1,721,176 research outputs found
Learning Discrete-Time Markov Chains Under Concept Drift
Full text linkLearning under concept drift is a novel and promising research area aiming at designing learning algorithms able to deal with nonstationary data-generating processes. In this research field, most of the literature focuses on learning nonstationary probabilistic frameworks, while some extensions about learning graphs and signals under concept drift exist. For the first time in the literature, this paper addresses the problem of learning discrete-time Markov chains (DTMCs) under concept drift. More specifically, following a hybrid active/passive approach, this paper introduces both a family of change-detection mechanisms (CDMs), differing in the required assumptions and performance, for detecting changes in DTMCs and an adaptive learning algorithm able to deal with DTMCs under concept drift. The effectiveness of both the proposed CDMs and the adaptive learning algorithm has been extensively tested on synthetically generated experiments and real data sets
Reducing the Computation Load of Convolutional Neural Networks through Gate Classification
No full textReducing the computational load of Convolutional Neural Networks (CNNs) is of utmost importance to allow their execution in computing systems characterized by constraints on computation and energy (e.g., embedded and cyber-physical systems and Internet-of-Things). To address this problem, which has been rarely addressed in the related literature, this paper introduces the Gate-Classification CNNs. The core of this novel family of CNNs is the presence of Gate-Classification layers that allow to incrementally process the input image through the CNN layers and take a decision as soon as 'enough confidence' about the classification is gained, hence not requiring the processing of the whole CNN when not needed. The Gate-Classification CNNs rely on the ability of CNNs to process features characterized by increasing complexity and meaning and, in particular, the Gate-Classification layers allow to select the path within the CNN according to the information content provided by the input image and the processed features. A wide experimental campaign on public-available datasets supports the effectiveness of the proposed solution
Privacy-preserving time series prediction with temporal convolutional neural networks
No full textDesigning and developing machine and deep learning solutions able to guarantee the privacy of users' data is a novel and promising research area. Homomorphic Encryption (HE) is playing a primary role in this area thanks to its ability to support the processing of machine and deep learning solutions directly on encrypted data. Currently, the research in this field focuses on HE-based machine and deep learning solutions for the processing of images and text, while the privacy-preserving processing of time series has been mostly left unattended due to the strong constraints imposed by HE on the machine and deep learning forecasting models. This paper introduces, for the first time in the literature, a general privacy-preserving solution for time series prediction based on HE and Temporal Convolutional Neural Networks. The novel content brought by the paper is twofold. From the algorithmic point of view, this paper introduces a family of Temporal Convolutional Neural Networks, called PINPOINT, which is integrated with a HE scheme to support the privacy-preserving time series prediction. From the technical point of view, this paper introduces and details a Cloud-based privacy-preserving system for the forecasting-as-a-service based on the proposed PINPOINT models. Experimental results on publicly available benchmarks show the effectiveness of the proposed solution for privacy-preserving time series prediction
T4C: A Framework for Time-Series Clustering-as-a-Service
Full text linkTime-series clustering-as-a-service is an innovative and promising research area. Its main goal is to design Cloud-based platforms and services able to provide efficient and effective time-series clustering directly to final users. This paper introduces T4C, an open-source Python-based framework for time-series clustering-as-a-service. T4C integrates some of the most used time-series clustering models and techniques, and it is able to generate on-the-fly websites where users can explore the result of the clustering procedure on their previously uploaded time-series
Incremental On-Device Tiny Machine Learning
No full textTiny Machine Learning (TML) is a novel research area aiming at designing and developing Machine Learning (ML) techniques meant to be executed on Embedded Systems and Internet-of-Things (IoT) units. Such techniques, which take into account the constraints on computation, memory, and energy characterizing the hardware platform they operate on, exploit approximation and pruning mechanisms to reduce the computational load and the memory demand of Machine and Deep Learning (DL) algorithms. Despite the advancement of the research, TML solutions present in the literature assume that Embedded Systems and IoT units support only the inference of ML and DL algorithms, whereas their training is confined to more-powerful computing units (due to larger computational load and memory demand). This also prevents such pervasive devices from being able to learn in an incremental way directly from the field to improve the accuracy over time or to adapt to new working conditions. The aim of this paper is to address such an open challenge by introducing an incremental algorithm based on transfer learning and k-nearest neighbor to support the on-device learning (and not only the inference) of ML and DL solutions on embedded systems and IoT units. Moreover, the proposed solution is general and can be applied to different application scenarios. Experimental results on image/audio benchmarks and two off-The-shelf hardware platforms show the feasibility and effectiveness of the proposed solution
Learning Convolutional Neural Networks in presence of Concept Drift
No full textDesigning adaptive machine learning systems able to operate in nonstationary conditions, also called concept drift, is a novel and promising research area. Convolutional Neural Networks (CNNs) have not been considered a viable solution for such adaptive systems due to the high computational load and the high number of images they require for the training. This paper introduces an adaptive mechanism for learning CNNs able to operate in presence of concept drift. Such an adaptive mechanism follows an "active approach", where the adaptation is triggered by the detection of a concept drift, and relies on the "transfer learning" paradigm to transfer (part of the) knowledge from the CNN operating before the concept drift to the one operating after. The effectiveness of the proposed solution has been evaluated on two types of CNNs and two real-world image benchmarks
New facies interpretation of the Messinian evaporites in the Mediterranean
No full textIntroductionThe study of the Messinian salinity crisis (MSC) in the Mediterranean has generated a controversy with a long-term discussion on many aspects. One of the problem faced by the scientists is that the interpretation of evaporite sediments can be very complicated. Evaporite sediments are among the most elusive for facies reconstruction and correlation and most data in the literature were not correctly placed in a reliable stratigraphic framework. Some examples of these difficulties are the following: a) deposits traditionally included in the Lower Evaporites are actually clastic sediments that, as we have shown, derive from the dismantlement of autochthonous Lower Evaporites (Manzi et al., 2005; Roveri et al., 2006);b) many of the Lower Evaporites outcrops in Italy are actually large-scale blocks emplaced by extensive mass-waste movements (Roveri et al., 2003; Roveri et al., 2006); some of these chaotic complexes were interpreted as collapse deposits due to halite dissolution (Caruso and Rouchy, 2006);c) some Lower Evaporites outcrops were commonly considered Upper Evaporites and vice-versa;d) laminated clastic sulphate sediments were commonly mistaken for primary cumulate deposits and vice-versa;e) the significance of halite deposition, which actually bears the only unequivocal sign of exposure found within the Messinian evaporites, has been overlooked;f) the Calcare di Base carbonates show commonly evidence of resedimentation;g) the Calcare di Base is never found at the base of the Lower Evaporites primary in situ selenites;h) lateral transitions between carbonate, gypsum or halite primary evaporites cannot be directly observed and must be considered speculation;For these reasons we devoted our efforts to provide a detailed stratigraphic and facies analyses of all the Messinian evaporites and criteria for distinguishing the Lower from the Upper Evaporites. The main aim was to discuss new possible stratigraphic markers to correlate the elusive evaporite sediments across the Mediterranean during the MSC and to correctly place the Messinian units into a reliable stratigraphic framework.Finally, as the peculiar and restricted setting of evaporite deposition makes the use of geochemical data problematic, we need to integrate the geochemical data in a detailed stratigraphic and facies framework. This because many of the available isotope data are scattered and commonly obtained from sections whose stratigraphy is not well constrained in the regional framework. An extensive study of the literature showed that the majority of the geochemical data were provided without a reliable record even for the local stratigraphy.This presentation illustrates our studies on a new evaporite facies interpretation that may be useful for large-scale correlations. The effort is to provide a new reliable facies, isotope and stratigraphic framework for the understanding of the Salinity Crisis in the Mediterranean
Artificial ageing of photocatalytic nanocomposites for the protection of natural stones
Full text linkDuring the last ten years, photocatalytic nanocomposites combining titania nanoparticles with silicon-based matrices have received increasing attention in the stone conservation research field, because they oer an eective multifunctional approach to the issue of stone protection. However, much work still has to be done in studying the behaviour of these nanocomposites in real environmental conditions and understanding to what extent they are able to retain their eectiveness and compatibility once applied on outdoor surfaces. The latter is a key information that should lie at the basis of any successful conservation and maintenance campaign. The present study provides insight into this relevant topic trough laboratory testing by assessing the artificial ageing of two silane-based photocatalytic nanocomposites, previously selected through an accurate testing on dierent natural stones. Three accelerated ageing procedures, based on artificial solar irradiation, heating and rain wash-out, allowed simulating about two years of outdoor exposure to some of the weathering factors to which stones are normally subjected. The results provided quite accurate information about the long-term behaviour of the products and on the role that the stone properties play therein. It was shown that, when the products are able to penetrate deeply enough inside the stone pores, they retain much of their hydrophobising and photocatalytic properties and maintain a good compatibility with the stone substrates, even after partial chemical degradation of the alkyl-silica matrices has occurred on the very stone surface
- …
