-
AI-Assisted Cervical Cancer Screening
Authors:
Kanchan Poudel,
Lisasha Poudel,
Prabin Raj Shakya,
Atit Poudel,
Archana Shrestha,
Bishesh Khanal
Abstract:
Visual Inspection with Acetic Acid (VIA) remains the most feasible cervical cancer screening test in resource-constrained settings of low- and middle-income countries (LMICs), which are often performed screening camps or primary/community health centers by nurses instead of the preferred but unavailable expert Gynecologist. To address the highly subjective nature of the test, various handheld devi…
▽ More
Visual Inspection with Acetic Acid (VIA) remains the most feasible cervical cancer screening test in resource-constrained settings of low- and middle-income countries (LMICs), which are often performed screening camps or primary/community health centers by nurses instead of the preferred but unavailable expert Gynecologist. To address the highly subjective nature of the test, various handheld devices integrating cameras or smartphones have been recently explored to capture cervical images during VIA and aid decision-making via telemedicine or AI models. Most studies proposing AI models retrospectively use a relatively small number of already collected images from specific devices, digital cameras, or smartphones; the challenges and protocol for quality image acquisition during VIA in resource-constrained camp settings, challenges in getting gold standard, data imbalance, etc. are often overlooked. We present a novel approach and describe the end-to-end design process to build a robust smartphone-based AI-assisted system that does not require buying a separate integrated device: the proposed protocol for quality image acquisition in resource-constrained settings, dataset collected from 1,430 women during VIA performed by nurses in screening camps, preprocessing pipeline, and training and evaluation of a deep-learning-based classification model aimed to identify (pre)cancerous lesions. Our work shows that the readily available smartphones and a suitable protocol can capture the cervix images with the required details for the VIA test well; the deep-learning-based classification model provides promising results to assist nurses in VIA screening; and provides a direction for large-scale data collection and validation in resource-constrained settings.
△ Less
Submitted 18 March, 2024;
originally announced March 2024.
-
Scaling While Privacy Preserving: A Comprehensive Synthetic Tabular Data Generation and Evaluation in Learning Analytics
Authors:
Qinyi Liu,
Mohammad Khalil,
Ronas Shakya,
Jelena Jovanovic
Abstract:
Privacy poses a significant obstacle to the progress of learning analytics (LA), presenting challenges like inadequate anonymization and data misuse that current solutions struggle to address. Synthetic data emerges as a potential remedy, offering robust privacy protection. However, prior LA research on synthetic data lacks thorough evaluation, essential for assessing the delicate balance between…
▽ More
Privacy poses a significant obstacle to the progress of learning analytics (LA), presenting challenges like inadequate anonymization and data misuse that current solutions struggle to address. Synthetic data emerges as a potential remedy, offering robust privacy protection. However, prior LA research on synthetic data lacks thorough evaluation, essential for assessing the delicate balance between privacy and data utility. Synthetic data must not only enhance privacy but also remain practical for data analytics. Moreover, diverse LA scenarios come with varying privacy and utility needs, making the selection of an appropriate synthetic data approach a pressing challenge. To address these gaps, we propose a comprehensive evaluation of synthetic data, which encompasses three dimensions of synthetic data quality, namely resemblance, utility, and privacy. We apply this evaluation to three distinct LA datasets, using three different synthetic data generation methods. Our results show that synthetic data can maintain similar utility (i.e., predictive performance) as real data, while preserving privacy. Furthermore, considering different privacy and data utility requirements in different LA scenarios, we make customized recommendations for synthetic data generation. This paper not only presents a comprehensive evaluation of synthetic data but also illustrates its potential in mitigating privacy concerns within the field of LA, thus contributing to a wider application of synthetic data in LA and promoting a better practice for open science.
△ Less
Submitted 12 January, 2024;
originally announced January 2024.
-
Active Learning for Neural Machine Translation
Authors:
Neeraj Vashistha,
Kriti Singh,
Ramakant Shakya
Abstract:
The machine translation mechanism translates texts automatically between different natural languages, and Neural Machine Translation (NMT) has gained attention for its rational context analysis and fluent translation accuracy. However, processing low-resource languages that lack relevant training attributes like supervised data is a current challenge for Natural Language Processing (NLP). We incor…
▽ More
The machine translation mechanism translates texts automatically between different natural languages, and Neural Machine Translation (NMT) has gained attention for its rational context analysis and fluent translation accuracy. However, processing low-resource languages that lack relevant training attributes like supervised data is a current challenge for Natural Language Processing (NLP). We incorporated a technique known Active Learning with the NMT toolkit Joey NMT to reach sufficient accuracy and robust predictions of low-resource language translation. With active learning, a semi-supervised machine learning strategy, the training algorithm determines which unlabeled data would be the most beneficial for obtaining labels using selected query techniques. We implemented two model-driven acquisition functions for selecting the samples to be validated. This work uses transformer-based NMT systems; baseline model (BM), fully trained model (FTM) , active learning least confidence based model (ALLCM), and active learning margin sampling based model (ALMSM) when translating English to Hindi. The Bilingual Evaluation Understudy (BLEU) metric has been used to evaluate system results. The BLEU scores of BM, FTM, ALLCM and ALMSM systems are 16.26, 22.56 , 24.54, and 24.20, respectively. The findings in this paper demonstrate that active learning techniques helps the model to converge early and improve the overall quality of the translation system.
△ Less
Submitted 30 December, 2022;
originally announced January 2023.
-
Modified SI Epidemic Model for Combating Virus Spread in Spatially Correlated Wireless Sensor Networks
Authors:
Rajeev K. Shakya
Abstract:
In wireless sensor networks (WSNs), main task of each sensor node is to sense the physical activity (i.e., targets or disaster conditions) and then to report it to the control center for further process. For this, sensor nodes are attached with many sensors having ability to measure the environmental information. Spatial correlation between nodes exists in such wireless sensor network based on com…
▽ More
In wireless sensor networks (WSNs), main task of each sensor node is to sense the physical activity (i.e., targets or disaster conditions) and then to report it to the control center for further process. For this, sensor nodes are attached with many sensors having ability to measure the environmental information. Spatial correlation between nodes exists in such wireless sensor network based on common sensory coverage and then the redundant data communication is observed. To study virus spreading dynamics in such scenario, a modified SI epidemic model is derived mathematically by incorporating WSN parameters such as spatial correlation, node density, sensing range, transmission range, total sensor nodes etc. The solution for proposed SI model is also determined to study the dynamics with time. Initially, a small number of nodes are attacked by viruses and then virus infection propagates through its neighboring nodes over normal data communication. Since redundant nodes exists in correlated sensor field, virus spread process could be different with different sensory coverage. The proposed SI model captures spatial and temporal dynamics than existing ones which are global. The infection process leads to network failure. By exploiting spatial correlation between nodes, spread control scheme is developed to limit the further infection in the network. Numerical result analysis is provided with comparison for validation.
△ Less
Submitted 15 January, 2018;
originally announced January 2018.
-
DSAT-MAC : Dynamic Slot Allocation based TDMA MAC protocol for Cognitive Radio Networks
Authors:
Rajeev K. Shakya,
Satyam Agarwal,
Y. N. Singh,
Nishchal K. Verma,
Amitabha Roy
Abstract:
Cognitive Radio Networks (CRN) have enabled us to efficiently reuse the underutilized radio spectrum. The MAC protocol in CRN defines the spectrum usage by sharing the channels efficiently among users. In this paper we propose a novel TDMA based MAC protocol with dynamically allocated slots. Most of the MAC protocols proposed in the literature employ Common Control Channel (CCC) to manage the reso…
▽ More
Cognitive Radio Networks (CRN) have enabled us to efficiently reuse the underutilized radio spectrum. The MAC protocol in CRN defines the spectrum usage by sharing the channels efficiently among users. In this paper we propose a novel TDMA based MAC protocol with dynamically allocated slots. Most of the MAC protocols proposed in the literature employ Common Control Channel (CCC) to manage the resources among Cognitive Radio (CR) users. Control channel saturation in case of large number of CR users is one of the main drawbacks of the CCC based MAC protocols. In contrast with CCC based MAC protocols, DSAT-MAC protocol is based on the TDMA mechanism, without using any CCC for control information exchange. The channels are divided into time slots and CR users send their control or data packets over their designated slot. The protocol ensures that no slot is left vacant. This guarantees full use of the available spectrum. The protocol includes the provision for Quality of Service, where real-time and safety critical data is transmitted with highest priority and least delay. The protocol also ensures a fair sharing of available spectrum among the CR users, with the mechanism to regulate the transmission of malicious nodes. Energy saving techniques are also presented for longer life of battery operated CR nodes. Theoretical analysis and simulations over ns-2 of the proposed protocol reveal that the protocol performs better in various CR adhoc network applications.
△ Less
Submitted 17 January, 2013;
originally announced January 2013.
-
TTMA: Traffic-adaptive Time-division Multiple Access Protocol Wireless Sensor Networks
Authors:
Rajeev K. Shakya
Abstract:
This paper has been withdrawn by arXiv. arXiv admin note: author list truncated due to disputed authorship and content. This submission repeats large portions of text from this http URL by other authors. Duty cycle mode in WSN improves energy-efficiency, but also introduces packet delivery latency. Several duty-cycle based MAC schemes have been proposed to reduce latency, but throughput is limite…
▽ More
This paper has been withdrawn by arXiv. arXiv admin note: author list truncated due to disputed authorship and content. This submission repeats large portions of text from this http URL by other authors. Duty cycle mode in WSN improves energy-efficiency, but also introduces packet delivery latency. Several duty-cycle based MAC schemes have been proposed to reduce latency, but throughput is limited by duty-cycled scheduling performance. In this paper, a Traffic-adaptive Time-division Multiple Access (TTMA), a distributed TDMA-based MAC protocol is introduced to improves the throughput by traffic-adaptive time-slot scheduling that increases the channel utilisation efficiency. The proposed time-slot scheduling method first avoids time-slots assigned to nodes with no traffic through fast traffic notification. It then achieves better channel utilisation among nodes having traffic through an ordered schedule negotiation scheme. By decomposing traffic notification and data transmission scheduling into two phases leads each phase to be simple and efficient. The performance evaluation shows that the two-phase design significantly improves the throughput and outperforms the time division multiple access (TDMA) control with slot stealing.
△ Less
Submitted 9 August, 2012; v1 submitted 8 June, 2012;
originally announced June 2012.
-
Optimizing Channel Access for Event-Driven Wireless Sensor Networks: Analysis and Enhancements
Authors:
Rajeev K. Shakya,
Yatindra Nath Singh,
Nishchal K. Verma
Abstract:
We study the problem of medium access control in domain of event-driven wireless sensor networks (WSNs). In this kind of WSN, sensor nodes send data to sink node only when an event occurs in the monitoring area. The nodes in this kind of WSNs encounter correlated traffic as a subset of nodes start sending data by sensing a common event simultaneously. We wish to rethink of medium access control (M…
▽ More
We study the problem of medium access control in domain of event-driven wireless sensor networks (WSNs). In this kind of WSN, sensor nodes send data to sink node only when an event occurs in the monitoring area. The nodes in this kind of WSNs encounter correlated traffic as a subset of nodes start sending data by sensing a common event simultaneously. We wish to rethink of medium access control (MAC) for this type of traffic characteristics. For WSNs, many existing MAC protocols utilize the basic CSMA/CA strategy such as IEEE 802.11 Binary Exponential Backoff (BEB) algorithm to handle the collisions among packets when more than one node need to access the channel. We show that this BEB algorithm does not work well without incurring access delay or performance degradation due to increased number of collisions and retransmissions when nodes encounter correlated traffic. Based on above observations in mind, We present a Adaptive Random Backoff (ARB) algorithm that is capable of mitigating the impact of correlated traffic and capable of minimizing the chance of collisions. ARB is based on minor modifications of BEB. We show using numerical analysis that our proposals improve the channel access in terms of latency, throughput, and frame drop** probability as compared with IEEE 802.11 DCF. Simulations using NS-2 network simulator are conducted to validate the analytical results.
△ Less
Submitted 20 May, 2012; v1 submitted 27 March, 2012;
originally announced March 2012.
