Showing items 101 - 120 of 2083

Your selections:

Modelling and experimental assessment of high-frequency oscillation in DC microgrid

- Habibullah, Mohammad, Shah, Rakibuzzaman, Mithulananthan, Nadarajah, Islam, Syed, Islam, Rabiul

Neighbourhood houses and men’s sheds, adult learning in community spaces. Transforming the future of education : the role of research

- Smith, Erica, Ollis, Tracey

Overview of business hyper-automation

- Zhao, Xiaohui, Oseni, Taiwo, Medishetty, Bhanu

Physics-informed graph learning

- Peng, Ciyuan, Xia, Feng, Saikrishna, Vidya, Liu, Huan

Power quality improvement of a solar powered bidirectional smart grid and electric vehicle integration system

- Paidimukkala, Nikitha, Das, Narottam, Islam, Syed

  • Authors: Paidimukkala, Nikitha , Das, Narottam , Islam, Syed
  • Date: 2022
  • Type: Text , Conference paper
  • Relation: 4th IEEE Sustainable Power and Energy Conference, iSPEC 2022, Virtual, online, 4-7 December 2022, Proceedings - 2022 IEEE Sustainable Power and Energy Conference, iSPEC 2022
  • Full Text: false
  • Reviewed:
  • Description: This research paper mainly focuses on photovoltaic (PV) smart grid (SG) and electric vehicles (EVs) integration with two-way power flow capabilities by charging/discharging and power quality improvements using the power converters. Due to the increase of energy demand by rapid growth of population, it is necessary to modernize the power grid with improved power quality, such that the energy converted by solar PV (SPV) can be transmitted and stored as the excess amount of power in terms of batteries to use at peak load demand. The batteries of the EVs charge at a low level of demand and discharged at peak demand. The EVs can function both as a load and an energy supplier to the SG. The simulation results demonstrate the functioning of interfaced smart G2V system by observing and improving the factors such as power factor, power regulation and elimination of harmonics by constructing a power electronic network which can perform bidirectional power flow and balancing the network using MATLAB/Simulink software. Subsequentially, the improvement of power quality of the integrated system by analyzing power compensation, voltage regulation and harmonics mitigation of the integration system has been examined in detail. © 2022 IEEE.

Prediction of positive cloud-to-ground lightning striking zones for tilted thundercloud based on line charge model

- Barman, Surajit, Shah, Rakibuzzaman, Islam, Syed, Kumar, Apurv

  • Authors: Barman, Surajit , Shah, Rakibuzzaman , Islam, Syed , Kumar, Apurv
  • Date: 2022
  • Type: Text , Conference paper
  • Relation: 14th IEEE PES Asia-Pacific Power and Energy Engineering Conference, APPEEC 2022, Melbourne, 20-23 November 2023, 2022 IEEE PES 14th Asia-Pacific Power and Energy Engineering Conference (APPEEC) Vol. 2022-November
  • Full Text: false
  • Reviewed:
  • Description: Bushfire is known as one of the ascendant factors to create pyrocumulus thundercloud that causes the ignition of new fires by pyrocumulonimbus (pyroCb) lightning strikes mostly of positive polarity, and causes massive damage to nature and infrastructure. A conceptual model-based risk planning would be beneficial to predict the lightning striking zones on the surface of the earth underneath the pyroCb thundercloud. In this paper, a simple line charge structured thundercloud model is constructed in 2-D coordinates using the method of images to predict the probable +CG (positive cloud-To-ground) lightning striking zones on the earth's surface for tilted dipole thundercloud charge configuration. The electric potential distribution and ground surface charge density for tilted dipole thundercloud is investigated via continuously adjusting the position and charge density of its charge regions. Simulation results confirm the initiation of negative charged density for the wind shear extension of upper positive charge region by 2 to 8 km, and would expect +CG lightning to strike within 7.88 to 20 km around the earth periphery particularly in the direction of the cloud's forward flank. The proposed model would serve as the foundation to identify the probable lightning affected area as well as can also be extended to analyze the hazardous situation appears in wind energy farms or agricultural fencing situated nearby the power grid during pyroCb events. © 2022 IEEE.

Prediction of rockburst using supervised machine learning

- Kishore, Tharun, Khandelwal, Manoj

  • Authors: Kishore, Tharun , Khandelwal, Manoj
  • Date: 2022
  • Type: Text , Conference paper
  • Relation: International Conference on Geotechnical challenges in Mining, Tunneling and Underground structures, ICGMTU 2021, Virtual, online, 20-21 December 2021, Lecture Notes in Civil Engineering Vol. 228, p. 133-154
  • Full Text: false
  • Reviewed:
  • Description: The rockburst is one of the serious mining hazards, that cause injury, death, damage to mining equipment, and leads to financial problems in mining constructions. At the same time, mining is the major and essential resource of mineral commodities, that all mining locations uncover important for sustaining and enhancing their needs of dwelling. Moreover, mining contributes a significant portion of the GDP (Gross Domestic Production) growth of nations such as China, Australia, Russia, and the USA. These countries play major roles in exporting mineral commodities, which has been creating a huge demand to find out conventional methods for predicting the rockburst occurrence during mining. Moreover, determining rockburst occurrence using formulation or machine-based equipment has not been ideal in showcasing the result, due to changing geological parameters and the mining environment. In recent years, various soft computing and machine learning tools have been developed significantly, which help us to improve predicting models more accurate. In this research, original Rockburst data set have been collected in the last 20 years. Then standard empirical formulation method which is currently used in mining and supervised machine learning models has been used to predict rockburst occurrence. Finally, an interacting machine learning model developed in a way to predict the Rockburst occurrence result in three Indicators (high, medium and low Rockburst) are presented. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Prioritization of clinical alarms using semantic features of vital signs in remote patient monitoring

- Arora, Teena, Balasubramanian, Venki, Mai, Shenhan

  • Authors: Arora, Teena , Balasubramanian, Venki , Mai, Shenhan
  • Date: 2022
  • Type: Text , Conference paper
  • Relation: 2022 Australasian Computer Science Week, ACSW 2022, Virtual, Online, 14-17 February 2022, ACM International Conference Proceeding Series p. 242-245
  • Full Text: false
  • Reviewed:
  • Description: In recent years remote patient monitoring applications have emerged that can monitor the patient continuously and remotely with the help of wearable sensors that collect physiological data and send it to a telemedicine platform. Sensitivity of the sensor, patient's movement, electromagnetic interference, and data processing algorithms are a few factors that affect the collected data, leading to false alarms, and consequent false alarm leads to alarm fatigue. This study presents novel factors such as trust, frequency, slope, and trend that transform the vital signs raw data from the sensors into semantic data in a remote monitoring application. Experimental results have shown that data transformations lead to a reduction in clinically non-significant alarms and the prioritization of clinically significant alarms. © 2022 ACM.

Psychoinformatics : the behavioral analytics

- Nimje, Sparsh, Katade, Jayesh, Dunbray, Nachiket, Mavale, Shreyas, Kulkarni, Siddhivinayak, Firmin, Sally

  • Authors: Nimje, Sparsh , Katade, Jayesh , Dunbray, Nachiket , Mavale, Shreyas , Kulkarni, Siddhivinayak , Firmin, Sally
  • Date: 2022
  • Type: Text , Conference paper
  • Relation: 3rd International Conference on Communication, Computing and Electronics Systems, ICCCES 2021, Coimbatore, India, 28-29 October 2021, Proceedings of Third International Conference on Communication, Computing and Electronics Systems Vol. 844, p. 547-562
  • Full Text: false
  • Reviewed:
  • Description: Human behavior is very complex and cannot be explained using traditional mathematical models. Intermediate forms, such as those obtained from personality data, can be used to predict behavioral aspects of a person, creating the hypothesis that arbitrating psychological models can be drawn directly from recordings of behavior. In recent years, smartphone addiction has increased to a great extent. Since the excessive use of smartphones has negatively affected our daily life, many applications to reduce dependence on smartphones have been developed around the world. Personal attributes or personality types can be extracted from data obtained directly from smart phones without the interaction of participants who may have social or health interventions. Many people who excessively use their smartphones have an uncontrollable urge to use the Internet. Internet addiction refers to uncontrolled use of the Internet which causes hindrance in our daily life. Due to its negative impact on the education and lives of people, it is necessary to detect tendencies of people toward addictive behavior and provide them with preventative support and treatment. Similarly, the development of social media has seen rapid growth in its usage. People often find themselves overusing utilities such as virtual communication, texting, and sharing information which have also caused various behavioral problems. This study provides a summary of the various methods and studies done on these behavioral problems and to analyze different techniques, and machine learning models are used to predict addictive personality types. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

PyHENet : a generic framework for privacy-preserving DL inference based on fully homomorphic encryption

- Chen, Qian, Yao, Lin, Wu, Yulin, Wang, Xuan, Zhang, Weizhe, Jiang, Zoe, Liu, Yang, Alazab, Mamoun

  • Authors: Chen, Qian , Yao, Lin , Wu, Yulin , Wang, Xuan , Zhang, Weizhe , Jiang, Zoe , Liu, Yang , Alazab, Mamoun
  • Date: 2022
  • Type: Text , Conference paper
  • Relation: 4th International Conference on Data Intelligence and Security, ICDIS 2022, Shenzhen, China, 24-26 August 2022, Proceedings 2022 4th International Conference on Data Intelligence and Security ICDIS 2022 p. 127-133
  • Full Text: false
  • Reviewed:
  • Description: Deep learning inference provides inference service by service provider with model for client with input of personal data. Due to the huge commercial value inside, on one hand, both client's original data and inference output should be kept secret from others, even including service provider. On the other hand, service provider's model should be kept secret, especially from his competitor. Current research on privacy-preserving deep learning inference focuses on building models with specific data. This paper proposes a generic framework PyHENet of privacy-preserving deep learning inference based on Pytorch and lattice-based FHE, such that crypto library can be flexibly embedded into network. Firstly, raw data is encrypted by lattice-based FHE and uploaded to service provider. Secondly, convolutional computation over float-point ciphertext data is proposed for service provider to execute low accuracy loss inference with aided parallel method SIMD. Thirdly, inference result in ciphertext format is sent back to client for decryption. To improve efficiency, inference procedure can be further divided into two phases. All the computations during the second phase are in plaintext format with GPU acceleration, while keeping the first phase unchanged. Using the same model and parameters, the relative accuracy of PyHENet is almost 100% compared to the plaintext inference. This paper is the first to propose a general framework of neural networks for fully homomorphic cryptographic inference, and is based on mainstream deep learning frameworks, which is both secure and more conducive to development. © 2022 IEEE.

Radiation analysis of a particle curtain using polydisperse particel size eulerian granular CFD modelling

- Patel, Smitkumar, Chen, Jingling, Coventry, Joe, Lipinski, Wojciech, Kumar, Apurv

  • Authors: Patel, Smitkumar , Chen, Jingling , Coventry, Joe , Lipinski, Wojciech , Kumar, Apurv
  • Date: 2022
  • Type: Text , Conference paper
  • Relation: 7th Thermal and Fluids Engineering Conference, TFEC 2022, Las Vegas USA, 15-18 May 2022, Proceedings of the Thermal and Fluids Engineering Summer Conference Vol. 2022-May, p. 969-981
  • Full Text: false
  • Reviewed:
  • Description: A solid particle receiver is a principal element in concentrated solar power technology in which falling ceramic particles in a cavity are exposed to extremely concentrated solar irradiation. Such particle receiver holds great promise in achieving high thermal efficiencies due to the possibility of reaching temperatures as high as 1000° C. To accurately model the hydrodynamics and the radiation heat transfer, it is imperative to simulate particle-gas interactions more realistically. All the particle receiver modelling till date has used only monodisperse size assumption to simplify the simulation. However, the radiation interaction with the particle curtain is greatly dependent on the size-dependent radiation properties. In the present work, we aim to model the two-dimensional mass, momentum and radiative transfer equations using a Eulerian-Eulerian multiphase granular model and the discrete ordinates model with a particle size distribution in the falling particles. Gaussian distribution is assumed as a representative size distribution spread around a mean particle size of magnitudes generally used in particle receivers (~100-500 µm). The distribution is then split into n size bands and the Eulerian granular flow is modelled for n secondary phases (up to 3 in this study) with a corresponding concentration to simulate the particle size distribution. Each particle size band was prescribed unique size dependent radiation absorption and scattering coefficient for solving the discrete ordinates radiative transfer equation. Finally, a parametric study is carried out to understand the effect of different particle sizes and their concentration on the volume fraction distribution, particle velocities and radiation absorption by the curtain inside the receiver. © 2022 Begell House Inc.. All rights reserved.

Resilience of stablecoin reserve for distributed energy trading

- Islam, Md Ezazul, Chetty, Madhu, Lim, Suryani, Chadhar, Mehmood, Islam, Syed

  • Authors: Islam, Md Ezazul , Chetty, Madhu , Lim, Suryani , Chadhar, Mehmood , Islam, Syed
  • Date: 2022
  • Type: Text , Conference paper
  • Relation: 14th IEEE PES Asia-Pacific Power and Energy Engineering Conference, APPEEC 2022, Melbourne, Australia, 20-23 November 2022, Asia-Pacific Power and Energy Engineering Conference, APPEEC Vol. 2022-November
  • Full Text: false
  • Reviewed:
  • Description: For payment settlement, a blockchain-based Peer-To-Peer (P2P) energy trading requires a stable medium of exchange with little price volatility. Stablecoins, the most suitable medium of exchange, gaining concentration even from central banks. A consortium of central banks recommends complying with capital and liquidity standards for the high-quality liquid asset (HQLA) for the solvency of banks or financial institutions. Stablecoin as HQLA requires the adaption of such standards in P2P energy trading for reserve resilience. We propose a mechanism (NF90) that controls the inflow of stablecoins responding to Liquidity Coverage Ratio (LCR) for reserve resilience. Basel III accord recommends 100% of LCR. We measure the efficiency of NF90 concerning LCR as a metric. We simulate the proposed mechanism in Hyperledger Fabric as a permissioned blockchain platform for decentralisation, data storage, and smart contract. Experiments show that NF90 is the most efficient inflow control mechanism compared to other simulated mechanisms. © 2022 IEEE.

Revisiting social media in health care : a Bakhtinian carnival perspective

- Ukoha, Chukwuma, Stranieri, Andrew

  • Authors: Ukoha, Chukwuma , Stranieri, Andrew
  • Date: 2022
  • Type: Text , Conference paper
  • Relation: 2022 Australasian Computer Science Week, ACSW 2022, Virtual, Online, 14-17 February 2022, ACM International Conference Proceeding Series p. 254-256
  • Full Text:
  • Reviewed:
  • Description: Understanding the value of social media in health care has been a conundrum. Much of the literature in this area focuses on the use of social media for promotion, with very few studies seeking to elucidate how social media yields value in health care settings. This article draws on concepts from 18th century linguist Mikhail Bahktin to explain that social media acts like a Carnival in suspension of behavioral norms, and the provision of a forum for the proliferation of diverse dialogues. As a Carnival, social media plays an important role in encouraging dialogues that would not be appropriate within other spaces in the health care system. As such, social media is playing a pivotal role in changing norms toward shared care and patient empowerment. © 2022 ACM.

Risk and resiliency assessments of renewable dominated edge of grid under high-impact low-probability events -a Review

- Surinkaew, Tossaporn, Shah, Rakibuzzaman, Islam, Syed

Security of Internet of Things devices : ethical hacking a drone and its mitigation strategies

- Karmakar, Gour, Petty, Mark, Ahmed, Hassan, Das, Rajkumar, Kamruzzaman, Joarder

  • Authors: Karmakar, Gour , Petty, Mark , Ahmed, Hassan , Das, Rajkumar , Kamruzzaman, Joarder
  • Date: 2022
  • Type: Text , Conference paper
  • Relation: 2022 IEEE Asia-Pacific Conference on Computer Science and Data Engineering, CSDE 2022, Gold Coast, Australia, 18-20 December 2022, Proceedings of IEEE Asia-Pacific Conference on Computer Science and Data Engineering, CSDE 2022
  • Full Text: false
  • Reviewed:
  • Description: Internet of Things (IoT) is enabling us to introduce cost-effective, innovative and intelligent services in business, industrial, and government application domains. Despite these huge potential benefits of IoT applications, since the backbone of IoT is Internet and IoT connects numerous heterogeneous devices, IoT is vulnerable to many different attacks and thus has been a honey pot to the cybercriminals and hackers. For this reason, the attacks against IoT devices are increasing sharply in recent years. To prevent and detect these attacks, ethical hacking of different IoT devices are of paramount importance. This is because the lesson learnt from these ethical hackings can be exploited to develop effective and robust strategies and mitigation approaches to protect IoT devices from these attacks. There exist a few ethical hacking techniques reported in the literature such as hacking Android phones, Windows XP virtual machine and a DNS rebinding attack on IoT devices. In this paper, we implement an approach for the ethical hacking of a Drone and then hijack it. As an outcome of lesson learnt, the mitigation approaches on how to reduce the hacking on a drone is presented in this paper. © 2022 IEEE.

SEeMS : advanced artificial neural networks for employee learning motivation prediction

- Sin, Audrey, Islam, Sardar, Prentice, Catherine, Xia, Feng

  • Authors: Sin, Audrey , Islam, Sardar , Prentice, Catherine , Xia, Feng
  • Date: 2022
  • Type: Text , Conference paper
  • Relation: 7th IEEE International conference for Convergence in Technology, I2CT 2022, Pune, India, 7-9 April 2022, Proceedings 2022 IEEE 7th International conference for Convergence in Technology (I2CT)
  • Full Text: false
  • Reviewed:
  • Description: Employee learning motivation is vital for employee professional development and organisational success. However, worldwide statistics show that employees are generally unmotivated to learn. This study aims to examine employee learning motivation signals to determine the best-fit model for early intervention. In this paper, we present SEeMS a Smart Employee learning Motivation System to predict employee learning motivation autonomously. An Advanced Artificial Neural Networks (AANN) with a blended activation function of Sigmoid and ReLu (bSigReLu) is proposed and compared with other learning models. Experimental results demonstrate that the proposed model outperformed conventional state-of-art models. This novel study contributes to the field of organisational behaviour and data science by extending the usage of kernels and customised activation functions to solve the employee learning motivation problem. The superiority of the algorithm makes SEeMS ideal for practical deployment. According to the predictions, organisations could design better strategies to improve employee learning motivation for targeted employees. It is the first step towards achieving an eco-system of self-motivated employee learning that contributes to employee job satisfaction, performance, and well-being, indirectly contributing to employer competitiveness. © 2022 IEEE.

sGrid++ : revising simple grid based density estimator for mining outlying aspect

- Samariya, D., Ma, J., Aryal, S.

  • Authors: Samariya, D. , Ma, J. , Aryal, S.
  • Date: 2022
  • Type: Text , Conference paper
  • Relation: 23rd International Conference on Web Information Systems Engineering, WISE 2021, Biarritz, France, 1-3 November 2022, Web Information Systems Engineering – WISE 2022 23rd International Conference, Biarritz, France, November 1–3, 2022, Proceedings Vol. 13724 LNCS, p. 194-208
  • Full Text: false
  • Reviewed:
  • Description: In this paper, we address the problem of outlying aspect mining, which aims to identify a set of features (subspace(s) a.k.a aspect(s)) where a given data object stands out from the rest of the data. To detect the most outlying aspect of a given data object, outlying aspect mining algorithms need to compare and rank subspaces with different dimensionality. Thus, they require a fast and dimensionally unbias scoring measure. Existing measures use density or distance to compute the outlyingness of the query in each subspace. Density and distance are dimensionally bias, i.e. density decreases as the dimension of subspace increases. To make them comparable (dimensionally unbias), Z-score normalization is used in the previous works. However, to compute Z-score normalization, we need to compute the outlyingness of each data point in each subspace, which adds significant computational overhead on top of the already expensive density or distance computation. Recently developed measure called sGrid is a simple and efficient density estimator which allows a fast systemic search. While it is efficient compared to other distance and density-based measures, it is also a dimensionally bias measure and it requires to use Z-score normalization to make it dimensionality unbiased, which makes it computationally expensive. In this paper, we propose a simpler version of sGrid called sGrid++ that is not only efficient and effective but also dimensionality unbiased. It does not require Z-score normalization. We demonstrate the effectiveness and efficiency of the proposed scoring measure in outlying aspect mining using synthetic and real-world datasets. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Spam email categorization with nlp and using federated deep learning

- Ul Haq, Ikram, Black, Paul, Gondal, Iqbal, Kamruzzaman, Joarder, Watters, Paul, Kayes, A.

  • Authors: Ul Haq, Ikram , Black, Paul , Gondal, Iqbal , Kamruzzaman, Joarder , Watters, Paul , Kayes, A.
  • Date: 2022
  • Type: Text , Conference paper
  • Relation: 18th International Conference on Advanced Data Mining and Applications, ADMA 2022, Brisbane, Australia, 28-30 November 2022, Advanced Data Mining and Applications, 18th International Conference, ADMA 2022 Vol. 13726 LNAI, p. 15-27
  • Full Text: false
  • Reviewed:
  • Description: Emails are the most popular and efficient communication method that makes them vulnerable to misuse. Federated learning (FL) provides a decentralized machine learning (ML) model, where a central server coordinates clients that collaboratively train a shared ML model. This paper proposes Federated Phishing Filtering (FPF) technique based on federated learning, natural language processing, and deep learning. FL for intelligent algorithms fuses trained models of ML algorithms from multiple sites for collective learning. This approach improves ML performance by utilizing large collective training data sets across the corporate client base, resulting in higher phishing email detection accuracy. FPF techniques preserve email privacy using local feature extraction on client email servers. Thus, the contents of emails do not need to be transmitted across the network or stored on third-party servers. We have applied FL and Natural Language Processing (NLP) for email phishing detection. This technique provides four training modes that perform FL without sharing email content. Our research categorizes emails as benign, spam, and phishing. Empirical evaluations with publicly available datasets show that accuracy is improved by the use of our Federated Deep Learning model. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

The significance and impact of winning an academic award : a study of early career academics

- Ren, Jing, Shi, Yajie, Shatte, Adrian, Kong, Xiangjie, Xia, Feng

  • Authors: Ren, Jing , Shi, Yajie , Shatte, Adrian , Kong, Xiangjie , Xia, Feng
  • Date: 2022
  • Type: Text , Conference paper
  • Relation: 22nd ACM/IEEE Joint Conference on Digital Libraries, JCDL 2022, Virtual, online, 20-24 June 2022, Proceedings of the ACM/IEEE Joint Conference on Digital Libraries
  • Full Text: false
  • Reviewed:
  • Description: Academic award plays an important role in an academic's careerparticularly for early career academics. Previous studies have primarilyfocused on the impact of awards conferred to academics whoe made outstanding contributions to a specific research field, such as the Nobel Prize. In contrast, this paper aims to investigatethe effect of awards conferred to academics at an earlier careerstage, who have the potential to make a great impact in the future. We devise a metric named Award Change Factor (ACF), to evaluatethe change of a recipient's academic behavior after winningan academic award. Next, we propose a model to compare awardrecipients with academics who have similar performance beforewinning an academic award. In summary, we analyze the impact ofan award on the recipients' academic impact and their teams fromdifferent perspectives. Experimental results show that most recipientsdo have improvements in both productivity and citations afterwinning an academic award, while there is no significant impacton publication quality. In addition, receipt of an academic awardnot only expands recipients' collaboration network, but also has apositive effect on their team size. © 2022 Institute of Electrical and Electronics Engineers Inc.. All rights reserved.

Theoretical study and empirical investigation of sentence analogies

- Afantenos, Stergos, Lim, Suryani, Prade, Henri, Richard, Gilles

  • Authors: Afantenos, Stergos , Lim, Suryani , Prade, Henri , Richard, Gilles
  • Date: 2022
  • Type: Text , Conference paper
  • Relation: 1st Workshop on the Interactions between Analogical Reasoning and Machine Learning at 31st International Joint Conference on Artificial Intelligence - 25th European Conference on Artificial Intelligence, IARML@IJCAI-ECAI 2022, Vienna, Austria, 23 July 2022, CEUR Workshop Proceedings Vol. 3174, p. 15-28
  • Full Text:
  • Reviewed:
  • Description: Analogies between 4 sentences, “a is to b as c is to d”, are usually defined between two pairs of sentences (a, b) and (c, d) by constraining a relation R holding between the sentences of the first pair, to hold for the second pair. From a theoretical perspective, three postulates define an analogy - one of which is the “central permutation” postulate which allows the permutation of central elements b and c. This postulate is no longer appropriate in sentence analogies since the existence of R offers no guarantee in general for the existence of some relation S such that S also holds for the pairs (a, c) and (b, d). In this paper, the “central permutation” postulate is replaced by a weaker “internal reversal” postulate to provide an appropriate definition of sentence analogies. To empirically validate the aforementioned postulate, we build a LSTM as well as baseline Random Forest models capable of learning analogies based on quadruplets. We use the Penn Discourse Treebank (PDTB), the Stanford Natural Language Inference (SNLI) and the Microsoft Research Paraphrase (MSRP) corpora. Our experiments show that our models trained on samples of analogies between (a, b) and (c, d), recognize analogies between (b, a) and (d, c) when the underlying relation is symmetrical, validating thus the formal model of sentence analogies using “internal reversal” postulate. © 2022 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).