Author

Wagh

754 papers · 1975–2026

1975peak 86 in 20252026

Frequent co-authors

Papers

showing 1–25 of 754 RSS

SecureNN: 3-Party Secure Computation for Neural Network Training

Sameer Wagh, Divya Gupta, Nishanth Chandran

Artificial Intelligence · Proceedings on Privacy Enhancing Technologies · July 2019 · cited 404× · DOI ↗ · more like this ·

Abstract Neural Networks (NN) provide a powerful method for machine learning training and inference. To effectively train, it is desirable for multiple parties to combine their data – however, doing so conflicts with data privacy. In this work, we provide novel three-party secure computation protocols for various NN building blocks such as matrix multiplication, convolutions, Rectified Linear Units, Maxpool, normalization and so on. This enables us to construct three-party secure protocols for training and inference of several NN architectures such that no single party learns any information about the data. Experimentally, we implement our system over Amazon EC2 servers in different settings. Our work advances the state-of-the-art of secure computation for neural networks in three ways: 1. Scalability: We are the first work to provide neural network training on Convolutional Neural…

FALCON: Honest-Majority Maliciously Secure Framework for Private Deep Learning

Sameer Wagh, Shruti Tople, Fabrice Benhamouda +3 more

cs.CR · April 2020 · cited 402× · arXiv ↗ · more like this ·

We propose Falcon, an end-to-end 3-party protocol for efficient private training and inference of large machine learning models. Falcon presents four main advantages - (i) It is highly expressive with support for high capacity networks such as VGG16 (ii) it supports batch normalization which is important for training complex networks such as AlexNet (iii) Falcon guarantees security with abort against malicious adversaries, assuming an honest majority (iv) Lastly, Falcon presents new theoretical insights for protocol design that make it highly efficient and allow it to outperform existing secure deep learning solutions. Compared to prior art for private inference, we are about 8x faster than SecureNN (PETS'19) on average and comparable to ABY3 (CCS'18). We are about 16-200x more communication efficient than either of these. For private training, we are about 6x faster than SecureNN, 4.4x…

The Stages of Drug Discovery and Development Process

Amol Bhalchandra Deore, Jayprabha R Dhumane, Rushikesh Wagh +1 more

Computational Theory and Mathematics · Asian Journal of Pharmaceutical Research and Development · 15 Dec 2019 · cited 199× · DOI ↗ · more like this ·

Drug discovery is a process which aims at identifying a compound therapeutically useful in curing and treating disease. This process involves the identification of candidates, synthesis, characterization, validation, optimization, screening and assays for therapeutic efficacy. Once a compound has shown its significance in these investigations, it will initiate the process of drug development earlier to clinical trials. New drug development process must continue through several stages in order to make a medicine that is safe, effective, and has approved all regulatory requirements. One overall theme of our article is that the process is sufficiently long, complex, and expensive so that many biological targets must be considered for every new medicine ultimately approved for clinical use and new research tools may be needed to investigate each new target. From initial discovery to a…

Effective deep learning approaches for summarization of legal texts

Deepa Anand, Rupali Sunil Wagh

Artificial Intelligence · Journal of King Saud University - Computer and Information Sciences · 7 Dec 2019 · cited 140× · DOI ↗ · more like this ·

The availability of legal judgment documents in digital form offers numerous opportunities for information extraction and application. Automatic summarization of these legal texts is a crucial and a challenging task due to the unusual structure and high complexity of these documents. Previous approaches in this direction have relied on huge labelled datasets, using hand engineered features, leveraging on domain knowledge and focussed their attention on a narrow sub-domain for increased effectiveness. In this paper, we propose simple generic techniques using neural network for the summarization task for Indian legal judgment documents. We explore two neural network architectures for this task utilizing the word and sentence embeddings for capturing the semantics. The main advantage of the proposed approaches is that they do not rely on hand crafted features, or domain specific knowledge…

Attendance system based on face recognition using eigen face and PCA algorithms

Priyanka Wagh, Roshani Thakare, Jagruti Chaudhari +1 more

Computer Vision and Pattern Recognition · October 2015 · cited 118× · DOI ↗ · more like this ·

The attendance maintaining system is difficult process if it is done manually. The smart and automated attendance system for managing the attendance can be implemented using the various ways of biometrics. Face recognition is one of them. By using this system, the issue of fake attendance and proxies can be solved. In the previous face recognition based attendance system, there were some disadvantages like intensity of light problem and head pose problem. Therefore to overcome these issues, various techniques like illumination invariant, Viola and Jones algorithm, Principle component analysis are used. The major steps in this system are detecting the faces and recognizing them. After these, the comparison of detected faces can be done by crosschecking with the database of student's faces. This smart system will be an effective way to maintain the attendance and records of students.

Survey on Sentiment Analysis using Twitter Dataset

Rasika Wagh, Payal K. Punde

Artificial Intelligence · 2018 Second International Conference on Electronics, Communication and Aerospace Technology (ICECA) · March 2018 · cited 104× · DOI ↗ · more like this ·

Social networking sites like twitter have millions of people share their thoughts day by day as tweets. As tweet is characteristic short and basic way of expression. So in this review paper we focused on sentiment analysis of Twitter data. The Sentiment Analysis sees as area of text data mining and NLP. The research of sentiment analysis of Twitter data can be performed in different aspects. This paper shows sentiment analysis types and techniques used to perform extraction of sentiment from tweets. In this survey paper, we have taken comparative study of different techniques and approaches of sentiment analysis having twitter as a data.

ELSA: Secure Aggregation for Federated Learning with Malicious Actors

Mayank Rathee, Conghao Shen, Sameer Wagh +1 more

Artificial Intelligence · May 2023 · cited 95× · DOI ↗ · more like this ·

Federated learning (FL) is an increasingly popular approach for machine learning (ML) in cases where the training dataset is highly distributed. Clients perform local training on their datasets and the updates are then aggregated into the global model. Existing protocols for aggregation are either inefficient, or don’t consider the case of malicious actors in the system. This is a major barrier in making FL an ideal solution for privacy-sensitive ML applications. We present Elsa, a secure aggregation protocol for FL, which breaks this barrier - it is efficient and addresses the existence of malicious actors at the core of its design. Similar to prior work on Prio and Prio+, Elsa provides a novel secure aggregation protocol built out of distributed trust across two servers that keeps individual client updates private as long as one server is honest, defends against malicious clients, and…

Les Houches "Physics at TeV Colliders 2005'' Beyond the Standard Model working group: summary report

B. C. Allanach, C. Grojean, P. Skands +110 more

hep-ph · 21 Feb 2006 · cited 91× · arXiv ↗ · more like this ·

The work contained herein constitutes a report of the "Beyond the Standard Model'' working group for the Workshop "Physics at TeV Colliders", Les Houches, France, 2-20 May, 2005. We present reviews of current topics as well as original research carried out for the workshop. Supersymmetric and non-supersymmetric models are studied, as well as computational tools designed in order to facilitate their phenomenology.

Balancing Curricular and Pedagogical Needs in Computational Construction Kits: Lessons From the DeltaTick Project

Michelle Hoda Wilkerson, Aditi Wagh, Uri Wilensky

Computer Science Applications · Science Education · 15 Apr 2015 · cited 87× · DOI ↗ · more like this ·

To successfully integrate simulation and computational methods into K–12 STEM education, learning environments should be designed to help educators maintain balance between (a) addressing curricular content and practices and (b) attending to student knowledge and interests. We describe DeltaTick, a graphical simulation construction interface for the NetLogo modeling environment designed to make computational model construction a more accessible and responsive part of science and mathematics curricular activities through domain-specific, customizable construction libraries. With DeltaTick, learners assemble and reassemble predefined sets of “behavior blocks” to build simulations that represent a particular domain of study. When needed, blocks can be added, adjusted, or replaced to better reflect student knowledge, interests, or questions. We present coding analyses and vignettes from…

Towards Probabilistic Verification of Machine Unlearning

David Marco Sommer, Liwei Song, Sameer Wagh +1 more

cs.CR · March 2020 · cited 86× · arXiv ↗ · more like this ·

The right to be forgotten, also known as the right to erasure, is the right of individuals to have their data erased from an entity storing it. The status of this long held notion was legally solidified recently by the General Data Protection Regulation (GDPR) in the European Union. Consequently, there is a need for mechanisms whereby users can verify if service providers comply with their deletion requests. In this work, we take the first step in proposing a formal framework to study the design of such verification mechanisms for data deletion requests -- also known as machine unlearning -- in the context of systems that provide machine learning as a service (MLaaS). Our framework allows the rigorous quantification of any verification mechanism based on standard hypothesis testing. Furthermore, we propose a novel backdoor-based verification mechanism and demonstrate its effectiveness…

SecureNN: Efficient and Private Neural Network Training.

Sameer Wagh, Divya Gupta, Nishanth Chandran

Artificial Intelligence · IACR Cryptology ePrint Archive · 2018 · cited 80× · OpenAlex ↗ · more like this ·

Neural Networks (NN) provide a powerful method for machine learning training and inference. To effectively train, it is desirable for multiple parties to combine their data — however, doing so conflicts with data privacy. In this work, we provide novel three-party secure computation protocols for various NN building blocks such as matrix multiplication, convolutions, Rectified Linear Units, Maxpool, normalization and so on. This enables us to construct three-party secure protocols for training and inference of several NN architectures such that no single party learns any information about the data. Experimentally, we implement our system over Amazon EC2 servers in different settings. Our work advances the state-of-the-art of secure computation for neural networks in three ways: Scalability: We are the first work to provide neural network training on Convolutional Neural Networks (CNNs)…

A Comparative Study of SOAP Vs REST Web Services Provisioning Techniques for Mobile Host

Kishor S. Wagh, Ravindra C. Thool

Information Systems · Journal of information engineering and applications · 2012 · cited 71× · OpenAlex ↗ · more like this ·

Web services are moving towards mobile wireless world as a new emerging technology for mobile applications communication. Now mobile devices can operate as web service providers due to the advancement in mobile device capabilities and it has also become possible due to improvement in web service development technologies and wireless communication techniques. Today’s most important need is to provide uninterrupted, lightweight, continuous web services to resource constrained mobile device in wireless environment. This paper presents a detailed comparison between two frameworks used for providing web services through SOAP and REST and also discusses the problems and challenges in these two frameworks. With the help of comparison we can decide which frame work is most suitable for wireless environment and fulfills the current needs of accessing lightweight mobile web services continuously…

Priority Based Dynamic Resource Allocation in Cloud Computing

Chandrashekhar S. Pawar, R. B. Wagh

Information Systems · December 2012 · cited 70× · DOI ↗ · more like this ·

Today Cloud computing is on demand as it offers dynamic flexible resource allocation, for reliable and guaranteed services in pay-as-you-use manner, to Cloud service users. So there must be a provision that all resources are made available to requesting users in efficient manner to satisfy their needs. This resource provision is done by considering the Service Level Agreements (SLA) and with the help of parallel processing. Recent work considers various strategies with single SLA parameter. Hence by considering multiple SLA parameter and resource allocation by preemption mechanism for high priority task execution can improve the resource utilization in Cloud. In this paper we propose an algorithm which considered Preemptable task execution and multiple SLA parameters such as memory, network bandwidth, and required CPU time. An obtained experimental results show that in a situation where…

Impact of Remittanceson Poverty and Financial Development in Sub-Saharan Africa

Sanjeev Gupta, Catherine Pattillo, Smita Wagh +3 more

Sociology and Political Science · IMF Working Paper · 2007 · cited 66× · DOI ↗ · more like this ·

This paper assesses the impact of the steadily growing remittance flows to sub-Saharan Africa (SSA). Though the region receives only a small portion of the total recorded remittances to developing countries, and the volume of aid flows to SSA swamps remittances, this paper finds that remittances, which are a stable, private transfer, have a direct poverty mitigating effect, and promote financial development. These findings hold even after factoring in the reverse causality between remittances, poverty and financial development. The paper posits that formalizing such flows can serve as an effective access point for ""unbanked"" individuals and households, and that the effective use of such flows can mitigate the costs of skilled out-migration in SSA

Zero-inflated models and estimation in zero-inflated Poisson distribution

Yogita S. Wagh, Kirtee K. Kamalja

Statistics and Probability · Communications in Statistics - Simulation and Computation · 14 Jun 2017 · cited 65× · DOI ↗ · more like this ·

In this paper, we briefly overview different zero-inflated probability distributions. We compare the performance of the estimates of Poisson, Generalized Poisson, ZIP, ZIGP and ZINB models through Mean square error (MSE), bias and Standard error (SE) when the samples are generated from ZIP distribution. We propose a new estimator referred to as probability estimator (PE) of inflation parameter of ZIP distribution based on moment estimator (ME) of the mean parameter and compare its performance with ME and maximum likelihood estimator (MLE) through a simulation study. We use the PE along with ME and MLE to fit ZIP distribution to various zero-inflated datasets and observe that the results do not differ significantly. We recommend using PE in place of MLE since it is easy to calculate and the simulation study in this paper demonstrates that the PE performs as good as MLE irrespective of…

Robot path planning using intersecting convex shapes: Analysis and simulation

Jitendra Kumar Singh, M. Wagh

Computer Vision and Pattern Recognition · IEEE Journal on Robotics and Automation · April 1987 · cited 61× · DOI ↗ · more like this ·

An automated path planning algorithm for a mobile robot in a structured environment is presented. An algorithm based on the Quine-McCluskey method of finding prime implicants in a logical expression is used to isolate all the largest rectangular free convex areas in a specified environment. The free convex areas are represented as nodes in a graph, and a graph traversal strategy that dynamically allocates costs to graph paths is used. Complexity of the algorithm and a strategy to trade optimality for smaller computation time are discussed.

Bridging inquiry‐based science and constructionism: Exploring the alignment between students tinkering with code of computational models and goals of inquiry

Aditi Wagh, Kate Cook‐Whitt, Uri Wilensky

Computer Science Applications · Journal of Research in Science Teaching · 25 Jan 2017 · cited 59× · DOI ↗ · more like this ·

Abstract Research on the design of learning environments for K‐12 science education has been informed by two bodies of literature: inquiry‐based science and Constructionism. Inquiry‐based science has emphasized engagement in activities that reflect authentic scientific practices. Constructionism has focused on designing intuitively accessible authoring environments and microworlds that embody the structure of a content domain in program code. Learners build, extend, or explore artifacts to make sense of underlying mechanisms. In this paper, we bridge these bodies of work to argue that interacting with program code of a computational model can facilitate engagement in inquiry‐based science. Tinkering with code involves students playfully manipulating the code of a computational model to generate and pursue goals or questions in the model. We use data from video‐recorded interviews with…

Are Donor Countries Giving More or Less Aid?

Sanjeev Gupta, Catherine Pattillo, Smita Wagh

Development · Review of Development Economics · 25 Jul 2006 · cited 54× · DOI ↗ · more like this ·

Abstract The volume of foreign aid has increased during the last four decades, albeit with interruptions in certain years. Over time, the major recipients have changed: while the share of aid to Asia has diminished since the 1980s, that destined for sub‐Saharan Africa has grown. There is some evidence that, since the late 1990s, debt relief has assumed a larger share of the increased aid flows to sub‐Saharan Africa. The share of technical cooperation—a component of aid that is viewed as being driven by donors—has risen. More recently, there has been an increased emphasis on providing budget support to recipient governments, especially in the form of debt relief. Donor harmonization, national ownership of development plans, and sound policies on the part of the recipients are crucial for the aid to be effective in reducing poverty.

Athena: Probabilistic Verification of Machine Unlearning

David Sommer, Liwei Song, Sameer Wagh +1 more

Artificial Intelligence · Proceedings on Privacy Enhancing Technologies · July 2022 · cited 52× · DOI ↗ · more like this ·

The right to be forgotten, also known as the right to erasure, is the right of individuals to have their data erased from an entity storing it. The status of this long held notion was legally solidified recently by the General Data Protection Regulation (GDPR) in the European Union. As a consequence, there is a need for mechanisms whereby users can verify if service providers comply with their deletion requests. In this work, we take the first step in proposing a formal framework, called Athena, to study the design of such verification mechanisms for data deletion requests – also known as machine unlearning – in the context of systems that provide machine learning as a service (MLaaS). Athena allows the rigorous quantification of any verification mechanism based on hypothesis testing. Furthermore, we propose a novel verification mechanism that leverages backdoors and demonstrate its…

A class of translation invariant transforms

M. Wagh, Sharad V. Kanetkar

Computer Vision and Pattern Recognition · IEEE Transactions on Acoustics Speech and Signal Processing · April 1977 · cited 51× · DOI ↗ · more like this ·

A class of translation invariant transforms containing the R-transform is defined, and it is shown that a particular member of this class is superior to the R-transform for pattern recognition applications.

Priority based dynamic resource allocation in Cloud computing with modified waiting queue

Chandrashekhar S. Pawar, R. B. Wagh

Information Systems · March 2013 · cited 48× · DOI ↗ · more like this ·

Making Remittances Work for Africa

Sanjeev Gupta, Catherine Pattillo, Smita Wagh

Sociology and Political Science · Finance & development · 22 Jun 2007 · cited 48× · DOI ↗ · more like this ·

Remittances flowing into developing countries are attracting increasing attention because of their rising volume and their impact on recipient countries. In 2005, they totaled $188 billion—twice the amount of official assistance developing countries received. Moreover, there is evidence that such flows are underreported. Indeed, remittances through informal channels could add at least 50 percent to global recorded flows. Most of the reported flows go to regions other than sub-Saharan Africa (SSA), but SSA has still been part of the overall rising global trend. Between 2000 and 2005, remittances to the region increased by more than 55 percent, to nearly $7 billion, whereas they increased for developing countries as a group by 81 percent.

Building support for gender equality among young adolescents in school: findings from Mumbai India.

Pranita Achyut, Nazlı Bhatia, Singh Ak +9 more

Safety Research · 2011 · cited 43× · OpenAlex ↗ · more like this ·

There is increasing recognition that to reduce gender inequality -- a goal fundamental to improving a country’s overall health and development -- programs must start with youth. Yet there has been limited engagement of both girls and boys during early adolescence to challenge and shift gender norms that contribute to girls and women having less worth opportunities and decision-making ability than boys and men. Such inequitable gender norms can have a host of harmful consequences for girls and boys during childhood and beyond including poor sexual and reproductive health (SRH) outcomes violence and school drop out. In the last few years a growing body of evidence has emerged in numerous settings including India linking individual attitudes around gender to SRH behaviors and the use and experience of violence. Because gender socialization of both boys and girls begins early in India it is…

Healthcare Chatbot System using Artificial Intelligence

Nikita Vijay Shinde, Aniket Akhade, Pranali Bagad +3 more

Artificial Intelligence · 2021 5th International Conference on Trends in Electronics and Informatics (ICOEI) · 3 Jun 2021 · cited 41× · DOI ↗ · more like this ·

This paper presents Healthcare Chatbot using Artificial Intelligence that can make a human-system interaction to resolve basic queries regarding health parameters before consulting a doctor. The actual purpose behind this work is to work on the user's symptoms and to provide medical suggestions according to it, to reduce the time and cost required for the process. It is a system that communicates with users using Natural Language processing rather we can say it is a computer program that uses NLP for interaction. The chatbot works on provided input by the user, It takes sentence keywords and makes decisions to solve the user's query and answers it accordingly. Some of these calculations like Rank calculation and sentence similarity are performed with the help of TF-IDF, Stemming, n-grams and cosine similarity. Basically, our system uses machine learning techniques to solve user's…

Camouflage: Memory Traffic Shaping to Mitigate Timing Attacks

Yanqi Zhou, Sameer Wagh, Prateek Mittal +1 more

Artificial Intelligence · February 2017 · cited 38× · DOI ↗ · more like this ·

Information leaks based on timing side channels in computing devices have serious consequences for user security and privacy. In particular, malicious applications in multi-user systems such as data centers and cloud-computing environments can exploit memory timing as a side channel to infer a victim's program access patterns/phases. Memory timing channels can also be exploited for covert communications by an adversary. We propose Camouflage, a hardware solution to mitigate timing channel attacks not only in the memory system, but also along the path to and from the memory system (e.g. NoC, memory scheduler queues). Camouflage introduces the novel idea of shaping memory requests' and responses' interarrival time into a pre-determined distribution for security purposes, even creating additional fake traffic if needed. This limits untrusted parties (either cloud providers or coscheduled…