Image Captioning using Deep Stacked LSTMs Contextual Word Embeddings and Data Augmentation

Image Captioning, or the automatic generation of descriptions for images, is one of the core problems in Computer Vision . We propose to use Inception-ResNetConvolutional Neural Network as encoder to extract features from images . In addition to using ImageData Augmentation to avoid over-fitting, we useHorizontal and Vertical Flipping in addition to Perspective Transformations on the images .…

VisualGPT Data efficient Image Captioning by Balancing Visual Input and Linguistic Knowledge from Pretraining

Wepropose VisualGPT is a data-efficient image captioning model that leverages thelinguistic knowledge from a large pretrained language model (LM) When trained on 0.1%,0.5% and 1% of MSCOCO and Conceptual Captions, the proposed model, VisualGpt,surpasses strong image Captioning baselines . The proposed model outperforms the bestbaseline model by up to 10.8% CIDEr on MS COCO, up to 5.4% onConceptual Captionions .…

Actor Critic Method for High Dimensional Static Hamilton Jacobi Bellman Partial Differential Equations based on Neural Networks

We propose a novel numerical method for high dimensional Hamilton–Jacobi–Bellman (HJB) type elliptic partial differential equations . The HJB PDEs, reformulated as optimal control problems, are tackled by the actor-critic framework inspired by reinforcement learning, based on neuralnetwork parametrization of the value and control functions .…

Singular Euler Maclaurin expansion on multidimensional lattices

We extend the classical Euler-Maclaurin expansion to sums overmultidimensional lattices that involve functions with algebraic singularities . This offers a tool for the precise quantification of the effect of microscopicdiscreteness on macroscopic properties of a system . We connect our method to analytical number theory and show that all operator coefficients can be efficiently computed from derivatives of theEpstein zeta function .…

Improve Unscented Kalman Inversion With Low Rank Approximation and Reduced Order Model

The unscented Kalman inversion (UKI) is a generalderivative-free approach to solving the inverse problem . We aim to make the UKI more efficient in terms ofcomputational and memory costs for large scale inverse problems . We take advantage of the low-rank covariance structure to reduce the number of forwardproblem evaluations and the memory cost, related to the need to propagate largecovariance matrices .…

Elementary superexpressive activations

We call a finite family of activation functions superexpressive if anymultivariate continuous function can be approximated by a neural network that has a fixed architecture only depending on thenumber of input variables . To achieve any accuracy we only need to adjust weights, without increasing the number of neurons .…

Combining Spiking Neural Network and Artificial Neural Network for Enhanced Image Classification

Spiking neuralnetworks (SNNs) that more closely resemble biological brain synapses have attracted attention owing to their low power consumption . However, for continuous data values, they must employ a coding process to convert the valuesto spike trains . Thus, they have not yet exceeded the performance of artificialneural networks (ANNs), which handle such values directly .…

SoftTRR Protect Page Tables Against RowHammer Attacks using Software only Target Row Refresh

SoftTRR is motivated by DRAM-chip-based target row refresh (ChipTRR) but eliminates its main security limitation (i.e., ChipTRR tracks a limited number of rows and thus can be bypassed by many-sided hammer attacks . SofttrR protects an unlimited number of page tables by tracking memory accessesto the rows that are in close proximity to page-table rows and refreshing the rows once the tracked access count exceeds a pre-defined threshold .…

Software Architecture for Next Generation AI Planning Systems

Artificial Intelligence (AI) planning is a flourishing research and development discipline that provides powerful tools for searching a course of action that achieves some user goal . While these planning tools show excellent performance on benchmark planning problems, they represent challenging softwaresystems when it comes to their use and integration in real-world applications .…

Smart Contract Security a Practitioners Perspective

Smart contracts have been plagued by security incidents, which resulted insubstantial financial losses . We wondered how software practitioners build security into smart contracts in practice . We performed a mixture ofqualitative and quantitative studies with 13 interviewees and 156 surveyrespondents from 35 countries across six continents .…

Towards Evidence based Testability Measurements

Evaluating Software testability can assist software managers in optimizing budgets and identifying opportunities for refactoring . We introduce two novelevidence-based testability metrics, describe a prototype to compute them, and discuss initial findings on whether our measurements can reflect actual testability issues .…

Design techniques for a seamless information system architecture

The paper discusses design techniques for a seamless architecture ofinformation systems (IS) A seamless architecture is understood as such anarchitectural de-scription of an IS, that defines explicit connections betweenelements of architec-tural models of various architectural representations . Design techniques are based on the adaptive clustering method developed by the author, which allows one to bridge technological gaps between architecturalabstracts of different levels and to link architectural models in such a way .…

The Soccer Game bit by bit An information theoretic analyis

Soccer players are nodes discretely clustered into homogeneous groups . Players are grouped by physical proximity, supported by the intuitive notion that players use relative position as a key tactical tool to contribute to the team’s objectives . The model was applied to a set of matches from a major European national football league, with players’ coordinatessampled at 10Hz, resulting in approx.…

Hide and Seek Outwitting Community Detection Algorithms

Community affiliation of a node plays an important role in determining its position in a network . The current study focuses on hidingsuch sensitive communities so that the community affiliation of the targetednodes can be concealed . We introduce a novel method that greedily optimizes a node-centric objective function to determinethe rewiring strategy .…

Social Diffusion Sources Can Escape Detection

Influencing (and being influenced by) others indirectly through socialnetworks is fundamental to all human societies . Whether this happens through the diffusion of rumors, viruses, opinions, or know-how, finding the source is of persistent interest to people and an algorithmic challenge of much current research interest .…

On Stability and Convergence of Distributed Filters

A general distributed filter is constructed and itsestimation error dynamics is formulated . Conditions for achieving stable filtering operations are the same as those required in the centralized filtering setting . The concepts aredemonstrated in a Kalman filtering framework and validated using simulationexamples .…

SoftTRR Protect Page Tables Against RowHammer Attacks using Software only Target Row Refresh

SoftTRR is motivated by DRAM-chip-based target row refresh (ChipTRR) but eliminates its main security limitation (i.e., ChipTRR tracks a limited number of rows and thus can be bypassed by many-sided hammer attacks . SofttrR protects an unlimited number of page tables by tracking memory accessesto the rows that are in close proximity to page-table rows and refreshing the rows once the tracked access count exceeds a pre-defined threshold .…

SoftTRR Protect Page Tables Against RowHammer Attacks using Software only Target Row Refresh

SoftTRR is motivated by DRAM-chip-based target row refresh (ChipTRR) but eliminates its main security limitation (i.e., ChipTRR tracks a limited number of rows and thus can be bypassed by many-sided hammer attacks . SofttrR protects an unlimited number of page tables by tracking memory accessesto the rows that are in close proximity to page-table rows and refreshing the rows once the tracked access count exceeds a pre-defined threshold .…

SoftTRR Protect Page Tables Against RowHammer Attacks using Software only Target Row Refresh

SoftTRR is motivated by DRAM-chip-based target row refresh (ChipTRR) but eliminates its main security limitation (i.e., ChipTRR tracks a limited number of rows and thus can be bypassed by many-sided hammer attacks . SofttrR protects an unlimited number of page tables by tracking memory accessesto the rows that are in close proximity to page-table rows and refreshing the rows once the tracked access count exceeds a pre-defined threshold .…

Survey on Enterprise Internet of Things Systems E IoT A Security Perspective

As technology becomes more widely available, millions of users worldwide haveinstalled some form of smart device in their homes or workplaces . In contrast to commodity systems, purpose-builtEnterprise Internet-of-Things (E-IoT) systems offer a smart solution for more sophisticated applications (e.g., complete lighting control, A/V management, security) E-IiT systems are usually closed source, costly, require certified installers, and are overall more robust for their use cases .…

Demystifying Cryptocurrency Mining Attacks A Semi supervised Learning Approach Based on Digital Forensics and Dynamic Network Characteristics

Cryptocurrencies have emerged as a new form of digital money that has notescaped the eyes of cyber-attackers . Cyber-criminals have devised an exploitativetechnique of directly acquiring cryptocurrencies from benign users’ CPUswithout their knowledge through a process called crypto mining . The presence ofcrypto mining activities in a network is often an indicator of compromise of illegal usage of network resources for crypto mining purposes .…

A Ransomware Classification Framework Based on File Deletion and File Encryption Attack Structures

Ransomware has emerged as an infamous malware that has not escaped a lot of misinformation and inaccuracies from media hype . Victims are not sure whether or not topay a ransom demand without fully understanding the lurking consequences . The most severe categories CAT4 and CAT5 are better mitigated byexploiting encryption essentials while CAT3 can be effectively mitigated viareverse engineering.…

Content Confidentiality in Named Data Networking

Name-based access control scheme uses end-to-end encryption to protect data confidentiality . The scheme is based on the Ciphertext-policy attribute-based encryption (CP-ABE) We discuss the use of the scheme on the basis of two use-cases, and report overhead associated withit, based on our implementation .…

FASTEN Fair and Secure Distributed Voting Using Smart Contracts

Electing democratic representatives via voting has been a common mechanismsince the 17th century . Ballot voting, and in recent times,electronic voting via electronic voting machines (EVMs) improves fairness by relying on centralized trust . Homomorphic encryption-based voting protocols also assure fairness but cannot scale to large scale elections such as presidential elections .…

AI Augmented Behavior Analysis for Children with Developmental Disabilities Building Towards Precision Treatment

Autism spectrum disorder is a developmental disorder characterized bysignificant social, communication, and behavioral challenges . Individuals diagnosed with autism, intellectual, and developmental disabilities (AUIDD) typically require long-term care and targeted treatment and teaching . Successful integration of digitaltechnologies into clinical decision-making pipelines highlights the importance of augmenting teaching and treatments using novelalgorithms and high-fidelity sensors .…