Search

Since the development of semiconductor technologies, exascale computing and its associated applications have required increasing degrees of efficiency. Semiconductor-transistor-based circuits (STbCs) have struggled in increasing the GHz frequency. Emerging as an alternative to STbC, the superconducting electrons (SCE) technology promises higher-speed clock frequencies at ultra-low power consumption. The rapid single flux quantum (RSFQ) circuits have a theoretical potential for three orders of magnitude reduction in power while operating at clock frequencies higher than 100 GHz. Although the security in semiconductor technology has been extensively researched and developed, the security design in the superconducting field requires field demands attention. In this paper, C-SAR is presented that aims to protect the superconducting circuit electronics from Boolean satisfiability (SAT) based attacks. The SAT attack is an attack that can break all the existing combinational logic locking techniques. C-SAR can immunize against SAT attacks by increasing the key search space and prolonging the clock cycles of attack inputs. Even in the worst case of C-SAR, in face of S-SAT a specially designed SAT attack, C-SAR can also soar the attack cost exponentially with key bits first, then linearly with the length of camouflaged DFF array. We have shown in this work that the cost of C-SAR is manageable as it only linearly increases as a function of key bits.

Coupled partial differential equations (PDEs) are key tasks in modeling the complex dynamics of many physical processes. Recently, neural operators have shown the ability to solve PDEs by learning the integral kernel directly in Fourier/Wavelet space, so the difficulty for solving the coupled PDEs depends on dealing with the coupled mappings between the functions. Towards this end, we propose a \textit{coupled multiwavelets neural operator} (CMWNO) learning scheme by decoupling the coupled integral kernels during the multiwavelet decomposition and reconstruction procedures in the Wavelet space. The proposed model achieves significantly higher accuracy compared to previous learning-based solvers in solving the coupled PDEs including Gray-Scott (GS) equations and the non-local mean field game (MFG) problem. According to our experimental results, the proposed model exhibits a $2\times \sim 4\times$ improvement relative $L$2 error compared to the best results from the state-of-the-art models., Comment: Accepted to ICLR 2023

Novel double cross-linked (DC) hydrogels with pH-/temperature-sensitive properties were designed and developed. Therefore, linear pH-sensitive poly(methyl vinyl ether-alt-maleic acid) (P(VME/MA)) macromolecules were absorbed within a thermosensitive poly(N-isopropylacrylamide-co-hydroxyethylacrylamide)-hydrogel (PNH) and, subsequently, cross-linked together through a solvent-free thermal method. As a novelty, double cross-linked hydrogels were obtained from previously purified polymers in the absence of any solvent or cross-linking agent, which are generally harmful for the body. The new DC structures were characterized by FT–IR spectroscopy, SEM, swelling kinetic measurements, and mechanical tests. The resulting scaffolds exhibited interconnected pores and a flexible pattern, compared to the brittle structure of conventional PNH. The swelling kinetics of DC hydrogels were deeply affected by temperature (25 and 37 °C) and pH (7.4 and 1.2). Furthermore, the hydrogels absorbed a great amount of water in a basic environment and displayed improved mechanical properties. Metoclopramide (Met) was loaded within DC hydrogels as a model drug to investigate the ability of the support to control the drug release rate. The results obtained recommended them as convenient platforms for the oral administration of drugs, with the release of the largest part of the active principle occurring in the colon.

Thermally-induced gelling systems based on Poloxamer 407 (PL) and polysaccharides are known for their biomedical applications; however, phase separation frequently occurs in mixtures of poloxamer and neutral polysaccharides. In the present paper, the carboxymethyl pullulan (CMP) (here synthesized) was proposed for compatibilization with poloxamer (PL). The miscibility between PL and CMP in dilute aqueous solution was studied by capillary viscometry. CMP with substitution degrees higher than 0.5 proved to be compatible with PL. The thermogelation of concentrated PL solutions (17%) in the presence of CMP was monitored by the tube inversion method, texture analysis and rheology. The micellization and gelation of PL in the absence or in the presence of CMP were also studied by dynamic light scattering. The critical micelle temperature and sol–gel transition temperature decrease with the addition of CMP, but the concentration of CMP has a peculiar influence on the rheological parameters of the gels. In fact, low concentrations of CMP decrease the gel strength. With a further increase in polyelectrolyte concentration, the gel strength increases until 1% CMP, then the rheological parameters are lowered again. At 37 °C, the gels are able to recover the initial network structure after high deformations, showing a reversible healing process.

Assessing the stability of biological system models has aided in uncovering a plethora of new insights in genetics, neuroscience, and medicine. In this paper, we focus on analyzing the stability of neurological signals, including electroencephalogram (EEG) signals. Interestingly, spatiotemporal discrete-time linear fractional-order systems (DTLFOS) have been shown to accurately and efficiently represent a variety of neurological and physiological signals. Here, we leverage the conditions for stability of DTLFOS to assess a real-world EEG data set. By analyzing the stability of EEG signals during movement and rest tasks, we provide evidence of the usefulness of the quantification of stability as a bio-marker for cognitive motor control.

To enable heterogeneous computing systems with autonomous programming and optimization capabilities, we propose a unified, end-to-end, programmable graph representation learning (PGL) framework that is capable of mining the complexity of high-level programs down to the universal intermediate representation, extracting the specific computational patterns and predicting which code segments would run best on a specific core in heterogeneous hardware platforms. The proposed framework extracts multi-fractal topological features from code graphs, utilizes graph autoencoders to learn how to partition the graph into computational kernels, and exploits graph neural networks (GNN) to predict the correct assignment to a processor type. In the evaluation, we validate the PGL framework and demonstrate a maximum speedup of 6.42x compared to the thread-based execution, and 2.02x compared to the state-of-the-art technique.

Reinforcement learning (RL) is a technique to learn the control policy for an agent that interacts with a stochastic environment. In any given state, the agent takes some action, and the environment determines the probability distribution over the next state as well as gives the agent some reward. Most RL algorithms typically assume that the environment satisfies Markov assumptions (i.e. the probability distribution over the next state depends only on the current state). In this paper, we propose a model-based RL technique for a system that has non-Markovian dynamics. Such environments are common in many real-world applications such as in human physiology, biological systems, material science, and population dynamics. Model-based RL (MBRL) techniques typically try to simultaneously learn a model of the environment from the data, as well as try to identify an optimal policy for the learned model. We propose a technique where the non-Markovianity of the system is modeled through a fractional dynamical system. We show that we can quantify the difference in the performance of an MBRL algorithm that uses bounded horizon model predictive control from the optimal policy. Finally, we demonstrate our proposed framework on a pharmacokinetic model of human blood glucose dynamics and show that our fractional models can capture distant correlations on real-world datasets., 14 pages, 3 figures, CDC2021

Learning task-specific representations of persistence diagrams is an important problem in topological data analysis and machine learning. However, current state of the art methods are restricted in terms of their expressivity as they are focused on Euclidean representations. Persistence diagrams often contain features of infinite persistence (i.e., essential features) and Euclidean spaces shrink their importance relative to non-essential features because they cannot assign infinite distance to finite points. To deal with this issue, we propose a method to learn representations of persistence diagrams on hyperbolic spaces, more specifically on the Poincare ball. By representing features of infinite persistence infinitesimally close to the boundary of the ball, their distance to non-essential features approaches infinity, thereby their relative importance is preserved. This is achieved without utilizing extremely high values for the learnable parameters, thus the representation can be fed into downstream optimization methods and trained efficiently in an end-to-end fashion. We present experimental results on graph and image classification tasks and show that the performance of our method is on par with or exceeds the performance of other state of the art methods.

We propose a framework for the design and optimization of a secure self-optimizing, self-adapting system-on-chip (S4oC) architecture. The goal is to minimize the impact of attacks such as hardware Trojan and side-channel, by making real-time adjustments. S4oC learns to reconfigure itself, subject to various security measures and attacks, some of which possibly unknown at design time. Furthermore, the data types and patterns of the target applications, environmental conditions, and sources of variations are incorporated. S4oC is a manycore system, modeled as a four-layer graph, representing the model of computation (MoCp), model of connection (MoCn), model of memory (MoM) and model of storage (MoS), with a large number of elements including heterogeneous reconfigurable processing elements in MoCp, and memory elements in the MoM layer. Security driven community detection, and neural networks are utilized for application task clustering, and distributed reinforcement learning (RL) for task mapping., Comment: This is an invited paper to ISCAS 2020

Cloud computing has attracted both end-users and Cloud Service Providers (CSPs) in recent years. Improving resource utilization rate (RUtR), such as CPU and memory usages on servers, while maintaining Quality-of-Service (QoS) is one key challenge faced by CSPs with warehouse-scale data centers. Prior works proposed various algorithms to reduce energy cost or to improve RUtR, which either lack the fine-grained task scheduling capabilities, or fail to take a comprehensive system model into consideration. This article presents H2O-Cloud, a Hierarchical and Hybrid Online task scheduling framework for warehouse-scale CSPs, to improve resource usage effectiveness while maintaining QoS. H2O-Cloud is highly scalable and considers comprehensive information such as various workload scenarios, cloud platform configurations, user request information and dynamic pricing model. The hierarchy and hybridity of the framework, combined with its deep reinforcement learning (DRL) engines, enable H2O-Cloud to efficiently start on-the-go scheduling and learning in an unpredictable environment without pre-training. Our experiments confirm the high efficiency of the proposed H2O-Cloud when compared to baseline approaches, in terms of energy and cost while maintaining QoS. Compared with a state-of-the-art DRL-based algorithm, H2O-Cloud achieves up to 201.17% energy cost efficiency improvement, 47.88% energy efficiency improvement and 551.76% reward rate improvement., Comment: 12 pages, 5 figures. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems. 2019 Jul 23

Nowadays, the students’ perception is that everything should be available 24 hours/7 days a week. The exposure to limitless information and experiences has completely modified their view of the world in ways unfamiliar to other generations. Therefore educators must adapt the content and teaching/lecturing methods according to their expectations. One way of enhancing the learning experience is to be aware of the current attitudes and behaviors of today’s students and to rely more and more on the technological improvements in educational strategies. Goal oriented and intellectually inquisitive students prefer experimental activities, which are structured, involve teamwork and virtual collaboration instead of simple narrations and monologues. In addition, students are sometimes involved in part time or even full time jobs that significantly reduce their time of preparation for educational activities. The paper presents an innovative cloud computing approach and its implementation within HyperTraining, a collaborative learning framework for career development and internship programmes. It also illustrates how such a system increases the performance in special education and training programmes, as well as the manner the private companies can help universities and colleges in the education act. Nevertheless, it highlights the importance of blended learning concepts and tools used for knowledge building in the individual study process or technical skills and competences developed within practice activities during special educational processes such as laboratories, projects, even internship programmes, continuing the work started in [1], [2].

We consider the problem of maximizing non-negative non-decreasing set functions. Although most of the recent work focus on exploiting submodularity, it turns out that several objectives we encounter in practice are not submodular. Nonetheless, often we leverage the greedy algorithms used in submodular functions to determine a solution to the non-submodular functions. Hereafter, we propose to address the original problem by \emph{approximating} the non-submodular function and analyze the incurred error, as well as the performance trade-offs. To quantify the approximation error, we introduce a novel concept of $\delta$-approximation of a function, which we used to define the space of submodular functions that lie within an approximation error. We provide necessary conditions on the existence of such $\delta$-approximation functions, which might not be unique. Consequently, we characterize this subspace which we refer to as \emph{region of submodularity}. Furthermore, submodular functions are known to lead to different sub-optimality guarantees, so we generalize those dependencies upon a $\delta$-approximation into the notion of \emph{greedy curvature}. Finally, we used this latter notion to simplify some of the existing results and efficiently (i.e., linear complexity) determine tightened bounds on the sub-optimality guarantees using objective functions commonly used in practical setups and validate them using real data., Comment: 23 pages, 8 figures, submitted to journal

Microservices deployment of Rich Internet Applications is not a new research topic; many articles have been published and related innovative technologies released in the last years. Thus, the research work presented in this article is focused on surveying the current collaborative learning concepts and technologies and proposing a new virtual classware approach, implemented and deployed as a set of independent components responsible for specific collaborative learning services. While engineering education methodologies are complex, flexibility or openness should be considered as high priorities. Thus, an actual implementation, according to blended training scenarios, has been presented.

Supplemental material

The proliferation of complex phenomena and the tightening competition for limited resources are two fundamental challenges for the modeling, analysis, and optimization of dynamical processes taking place in networked environments/architectures. Modes of collective and competitive behavior can be noticed across a wide array of social, biological, and technological contexts. From urban crowds to bacterial colonies, from brain neurons to human cells and even electron-hole interactions in semiconductors, dynamical phase transitions influence the macroscopic behavior of complex networks. To address these challenges, we focus on understanding, modeling, analyzing, and optimizing large-scale interconnected systems, such as future thousand-core Networks-on-Chip (NoC), biologically propelled Cyber-Physical Systems (CPS) consisting of micro-robotic swarms, or biological networks of stem cells, for performance, power, or fault-tolerance. Enabled by recent advances in CMOS technology, the integration of tens and soon thousands of heterogeneous processing cores communicating via the NoC paradigm brings into the discussion the traffic modeling problem. Traffic modeling is of crucial importance for both static and dynamic NoC design and optimization problems such as topology selection and resource allocation, mapping, scheduling, routing or power management. The approaches proposed so far exhibit major limitations. For example, many of the queuing theory based modeling and optimization approaches proposed for NoC architectures ignore the traffic characteristics (e.g., non-stationarity, fractality). In many situations, these models lead to buffer overflow or deadline missing situations and so poor (non-optimal) performance levels. In this thesis, we show how this state of affairs can be changed by embracing a statistical physics inspired approach in order to insure accurate network traffc characterization. By using an analogy between a thermodynamic gas and a networked multicore architecture, our model captures the relevant traffc characteristics (e.g., fractality, non-stationarity) via a dynamical master equation. Our approach not only leads to a more accurate estimation of performance metrics over Markovian models, but also helps at defining new state space model for dynamical systems that can be used for online optimization. Besides being an important contribution by itself, this radically new approach enhances the efficiency of resource allocation in nanoscale networks and overcomes the prior limitations of performance analysis approaches based on queuing models. Building on statistical physics grounds, we model and analyze a biologically inspired communication protocol, namely the stochastic communication protocol. Under the stochastic communication protocol aiming at mitigating the nanoscale challenges (e.g., particle hits, cross-talk) in multi-core platforms, each node in the network disseminates packets multiple times via multiple paths. Hence, fault-tolerance is enforced at system-level by exploiting path diversity. To characterize such a dynamic behavior, we concentrate on estimating various performance metrics via a master equation approach capturing communication events such as packet duplication, packet transmission, and packet corruption. The proposed statistical physics model allows us to identify the benefits of this protocol for future communication fabrics. To address the power consumption issues in large scale on-chip networks, we formulate the power and peak temperature management of heterogeneous NoC platforms as constrained finite horizon fractal optimal control problem. We show not only that fractal characteristics can be accounted for via fractional calculus state space models, but also that the online controller can be reduced to a linear program and efficiently computed via parallel algorithms. Our approach not only leads to significant power savings, but it also opens new avenues for dynamic optimization of large-scale systems exhibiting fractal dynamics. Based on the proposed framework for modeling, analysis, and optimization of dynamical processes taking place on networked architectures, we formulate general guidelines for CPS design. As a concrete CPS example, we consider the design problem of the control algorithm of a pacemaker, which takes into account at run-time the fractal characteristics exhibited by heart rate variability. In summary, this thesis offers a statistical physics view on using the network-paradigm in multicore and cyber-physical system design. The results and discussion presented herein can be further extended to other classes of systems and applications. One research direction is represented by modeling and optimization of bacteria propelled micro-robotic swarms. Another research direction concerns the modeling of human dynamic processes such as car traffic, which can enable road structure optimization. Moreover, by relying on this statistical physics inspired framework, we can define models for biological communication and heterogeneous population growth with applications in regenerative medicine.

Normalized cross-correlation is a standard way of looking for examples of a target in an image. A way to avoid scanning the entire image is explored. It exploits the observation that in the locality of a target object, the watershed of a smoothed image is similar to the watershed of a smoothed image of the target. Therefore, it is only necessary to align the target watershed with the image watershed and track the one along the other. This can be faster than scanning the entire image.© (1995) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

A method of based on genetic algorithms for finding an optima soft morphological filter for specific situation is presented. The behavior of soft morphological filters is illustrated by empirical results. Some optimal parameter will also be proposed for MAE and MSE error criteria.© (1994) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Watershed transformation is used in morphological image segmentation. This transformation could be considered as a topographic region growing method. Recently, fast watershed algorithms have been proposed for general purpose computers. They are based on immersion simulations of the image surface, which is considered as a topographic relief. In such a model, the greylevel values of pixels stand for altitude values on the relief. In this paper, the operation of the present fast watershed algorithms is analyzed and a new extension is proposed. Drawbacks of the present algorithms are pointed out, studied, and illustrated with test images. These problems lead, in several cases, to a loss of information about image details and structures or even to unprocessed areas in the image. The new watershed algorithm overcomes these deficiencies and preserves more information about image details. The new algorithm is based on a split-and-merge scheme. It constantly monitors the presence of isolated areas during the immersion simulation, considering them as new catchment basins. Application of the split-and-merge watershed algorithm to marker-based image segmentation is discussed.

In this paper a parallel implementation of Meyer watershed algorithm, which does not construct watersheds between catchment basins [2, 10], is developed. A single program multiple data (SPMD) approach is used. Although the watershed algorithm is highly sequential, we have found that much parallelism can be exploited in a loosely coupled system by dividing the task to a set of processes supervised by a master process. Specialized flooding and communication methods have been developed for an efficient parallel computation of watersheds.

This paper presents a fast center weighted rank order filtering algorithm with logarithmic worst case time complexity with respect to the window size. The method is based on a data structure, called an unbalanced "double heap" which naturally supports partition of the window samples into two different "heaps", greater and smaller than the rth sample. The algorithm is designed for any window size and any weight of the center point and needs linear memory dependent only on the window width. There are no restrictions here with respect to the window size, which may be any odd or even number.