<?xml version="1.0" encoding="utf-8"?>
<export-example>
  <doc>
    <id>28701</id>
    <completedYear/>
    <publishedYear>2022</publishedYear>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst>4</pageFirst>
    <pageLast>13</pageLast>
    <pageNumber/>
    <edition/>
    <issue>2</issue>
    <volume>16</volume>
    <type>articler</type>
    <publisherName/>
    <publisherPlace/>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>0</belongsToBibliography>
    <completedDate>2022-03-22</completedDate>
    <publishedDate>--</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">In-Memory Principal Component Analysis by Crosspoint Array of Rresistive Switching Memory</title>
    <abstract language="eng">In Memory Computing (IMC) is one of the most promising candidates&#13;
for data-intensive computing accelerators of machine learning (ML). A key ML algorithm for dimensionality reduction and classification is principal component analysis (PCA), which heavily relies on matrixvector multiplications (MVM) for which classic von Neumann architectures are not optimized. Here, we provide the experimental&#13;
demonstration of a new IMCbased PCA algorithm based on power iteration and deflation executed in a 4-kbit array of resistive switching random-access memory (RRAM). The classification accuracy of the Wisconsin Breast Cancer data set reaches 95.43%, close to floatingpoint implementation. Our simulations indicate a 250× improvement in energy efficiency compared to commercial GPUs, thus supporting IMC for energy-efficient ML in modern data-intensive computing.</abstract>
    <parentTitle language="eng">IEEE Nanotechnology Magazine</parentTitle>
    <identifier type="issn">1932-4510</identifier>
    <identifier type="doi">10.1109/MNANO.2022.3141515</identifier>
    <enrichment key="BTU">an der BTU erstellt / created at BTU</enrichment>
    <enrichment key="opus.source">publish</enrichment>
    <enrichment key="opus.doi.autoCreate">false</enrichment>
    <enrichment key="opus.urn.autoCreate">false</enrichment>
    <enrichment key="Fprofil">4 Künstliche Intelligenz und Sensorik / Artificial Intelligence and Sensor Technology</enrichment>
    <author>
      <firstName>Piergiulio</firstName>
      <lastName>Mannocci</lastName>
    </author>
    <submitter>
      <firstName>Christian</firstName>
      <lastName>Wenger</lastName>
    </submitter>
    <author>
      <firstName>Andrea</firstName>
      <lastName>Baroni</lastName>
    </author>
    <author>
      <firstName>Enrico</firstName>
      <lastName>Melacarne</lastName>
    </author>
    <author>
      <firstName>Cristian</firstName>
      <lastName>Zambelli</lastName>
    </author>
    <author>
      <firstName>Piero</firstName>
      <lastName>Olivo</lastName>
    </author>
    <author>
      <firstName>Eduardo</firstName>
      <lastName>Pérez</lastName>
    </author>
    <author>
      <firstName>Christian</firstName>
      <lastName>Wenger</lastName>
    </author>
    <author>
      <firstName>Daniele</firstName>
      <lastName>Ielmini</lastName>
    </author>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>RRAM</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Multilevel switching</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>neural network</value>
    </subject>
    <collection role="institutes" number="1521">FG Halbleitermaterialien</collection>
  </doc>
  <doc>
    <id>28913</id>
    <completedYear/>
    <publishedYear>2022</publishedYear>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst>401</pageFirst>
    <pageLast>417</pageLast>
    <pageNumber/>
    <edition/>
    <issue/>
    <volume/>
    <type>conferenceobject_ref</type>
    <publisherName>Springer</publisherName>
    <publisherPlace/>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>0</belongsToBibliography>
    <completedDate>2022-05-27</completedDate>
    <publishedDate>--</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">Mitigating the Effects of RRAM Process Variation on the Accuracy of Artifical Neural Networks</title>
    <abstract language="eng">Weight storage is a key challenge in the efficient implementation of artificial neural networks. Novel memory technologies such as RRAM are able to greatly improve density and introduce non-volatility and multibit capabilities to this component of ANN accelerators. The usage of RRAM in this domain comes with downsides, mainly caused by cycle-to-cycle and device-to-device variability leading to erroneous readouts, greatly affecting digital systems. ANNs have the ability to compensate for this by their inherent redundancy and usually exhibit a gradual deterioration in the accuracy of the task at hand. This means, that slight error rates can be acceptable for weight storage in an ANN accelerator. In this work we link device-to-device variability to the accuracy of an ANN for such an accelerator. From this study, we can estimate how strongly a certain net is affected by a certain device parameter variability. This methodology is then used to present three mitigation strategies and to evaluate how they affect the reaction of the network to variability: a) Dropout Layers b) Fault-Aware Training c) Redundancy. These mitigations are then evaluated by their ability to improve accuracy and to lower hardware overhead by providing data for a real-word example. We improved this network’s resilience in such a way that it could tolerate double the variation in one of the device parameters (standard deviation of the oxide thickness can be 0.4 nm instead of 0.2 nm while maintaining sufficient accuracy.)</abstract>
    <parentTitle language="eng">Embedded Computer Systems: Architectures, Modeling, and Simulation. SAMOS 2021. Lecture Notes in Computer Science</parentTitle>
    <identifier type="isbn">978-3-031-04579-0</identifier>
    <identifier type="doi">10.1007/978-3-031-04580-6_27</identifier>
    <identifier type="isbn">978-3-031-04580-6</identifier>
    <identifier type="issn">0302-9743</identifier>
    <identifier type="issn">1611-3349</identifier>
    <enrichment key="BTU">an der BTU erstellt / created at BTU</enrichment>
    <enrichment key="opus.source">publish</enrichment>
    <enrichment key="UBICOseries">Lecture Notes in Computer Science ; volume 13227</enrichment>
    <enrichment key="opus.doi.autoCreate">false</enrichment>
    <enrichment key="opus.urn.autoCreate">false</enrichment>
    <enrichment key="Fprofil">4 Künstliche Intelligenz und Sensorik / Artificial Intelligence and Sensor Technology</enrichment>
    <author>
      <firstName>Markus</firstName>
      <lastName>Fritscher</lastName>
    </author>
    <submitter>
      <firstName>Christian</firstName>
      <lastName>Wenger</lastName>
    </submitter>
    <author>
      <firstName>Johannes</firstName>
      <lastName>Knödtel</lastName>
    </author>
    <author>
      <firstName>Maen</firstName>
      <lastName>Mallah</lastName>
    </author>
    <author>
      <firstName>Stefan</firstName>
      <lastName>Pechmann</lastName>
    </author>
    <author>
      <firstName>Emilio</firstName>
      <lastName>Perez-Bosch Quesada</lastName>
    </author>
    <author>
      <firstName>Tommaso</firstName>
      <lastName>Rizzi</lastName>
    </author>
    <author>
      <firstName>Christian</firstName>
      <lastName>Wenger</lastName>
    </author>
    <author>
      <firstName>Marc</firstName>
      <lastName>Reichenbach</lastName>
    </author>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>RRAM</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>memristive device</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>neural network</value>
    </subject>
    <collection role="institutes" number="1521">FG Halbleitermaterialien</collection>
  </doc>
  <doc>
    <id>28821</id>
    <completedYear/>
    <publishedYear>2022</publishedYear>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst>3C.3-1</pageFirst>
    <pageLast>3C.3-7</pageLast>
    <pageNumber/>
    <edition/>
    <issue/>
    <volume/>
    <type>conferenceobject_ref</type>
    <publisherName>Institute of Electrical and Electronics Engineers (IEEE)</publisherName>
    <publisherPlace/>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>0</belongsToBibliography>
    <completedDate>2022-05-09</completedDate>
    <publishedDate>--</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">Statistical model of program/verify algorithms in resistive-switching memories for in-memory neural network accelerators</title>
    <abstract language="eng">Resistive-switching random access memory (RRAM) is a promising technology for in-memory computing (IMC) to accelerate training and inference of deep neural networks (DNNs). This work presents the first physics-based statistical model describing (i) multilevel RRAM device program/verify (PV) algorithms by controlled set transition, (ii) the stochastic cycle-to-cycle (C2C) and device-to-device (D2D) variations within the array, and (iii) the impact of such imprecisions on the accuracy of DNN accelerators. The model can handle the full chain from RRAM materials/device parameters to the DNN performance, thus providing a valuable tool for device/circuit codesign of hardware DNN accelerators.</abstract>
    <parentTitle language="eng">2022 IEEE International Reliability Physics Symposium (IRPS)</parentTitle>
    <identifier type="issn">2473-2001</identifier>
    <identifier type="doi">10.1109/IRPS48227.2022.9764497</identifier>
    <identifier type="isbn">978-1-6654-7950-9</identifier>
    <identifier type="isbn">978-1-6654-7951-6</identifier>
    <enrichment key="BTU">an der BTU erstellt / created at BTU</enrichment>
    <enrichment key="opus.source">publish</enrichment>
    <enrichment key="opus.doi.autoCreate">false</enrichment>
    <enrichment key="opus.urn.autoCreate">false</enrichment>
    <enrichment key="Fprofil">4 Künstliche Intelligenz und Sensorik / Artificial Intelligence and Sensor Technology</enrichment>
    <author>
      <firstName>Artem</firstName>
      <lastName>Glukhov</lastName>
    </author>
    <submitter>
      <firstName>Christian</firstName>
      <lastName>Wenger</lastName>
    </submitter>
    <author>
      <firstName>Valerio</firstName>
      <lastName>Milo</lastName>
    </author>
    <author>
      <firstName>Andrea</firstName>
      <lastName>Baroni</lastName>
    </author>
    <author>
      <firstName>Nicola</firstName>
      <lastName>Lepri</lastName>
    </author>
    <author>
      <firstName>Cristian</firstName>
      <lastName>Zambelli</lastName>
    </author>
    <author>
      <firstName>Piero</firstName>
      <lastName>Olivo</lastName>
    </author>
    <author>
      <firstName>Eduardo</firstName>
      <lastName>Pérez</lastName>
    </author>
    <author>
      <firstName>Christian</firstName>
      <lastName>Wenger</lastName>
    </author>
    <author>
      <firstName>Daniele</firstName>
      <lastName>Ielmini</lastName>
    </author>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>RRAM</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Multilevel switching</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>neural network</value>
    </subject>
    <collection role="institutes" number="1521">FG Halbleitermaterialien</collection>
  </doc>
  <doc>
    <id>29207</id>
    <completedYear/>
    <publishedYear>2022</publishedYear>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst>83</pageFirst>
    <pageLast>88</pageLast>
    <pageNumber/>
    <edition/>
    <issue/>
    <volume/>
    <type>conferenceobject_ref</type>
    <publisherName/>
    <publisherPlace/>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>0</belongsToBibliography>
    <completedDate>2022-08-02</completedDate>
    <publishedDate>--</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">Analytical Calculation of Inference in Memristor-based Stochastic Artificial Neural Networks</title>
    <abstract language="eng">The impact of artificial intelligence on human life&#13;
has increased significantly in recent years. However, as the&#13;
complexity of problems rose aswell, increasing system features&#13;
for such amount of data computation became troublesome due&#13;
to the von Neumann’s computer architecture. Neuromorphic&#13;
computing aims to solve this problem by mimicking the parallel&#13;
computation of a human brain. For this approach, memristive&#13;
devices are used to emulate the synapses of a human brain. Yet,&#13;
common simulations of hardware based networks require time&#13;
consuming Monte-Carlo simulations to take into account the&#13;
stochastic switching of memristive devices. This work presents&#13;
an alternative concept making use of the convolution of the&#13;
probability distribution functions (PDF) of memristor currents&#13;
by its equivalent multiplication in Fourier domain. An artificial&#13;
neural network is accordingly implemented to perform the&#13;
inference stage with handwritten digits.</abstract>
    <parentTitle language="eng">29th International Conference on Mixed Design of Integrated Circuits and System (MIXDES), 23-24 June 2022 , Wrocław, Poland</parentTitle>
    <identifier type="doi">10.23919/MIXDES55591.2022.9838321</identifier>
    <identifier type="isbn">978-83-63578-22-0</identifier>
    <identifier type="isbn">978-83-63578-21-3</identifier>
    <identifier type="isbn">978-1-6654-6176-4</identifier>
    <enrichment key="BTU">an der BTU erstellt / created at BTU</enrichment>
    <enrichment key="opus.source">publish</enrichment>
    <enrichment key="opus.doi.autoCreate">false</enrichment>
    <enrichment key="opus.urn.autoCreate">false</enrichment>
    <enrichment key="Fprofil">4 Künstliche Intelligenz und Sensorik / Artificial Intelligence and Sensor Technology</enrichment>
    <author>
      <firstName>Nicolas</firstName>
      <lastName>Bogun</lastName>
    </author>
    <submitter>
      <firstName>Christian</firstName>
      <lastName>Wenger</lastName>
    </submitter>
    <author>
      <firstName>Emilio</firstName>
      <lastName>Perez-Bosch Quesada</lastName>
    </author>
    <author>
      <firstName>Eduardo</firstName>
      <lastName>Pérez</lastName>
    </author>
    <author>
      <firstName>Christian</firstName>
      <lastName>Wenger</lastName>
    </author>
    <author>
      <firstName>Alexander</firstName>
      <lastName>Kloes</lastName>
    </author>
    <author>
      <firstName>Mike</firstName>
      <lastName>Schwarz</lastName>
    </author>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>RRAM</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>neural network</value>
    </subject>
    <collection role="institutes" number="1521">FG Halbleitermaterialien</collection>
  </doc>
  <doc>
    <id>29286</id>
    <completedYear/>
    <publishedYear>2022</publishedYear>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst>340</pageFirst>
    <pageLast>347</pageLast>
    <pageNumber/>
    <edition/>
    <issue>3</issue>
    <volume>22</volume>
    <type>articler</type>
    <publisherName/>
    <publisherPlace/>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>0</belongsToBibliography>
    <completedDate>2022-09-13</completedDate>
    <publishedDate>--</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">Low Conductance State Drift Characterization and Mitigation in Resistive Switching Memories (RRAM) for Artificial Neural Networks</title>
    <abstract language="eng">The crossbar structure of Resistive-switching random access memory (RRAM) arrays enabled the In-Memory Computing circuits paradigm, since they imply the native acceleration of a crucial operations in this scenario, namely the Matrix-Vector-Multiplication (MVM). However, RRAM arrays are affected by several issues materializing in conductance variations that might cause severe performance degradation. A critical&#13;
one is related to the drift of the low conductance states appearing&#13;
immediately at the end of program and verify algorithms that&#13;
are mandatory for an accurate multi-level conductance operation. In this work, we analyze the benefits of a new programming&#13;
algorithm that embodies Set and Reset switching operations to&#13;
achieve better conductance control and lower variability. Data&#13;
retention analysis performed with different temperatures for 168&#13;
hours evidence its superior performance with respect to standard programming approach. Finally, we explored the benefits&#13;
of using our methodology at a higher abstraction level, through&#13;
the simulation of an Artificial Neural Network for image recognition task (MNIST dataset). The accuracy achieved shows higher&#13;
performance stability over temperature and time.</abstract>
    <parentTitle language="eng">IEEE Transactions on Device and Materials Reliability</parentTitle>
    <identifier type="issn">1530-4388</identifier>
    <identifier type="doi">10.1109/TDMR.2022.3182133</identifier>
    <enrichment key="BTU">an der BTU erstellt / created at BTU</enrichment>
    <enrichment key="opus.source">publish</enrichment>
    <enrichment key="opus.doi.autoCreate">false</enrichment>
    <enrichment key="opus.urn.autoCreate">false</enrichment>
    <enrichment key="Fprofil">4 Künstliche Intelligenz und Sensorik / Artificial Intelligence and Sensor Technology</enrichment>
    <author>
      <firstName>Andrea</firstName>
      <lastName>Baroni</lastName>
    </author>
    <submitter>
      <firstName>Christian</firstName>
      <lastName>Wenger</lastName>
    </submitter>
    <author>
      <firstName>Artem</firstName>
      <lastName>Glukhov</lastName>
    </author>
    <author>
      <firstName>Eduardo</firstName>
      <lastName>Pérez</lastName>
    </author>
    <author>
      <firstName>Christian</firstName>
      <lastName>Wenger</lastName>
    </author>
    <author>
      <firstName>Daniele</firstName>
      <lastName>Ielmini</lastName>
    </author>
    <author>
      <firstName>Piero</firstName>
      <lastName>Olivo</lastName>
    </author>
    <author>
      <firstName>Cristian</firstName>
      <lastName>Zambelli</lastName>
    </author>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>RRAM</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>neural network</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Multilevel switching</value>
    </subject>
    <collection role="institutes" number="1521">FG Halbleitermaterialien</collection>
  </doc>
  <doc>
    <id>29534</id>
    <completedYear/>
    <publishedYear>2022</publishedYear>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst>1</pageFirst>
    <pageLast>5</pageLast>
    <pageNumber/>
    <edition/>
    <issue/>
    <volume/>
    <type>conferenceobject_ref</type>
    <publisherName/>
    <publisherPlace/>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>0</belongsToBibliography>
    <completedDate>2022-11-17</completedDate>
    <publishedDate>--</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">End-to-end modeling of variability-aware neural networks based on resistive-switching memory arrays</title>
    <abstract language="eng">Resistive-switching random access memory (RRAM) is a promising technology that enables advanced applications in the field of in-memory computing (IMC). By operating the memory array in the analogue domain, RRAM-based IMC architectures can dramatically improve the energy efficiency of deep neural networks (DNNs). However, achieving a high inference accuracy is challenged by significant variation of RRAM conductance levels, which can be compensated by (i) advanced programming techniques and (ii) variability-aware training (VAT) algorithms. In both cases, however, detailed knowledge and accurate physics-based statistical models of RRAM are needed to develop programming and VAT methodologies. This work presents an end-to-end approach to the development of highly-accurate IMC circuits with RRAM, encompassing the device modeling, the precise programming algorithm, and the VAT simulations to maximize the DNN classification accuracy in presence of conductance variations.</abstract>
    <parentTitle language="eng">Proc. 30th IFIP/IEEE International Conference on Very Large Scale Integration (VLSI-SoC 2022)</parentTitle>
    <identifier type="doi">10.1109/VLSI-SoC54400.2022.9939653</identifier>
    <enrichment key="BTU">an der BTU erstellt / created at BTU</enrichment>
    <enrichment key="opus.source">publish</enrichment>
    <enrichment key="opus.doi.autoCreate">false</enrichment>
    <enrichment key="opus.urn.autoCreate">false</enrichment>
    <enrichment key="Fprofil">4 Künstliche Intelligenz und Sensorik / Artificial Intelligence and Sensor Technology</enrichment>
    <author>
      <firstName>Artem</firstName>
      <lastName>Glukhov</lastName>
    </author>
    <submitter>
      <firstName>Christian</firstName>
      <lastName>Wenger</lastName>
    </submitter>
    <author>
      <firstName>Nicola</firstName>
      <lastName>Lepri</lastName>
    </author>
    <author>
      <firstName>Valerio</firstName>
      <lastName>Milo</lastName>
    </author>
    <author>
      <firstName>Andrea</firstName>
      <lastName>Baroni</lastName>
    </author>
    <author>
      <firstName>Cristian</firstName>
      <lastName>Zambelli</lastName>
    </author>
    <author>
      <firstName>Piero</firstName>
      <lastName>Olivo</lastName>
    </author>
    <author>
      <firstName>Eduardo</firstName>
      <lastName>Pérez</lastName>
    </author>
    <author>
      <firstName>Christian</firstName>
      <lastName>Wenger</lastName>
    </author>
    <author>
      <firstName>Daniele</firstName>
      <lastName>Ielmini</lastName>
    </author>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>RRAM</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>HfO2</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>neural network</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>memristive switching</value>
    </subject>
    <collection role="institutes" number="1521">FG Halbleitermaterialien</collection>
  </doc>
</export-example>
