptible, contributing to build a topography of memory. According to a mnemotopic vision, Communication Design can support the passage from a memory in which the observer participates only as an individual to a collective form of memory. A mnemotopic form of Communication Design can, through geolocation and content map-based systems, make chronology a topography rooted in the territory and practicable; it can be useful to understand how the perception of the memory of places changes over time, considering how to insert them in the contemporary world. Mnemotopes can be materialized in different format of translation, editing and narration and then involved in complex systems of communication. The memory of places, therefore, if stabilized by the tools offered by Communication Design, can make visible ruins and territorial stratifications, illuminating them with new communicative interests that can be shared and participated.
Abstract: As biometric systems become widely deployed, the security of identification systems can be easily attacked by various spoof materials. This paper contributes to finding a reliable and practical anti-spoofing method using Convolutional Neural Networks (CNNs) based on the types of loss functions and optimizers. The types of CNNs used in this paper include AlexNet, VGGNet, and ResNet. By using various loss functions including Cross-Entropy, Center Loss, Cosine Proximity, and Hinge Loss, and various loss optimizers which include Adam, SGD, RMSProp, Adadelta, Adagrad, and Nadam, we obtained significant performance changes. We realize that choosing the correct loss function for each model is crucial since different loss functions lead to different errors on the same evaluation. By using a subset of the Livdet 2017 database, we validate our approach to compare the generalization power. It is important to note that we use a subset of LiveDet and the database is the same across all training and testing for each model. This way, we can compare the performance, in terms of generalization, for the unseen data across all different models. The best CNN (AlexNet) with the appropriate loss function and optimizers result in more than 3% of performance gain over the other CNN models with the default loss function and optimizer. In addition to the highest generalization performance, this paper also contains the models with high accuracy associated with parameters and mean average error rates to find the model that consumes the least memory and computation time for training and testing. Although AlexNet has less complexity over other CNN models, it is proven to be very efficient. For practical anti-spoofing systems, the deployed version should use a small amount of memory and should run very fast with high anti-spoofing performance. For our deployed version on smartphones, additional processing steps, such as quantization and pruning algorithms, have been applied in our final model.
Abstract: In agricultural production and breeding, implementing automatic picking robot in orchard farming to reduce human labour and error is challenging. The core function of it is automatic identification based on machine vision. This paper focuses on apple detection and counting in orchards and implements several deep learning methods. Extensive datasets are used and a semi-automatic annotation method is proposed. The proposed deep learning models are in state-of-the-art YOLO family. In view of the essence of the models with various backbones, a multi-dimensional comparison in details is made in terms of counting accuracy, mAP and model memory, laying the foundation for realising automatic precision agriculture.
Abstract: Detecting subjectively biased statements is a vital task. This is because this kind of bias, when present in the text or other forms of information dissemination media such as news, social media, scientific texts, and encyclopedias, can weaken trust in the information and stir conflicts amongst consumers. Subjective bias detection is also critical for many Natural Language Processing (NLP) tasks like sentiment analysis, opinion identification, and bias neutralization. Having a system that can adequately detect subjectivity in text will boost research in the above-mentioned areas significantly. It can also come in handy for platforms like Wikipedia, where the use of neutral language is of importance. The goal of this work is to identify the subjectively biased language in text on a sentence level. With machine learning, we can solve complex AI problems, making it a good fit for the problem of subjective bias detection. A key step in this approach is to train a classifier based on BERT (Bidirectional Encoder Representations from Transformers) as upstream model. BERT by itself can be used as a classifier; however, in this study, we use BERT as data preprocessor as well as an embedding generator for a Bi-LSTM (Bidirectional Long Short-Term Memory) network incorporated with attention mechanism. This approach produces a deeper and better classifier. We evaluate the effectiveness of our model using the Wiki Neutrality Corpus (WNC), which was compiled from Wikipedia edits that removed various biased instances from sentences as a benchmark dataset, with which we also compare our model to existing approaches. Experimental analysis indicates an improved performance, as our model achieved state-of-the-art accuracy in detecting subjective bias. This study focuses on the English language, but the model can be fine-tuned to accommodate other languages.
Abstract: The Elliptic Curve Digital Signature algorithm-based X509v3 certificates are becoming more popular due to their short public and private key sizes. Moreover, these certificates can be stored in Internet of Things (IoT) devices, with limited resources, using less memory and transmitted in network security protocols, such as Internet Key Exchange (IKE), Transport Layer Security (TLS) and Secure Shell (SSH) with less bandwidth. The proposed method gives another advantage, in that it increases the performance of the above-mentioned protocols in terms of key exchange by saving one scalar multiplication operation.
Abstract: The cache has an important role in the reduction of access delay between a processor and memory in high-performance embedded systems. In these systems, the energy consumption is one of the most important concerns, and it will become more important with smaller processor feature sizes and higher frequencies. Meanwhile, the cache system dissipates a significant portion of energy compared to the other components of a processor. There are some elements that can affect the energy consumption of the cache such as replacement policy and degree of associativity. Due to these points, it can be inferred that selecting an appropriate configuration for the cache is a crucial part of designing a system. In this paper, we investigate the effect of different cache replacement policies on both cache’s performance and energy consumption. Furthermore, the impact of different Instruction Set Architectures (ISAs) on cache’s performance and energy consumption has been investigated.
Abstract: Nowadays, dialogue systems increasingly become the
way for humans to access many computer systems. So, humans
can interact with computers in natural language. A dialogue
system consists of three parts: understanding what humans say in
natural language, managing dialogue, and generating responses in
natural language. In this paper, we survey deep learning based
methods for dialogue management, response generation and dialogue
evaluation. Specifically, these methods are based on neural network,
long short-term memory network, deep reinforcement learning,
pre-training and generative adversarial network. We compare these
methods and point out the further research directions.
Abstract: Autonomous driving systems require high reliability to provide people with a safe and comfortable driving experience. However, despite the development of a number of vehicle sensors, it is difficult to always provide high perceived performance in driving environments that vary from time to season. The image segmentation method using deep learning, which has recently evolved rapidly, provides high recognition performance in various road environments stably. However, since the system controls a vehicle in real time, a highly complex deep learning network cannot be used due to time and memory constraints. Moreover, efficient networks are optimized for GPU environments, which degrade performance in embedded processor environments equipped simple hardware accelerators. In this paper, a semantic segmentation network, matrix multiplication accelerator network (MMANet), optimized for matrix multiplication accelerator (MMA) on Texas instrument digital signal processors (TI DSP) is proposed to improve the recognition performance of autonomous driving system. The proposed method is designed to maximize the number of layers that can be performed in a limited time to provide reliable driving environment information in real time. First, the number of channels in the activation map is fixed to fit the structure of MMA. By increasing the number of parallel branches, the lack of information caused by fixing the number of channels is resolved. Second, an efficient convolution is selected depending on the size of the activation. Since MMA is a fixed, it may be more efficient for normal convolution than depthwise separable convolution depending on memory access overhead. Thus, a convolution type is decided according to output stride to increase network depth. In addition, memory access time is minimized by processing operations only in L3 cache. Lastly, reliable contexts are extracted using the extended atrous spatial pyramid pooling (ASPP). The suggested method gets stable features from an extended path by increasing the kernel size and accessing consecutive data. In addition, it consists of two ASPPs to obtain high quality contexts using the restored shape without global average pooling paths since the layer uses MMA as a simple adder. To verify the proposed method, an experiment is conducted using perfsim, a timing simulator, and the Cityscapes validation sets. The proposed network can process an image with 640 x 480 resolution for 6.67 ms, so six cameras can be used to identify the surroundings of the vehicle as 20 frame per second (FPS). In addition, it achieves 73.1% mean intersection over union (mIoU) which is the highest recognition rate among embedded networks on the Cityscapes validation set.
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.
Abstract: Recent development of AI and edge computing plays a critical role to capture meaningful events such as detection of an unattended bag. One of the core problems is re-identification across multiple CCTVs. Immediately following the detection of a meaningful event is to track and trace the objects related to the event. In an extensive environment, the challenge becomes severe when the number of CCTVs increases substantially, imposing difficulties in achieving high accuracy while maintaining real-time performance. The algorithm that re-identifies cross-boundary objects for extensive tracking is referred to Extensive Re-Identification, which emphasizes the issues related to the complexity behind a great number of CCTVs. The Spatial-Temporal Awareness approach challenges the conventional thinking and concept of operations which is labor intensive and time consuming. The ability to perform Extensive Re-Identification through a multi-sensory network provides the next-level insights – creating value beyond traditional risk management.