Dimitrios Melissourgos, Hanzhi Gao, Chaoyi Ma, Shigang Chen, Sam S Wu
Artificial neural networks (ANNs) are changing the paradigm in medical diagnosis. However, it remains an open problem how to outsource the model training operations to the cloud while protecting the privacy of distributed patient data. Homomorphic encryption suffers from high overhead over data independently encrypted from numerous sources, differential privacy introduces a high level of noise which drastically increases the number of patient records needed to train a model, while federated learning requires all participants to perform synchronized local training that counters our goal of outsourcing all training operations to the cloud. This paper proposes to use matrix masking for outsourcing all model training operations to the cloud with privacy protection. After outsourcing their masked data to the cloud, the clients do not need to coordinate and perform any local training operations. The accuracy of the models trained by the cloud from the masked data is comparable to the accuracy of the optimal benchmark models that are trained directly from the original raw data. Our results are confirmed by experimental studies on privacy-preserving cloud training of medical-diagnosis neural network models based on real-world Alzheimer's disease data and Parkinson's disease data.
{"title":"Training Medical-Diagnosis Neural Networks on the Cloud with Privacy-Sensitive Patient Data from Multiple Clients.","authors":"Dimitrios Melissourgos, Hanzhi Gao, Chaoyi Ma, Shigang Chen, Sam S Wu","doi":"10.1145/3549206.3549291","DOIUrl":"https://doi.org/10.1145/3549206.3549291","url":null,"abstract":"<p><p>Artificial neural networks (ANNs) are changing the paradigm in medical diagnosis. However, it remains an open problem how to outsource the model training operations to the cloud while protecting the privacy of distributed patient data. Homomorphic encryption suffers from high overhead over data independently encrypted from numerous sources, differential privacy introduces a high level of noise which drastically increases the number of patient records needed to train a model, while federated learning requires all participants to perform synchronized local training that counters our goal of outsourcing all training operations to the cloud. This paper proposes to use matrix masking for outsourcing all model training operations to the cloud with privacy protection. After outsourcing their masked data to the cloud, the clients do not need to coordinate and perform any local training operations. The accuracy of the models trained by the cloud from the masked data is comparable to the accuracy of the optimal benchmark models that are trained directly from the original raw data. Our results are confirmed by experimental studies on privacy-preserving cloud training of medical-diagnosis neural network models based on real-world Alzheimer's disease data and Parkinson's disease data.</p>","PeriodicalId":72026,"journal":{"name":"... International Conference on Contemporary Computing. IC3 (Conference)","volume":"2022 ","pages":"502-508"},"PeriodicalIF":0.0,"publicationDate":"2022-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10155738/pdf/nihms-1892709.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"9479376","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Obviating Multiple Attacks with Enhanced Logic Locking","authors":"P. Anu, N. Mohankumar, M. N. Devi","doi":"10.1145/3549206.3549235","DOIUrl":"https://doi.org/10.1145/3549206.3549235","url":null,"abstract":"","PeriodicalId":72026,"journal":{"name":"... International Conference on Contemporary Computing. IC3 (Conference)","volume":"56 1","pages":"162-167"},"PeriodicalIF":0.0,"publicationDate":"2022-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75881561","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-08-01DOI: 10.1109/ic3.2019.8844903
I. Dhillon
Despite having remarkable performance on many sequence learning tasks, recurrent neural networks (RNNs) are hard to train with long sequences due to limited expressive power and the vanishing and exploding gradient issues. Previous work has focused on stabilizing the gradients by encouraging orthogonality of weight matrices via re-parameterization techniques. However, two major issues remain in these methods. First, the re-parameterization often relies on a chain of operations on small matrices or vectors that are not friendly to hardware accelerators. As a result, it becomes a source of performance bottleneck for training. Second, these methods fix the singular values of the transition matrix throughout the temporal dimension, which further restricts the expressive power of the model and wastes the potential of encoding useful information into the singular values. In this talk, I will present the Singular Value Gated RNN that can efficiently encode temporal information into singular values, as well as mitigate the vanishing and exploding gradient problems. In addition, we can design novel forward and backward propagation algorithms that are friendly to hardware accelerators. This leads to 3-4 times speedup on GPUs and greatly reduces memory cost. On contemporary applications like voice recognition and text summarization, where long term dependencies are hard to capture, the proposed method outperforms other recurrent models with similar or smaller model sizes. Joint work with Jiong Zhang of UT Austin
{"title":"Abstract of the Keynotes","authors":"I. Dhillon","doi":"10.1109/ic3.2019.8844903","DOIUrl":"https://doi.org/10.1109/ic3.2019.8844903","url":null,"abstract":"Despite having remarkable performance on many sequence learning tasks, recurrent neural networks (RNNs) are hard to train with long sequences due to limited expressive power and the vanishing and exploding gradient issues. Previous work has focused on stabilizing the gradients by encouraging orthogonality of weight matrices via re-parameterization techniques. However, two major issues remain in these methods. First, the re-parameterization often relies on a chain of operations on small matrices or vectors that are not friendly to hardware accelerators. As a result, it becomes a source of performance bottleneck for training. Second, these methods fix the singular values of the transition matrix throughout the temporal dimension, which further restricts the expressive power of the model and wastes the potential of encoding useful information into the singular values. In this talk, I will present the Singular Value Gated RNN that can efficiently encode temporal information into singular values, as well as mitigate the vanishing and exploding gradient problems. In addition, we can design novel forward and backward propagation algorithms that are friendly to hardware accelerators. This leads to 3-4 times speedup on GPUs and greatly reduces memory cost. On contemporary applications like voice recognition and text summarization, where long term dependencies are hard to capture, the proposed method outperforms other recurrent models with similar or smaller model sizes. Joint work with Jiong Zhang of UT Austin","PeriodicalId":72026,"journal":{"name":"... International Conference on Contemporary Computing. IC3 (Conference)","volume":"1 1","pages":"1-2"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81789298","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-08-01DOI: 10.1109/IC3.2018.8530461
Jasminder Kaur Sandhu, A. Verma, P. Rana
The data-driven technique uses real-world readings or simulated dataset to draw inference about the behavior of communication network. The design of the network is further optimized to enhance the performability according to the inference drawn. The performability of the network is dependent on the performance parameters of the network such as packet delivery ratio, packets dropped, delay, throughput, and data rate. The data rate prediction is carried out using different machine learning techniques. Further, the performability of the network is directly associated with its survivability. Better is the network performability, more is the survivability of that particular network. This work proposes a framework for survivable Wireless Sensor Network which predicts the data rate of the network. The past experience serves as an optimized way to traverse data in the network with efficient data rate. A primary dataset designed with the help of simulations is used for this work. Also, the robustness of best predictive model is checked with the help of N-fold cross-validation technique.
{"title":"A Data-Driven Framework for Survivable Wireless Sensor Networks","authors":"Jasminder Kaur Sandhu, A. Verma, P. Rana","doi":"10.1109/IC3.2018.8530461","DOIUrl":"https://doi.org/10.1109/IC3.2018.8530461","url":null,"abstract":"The data-driven technique uses real-world readings or simulated dataset to draw inference about the behavior of communication network. The design of the network is further optimized to enhance the performability according to the inference drawn. The performability of the network is dependent on the performance parameters of the network such as packet delivery ratio, packets dropped, delay, throughput, and data rate. The data rate prediction is carried out using different machine learning techniques. Further, the performability of the network is directly associated with its survivability. Better is the network performability, more is the survivability of that particular network. This work proposes a framework for survivable Wireless Sensor Network which predicts the data rate of the network. The past experience serves as an optimized way to traverse data in the network with efficient data rate. A primary dataset designed with the help of simulations is used for this work. Also, the robustness of best predictive model is checked with the help of N-fold cross-validation technique.","PeriodicalId":72026,"journal":{"name":"... International Conference on Contemporary Computing. IC3 (Conference)","volume":"1 1","pages":"1-6"},"PeriodicalIF":0.0,"publicationDate":"2018-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78383572","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-08-01DOI: 10.1109/IC3.2018.8530662
Vivek Singh Baghel, S. Bhavani
Multiple team formation problem is an interesting problem in the area of social network analysis which asks how experts can be assigned among multiple projects so that the effectiveness of allocation can be maximized. In this paper, we consider an approach based on genetic algorithm and sociometry for calculating the social relationship between experts in each project. Due to unavailability of data set, experimentation has been performed on synthetic data set given by J. H. Gutiérrez et al. [1] and results show that the allocation of experts for different projects is effective compared to the algorithm proposed in the literature.
{"title":"Multiple Team Formation Using an Evolutionary Approach","authors":"Vivek Singh Baghel, S. Bhavani","doi":"10.1109/IC3.2018.8530662","DOIUrl":"https://doi.org/10.1109/IC3.2018.8530662","url":null,"abstract":"Multiple team formation problem is an interesting problem in the area of social network analysis which asks how experts can be assigned among multiple projects so that the effectiveness of allocation can be maximized. In this paper, we consider an approach based on genetic algorithm and sociometry for calculating the social relationship between experts in each project. Due to unavailability of data set, experimentation has been performed on synthetic data set given by J. H. Gutiérrez et al. [1] and results show that the allocation of experts for different projects is effective compared to the algorithm proposed in the literature.","PeriodicalId":72026,"journal":{"name":"... International Conference on Contemporary Computing. IC3 (Conference)","volume":"96 1","pages":"1-6"},"PeriodicalIF":0.0,"publicationDate":"2018-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81054850","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2012-08-06DOI: 10.1007/978-3-642-32129-0_8
Manohar Pallantla, Alok Singh
{"title":"Grenade Explosion Method for Maximum Weight Clique Problem","authors":"Manohar Pallantla, Alok Singh","doi":"10.1007/978-3-642-32129-0_8","DOIUrl":"https://doi.org/10.1007/978-3-642-32129-0_8","url":null,"abstract":"","PeriodicalId":72026,"journal":{"name":"... International Conference on Contemporary Computing. IC3 (Conference)","volume":"25 1","pages":"20-27"},"PeriodicalIF":0.0,"publicationDate":"2012-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73761024","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2012-08-06DOI: 10.1007/978-3-642-32129-0_6
R. Hariharan
{"title":"Data-Driven Biology and Computation","authors":"R. Hariharan","doi":"10.1007/978-3-642-32129-0_6","DOIUrl":"https://doi.org/10.1007/978-3-642-32129-0_6","url":null,"abstract":"","PeriodicalId":72026,"journal":{"name":"... International Conference on Contemporary Computing. IC3 (Conference)","volume":"2 1","pages":"7"},"PeriodicalIF":0.0,"publicationDate":"2012-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"77603450","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2012-08-06DOI: 10.1007/978-3-642-32129-0_45
Anup Nandy, Soumik Mondal, P. Chakraborty, G. Nandi
{"title":"Development of a Robust Microcontroller Based Intelligent Prosthetic Limb","authors":"Anup Nandy, Soumik Mondal, P. Chakraborty, G. Nandi","doi":"10.1007/978-3-642-32129-0_45","DOIUrl":"https://doi.org/10.1007/978-3-642-32129-0_45","url":null,"abstract":"","PeriodicalId":72026,"journal":{"name":"... International Conference on Contemporary Computing. IC3 (Conference)","volume":"68 1","pages":"452-462"},"PeriodicalIF":0.0,"publicationDate":"2012-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79879293","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2012-08-06DOI: 10.1007/978-3-642-32129-0_7
Lavika Goel, D. Gupta, V. Panchal
{"title":"Dynamic Model of Blended Biogeography Based Optimization for Land Cover Feature Extraction","authors":"Lavika Goel, D. Gupta, V. Panchal","doi":"10.1007/978-3-642-32129-0_7","DOIUrl":"https://doi.org/10.1007/978-3-642-32129-0_7","url":null,"abstract":"","PeriodicalId":72026,"journal":{"name":"... International Conference on Contemporary Computing. IC3 (Conference)","volume":"15 1","pages":"8-19"},"PeriodicalIF":0.0,"publicationDate":"2012-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79030066","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2012-08-06DOI: 10.1007/978-3-642-32129-0_42
A. Abbas, K. A. A. Soufy
{"title":"Saturation Analysis of IEEE 802.11 EDCA for Ad Hoc Networks","authors":"A. Abbas, K. A. A. Soufy","doi":"10.1007/978-3-642-32129-0_42","DOIUrl":"https://doi.org/10.1007/978-3-642-32129-0_42","url":null,"abstract":"","PeriodicalId":72026,"journal":{"name":"... International Conference on Contemporary Computing. IC3 (Conference)","volume":"86 1","pages":"419-430"},"PeriodicalIF":0.0,"publicationDate":"2012-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79160667","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号