An energy-efficient in-memory computing architecture for survival data analysis based on resistive switching memories
One of the objectives fostered in medical science is the so-called precision medicine, which requires the analysis of a large amount of survival data from patients to deeply understand treatment options. Tools like machine learning (ML) and deep neural networks are becoming a de-facto standard. Nowa...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2022-08-01
|
| Series: | Frontiers in Neuroscience |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fnins.2022.932270/full |
| _version_ | 1818504681017573376 |
|---|---|
| author | Andrea Baroni Artem Glukhov Eduardo Pérez Christian Wenger Christian Wenger Enrico Calore Enrico Calore Sebastiano Fabio Schifano Sebastiano Fabio Schifano Piero Olivo Daniele Ielmini Cristian Zambelli |
| author_facet | Andrea Baroni Artem Glukhov Eduardo Pérez Christian Wenger Christian Wenger Enrico Calore Enrico Calore Sebastiano Fabio Schifano Sebastiano Fabio Schifano Piero Olivo Daniele Ielmini Cristian Zambelli |
| author_sort | Andrea Baroni |
| collection | DOAJ |
| description | One of the objectives fostered in medical science is the so-called precision medicine, which requires the analysis of a large amount of survival data from patients to deeply understand treatment options. Tools like machine learning (ML) and deep neural networks are becoming a de-facto standard. Nowadays, computing facilities based on the Von Neumann architecture are devoted to these tasks, yet rapidly hitting a bottleneck in performance and energy efficiency. The in-memory computing (IMC) architecture emerged as a revolutionary approach to overcome that issue. In this work, we propose an IMC architecture based on resistive switching memory (RRAM) crossbar arrays to provide a convenient primitive for matrix-vector multiplication in a single computational step. This opens massive performance improvement in the acceleration of a neural network that is frequently used in survival analysis of biomedical records, namely the DeepSurv. We explored how the synaptic weights mapping strategy and the programming algorithms developed to counter RRAM non-idealities expose a performance/energy trade-off. Finally, we discussed how this application is tailored for the IMC architecture rather than being executed on commodity systems. |
| first_indexed | 2024-12-10T21:40:24Z |
| format | Article |
| id | doaj.art-fc722f8ddb564ece8c127eb70dfd8593 |
| institution | Directory Open Access Journal |
| issn | 1662-453X |
| language | English |
| last_indexed | 2024-12-10T21:40:24Z |
| publishDate | 2022-08-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Neuroscience |
| spelling | doaj.art-fc722f8ddb564ece8c127eb70dfd85932022-12-22T01:32:32ZengFrontiers Media S.A.Frontiers in Neuroscience1662-453X2022-08-011610.3389/fnins.2022.932270932270An energy-efficient in-memory computing architecture for survival data analysis based on resistive switching memoriesAndrea Baroni0Artem Glukhov1Eduardo Pérez2Christian Wenger3Christian Wenger4Enrico Calore5Enrico Calore6Sebastiano Fabio Schifano7Sebastiano Fabio Schifano8Piero Olivo9Daniele Ielmini10Cristian Zambelli11IHP-Leibniz Institut fur Innovative Mikroelektronik, Frankfurt (Oder), GermanyDipartimento di Elettronica, Informazione e Bioingegneria, Politecnico di Milano and IU.NET, Milano, ItalyIHP-Leibniz Institut fur Innovative Mikroelektronik, Frankfurt (Oder), GermanyIHP-Leibniz Institut fur Innovative Mikroelektronik, Frankfurt (Oder), GermanyBTU Cottbus-Senftenberg, Cottbus, GermanyDipartimento di Fisica e Scienze Della Terra, Università Degli Studi di Ferrara, Ferrara, ItalyIstituto Nazionale di Fisica Nucleare (INFN), Ferrara, ItalyIstituto Nazionale di Fisica Nucleare (INFN), Ferrara, ItalyDipartimento di Scienze Dell'Ambiente e Della Prevenzione, Università Degli Studi di Ferrara, Ferrara, ItalyDipartimento di Ingegneria, Università Degli Studi di Ferrara, Ferrara, ItalyDipartimento di Elettronica, Informazione e Bioingegneria, Politecnico di Milano and IU.NET, Milano, ItalyDipartimento di Ingegneria, Università Degli Studi di Ferrara, Ferrara, ItalyOne of the objectives fostered in medical science is the so-called precision medicine, which requires the analysis of a large amount of survival data from patients to deeply understand treatment options. Tools like machine learning (ML) and deep neural networks are becoming a de-facto standard. Nowadays, computing facilities based on the Von Neumann architecture are devoted to these tasks, yet rapidly hitting a bottleneck in performance and energy efficiency. The in-memory computing (IMC) architecture emerged as a revolutionary approach to overcome that issue. In this work, we propose an IMC architecture based on resistive switching memory (RRAM) crossbar arrays to provide a convenient primitive for matrix-vector multiplication in a single computational step. This opens massive performance improvement in the acceleration of a neural network that is frequently used in survival analysis of biomedical records, namely the DeepSurv. We explored how the synaptic weights mapping strategy and the programming algorithms developed to counter RRAM non-idealities expose a performance/energy trade-off. Finally, we discussed how this application is tailored for the IMC architecture rather than being executed on commodity systems.https://www.frontiersin.org/articles/10.3389/fnins.2022.932270/fullresistive RAM (RRAM)driftin-memory computing (IMC)survival analysismulti level conductance |
| spellingShingle | Andrea Baroni Artem Glukhov Eduardo Pérez Christian Wenger Christian Wenger Enrico Calore Enrico Calore Sebastiano Fabio Schifano Sebastiano Fabio Schifano Piero Olivo Daniele Ielmini Cristian Zambelli An energy-efficient in-memory computing architecture for survival data analysis based on resistive switching memories Frontiers in Neuroscience resistive RAM (RRAM) drift in-memory computing (IMC) survival analysis multi level conductance |
| title | An energy-efficient in-memory computing architecture for survival data analysis based on resistive switching memories |
| title_full | An energy-efficient in-memory computing architecture for survival data analysis based on resistive switching memories |
| title_fullStr | An energy-efficient in-memory computing architecture for survival data analysis based on resistive switching memories |
| title_full_unstemmed | An energy-efficient in-memory computing architecture for survival data analysis based on resistive switching memories |
| title_short | An energy-efficient in-memory computing architecture for survival data analysis based on resistive switching memories |
| title_sort | energy efficient in memory computing architecture for survival data analysis based on resistive switching memories |
| topic | resistive RAM (RRAM) drift in-memory computing (IMC) survival analysis multi level conductance |
| url | https://www.frontiersin.org/articles/10.3389/fnins.2022.932270/full |
| work_keys_str_mv | AT andreabaroni anenergyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT artemglukhov anenergyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT eduardoperez anenergyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT christianwenger anenergyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT christianwenger anenergyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT enricocalore anenergyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT enricocalore anenergyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT sebastianofabioschifano anenergyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT sebastianofabioschifano anenergyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT pieroolivo anenergyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT danieleielmini anenergyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT cristianzambelli anenergyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT andreabaroni energyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT artemglukhov energyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT eduardoperez energyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT christianwenger energyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT christianwenger energyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT enricocalore energyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT enricocalore energyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT sebastianofabioschifano energyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT sebastianofabioschifano energyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT pieroolivo energyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT danieleielmini energyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories AT cristianzambelli energyefficientinmemorycomputingarchitectureforsurvivaldataanalysisbasedonresistiveswitchingmemories |