Chihiro Matsui 研究室

主宰者：Chihiro Matsui

東京大学

AI 要約（直近 5 年の研究成果）

本研究室は、人工知能の推論処理を効率化するためのハードウェア技術と、半導体製造プロセスの最適化に関する研究を展開しています。特に、抵抗変化型メモリ（ReRAM）を用いたメモリ内演算（Computation-in-Memory）に焦点を当て、ニューラルネットワークの演算を省エネルギーで実行する仕組みを開発しています。ReRAMは高密度で低消費電力という利点がある一方で、デバイスの特性ばらつきやノイズによる信号検出の誤りが課題であり、これらを補償・耐性化する手法の研究を進めています。具体的には、複数の異なるメモリタイプを組み合わせたヘテロジニアスなシステム設計、量子化やビット圧縮を組み込んだニューラルネットワークの訓練法、デバイスレベルのノイズ特性の詳細な分析と物理モデル化、さらには長期的なデータ保持性の確保に向けた補償手法を開発しています。これらの技術により、ロボットや自動運転車といったエッジデバイスに搭載可能な、低消費電力で高精度な人工知能システムの実現を目指しています。併行して、半導体製造における単一ウェハットの湿式エッチングプロセスの自動化に取り組んでいます。機械学習モデルを用いてエッチング結果を予測し、製造レシピを自動決定する手法を開発することで、スマート製造の実現を進めています。

※ AI（Claude）が、公開されている論文要旨から研究の問い・手法・主要な発見を事実情報として抽出・再構成して自動生成しています。誤りを含む可能性があるため、正確性は研究室公式情報でご確認ください。

外部リンク

研究成果（78 件）

[2026] Comparative analysis on Gaussian process regression and numerical model for accurate Si 3 N 4 wet etching rate prediction under small training data condition
DOI: https://doi.org/10.35848/1347-4065/ae4df6
[2026] Mapping strategy of quantized KV cache of LLMs into mixture of SLC and MLC ReRAM among various quantization methods
DOI: https://doi.org/10.35848/1347-4065/ae33cb
[2025] Accurate Si3N4 Wet Etching Rate Prediction with Small Training Data by Gaussian Process Regression
DOI: https://doi.org/10.7567/ssdm.2025.a-3-05
[2025] Truncated Carry Approximate Adder and Explicit Bit Pruning for SRAM-based Digital Computation-in-Memory
DOI: https://doi.org/10.7567/ssdm.2025.f-4-03
[2025] Hybrid-Domain SRAM Computation-in-Memory Architecture for Mixed-Precision FP/INT Operations
DOI: https://doi.org/10.7567/ssdm.2025.f-4-02
[2025] Column-wise Weight Value Inversion (Colwin)of ReRAM CiM to Reduce MAC Error during Data-retention
DOI: https://doi.org/10.7567/ssdm.2025.b-2-04
[2025] Fine-Grained Extraction Method of Discrete Shift and Spike RTN in 40nm TaOX-based Analog ReRAM for CiM Applications
DOI: https://doi.org/10.7567/ssdm.2025.b-2-05
[2025] ReRAM Resistance Design of LRS and HRS for Ultrahigh-Capacity Digital Memory and Analog Computation-in-Memory
DOI: https://doi.org/10.1587/transfun.2025vlp0011
[2025] Adaptive Oxygen Vacancy Diffusion Compensation in MLC Intermediate States for Over 10-Year Data-Retention of TaO X ReRAM Analog CiM Array
DOI: https://doi.org/10.1109/esserc66193.2025.11214081
[2025] Quantization methodology for edge Transformer Hetero-CiM with SRAM CiM in self-attention and MLC ReRAM CiM in linear and FC layers
DOI: https://doi.org/10.35848/1347-4065/ade2c8

続きを表示（残り 68 件）

[2025] VaLI: Variability-aware Fine-tuning with Low-rank Adapter and Iterative Training for ReRAM Computation-in-Memory
DOI: https://doi.org/10.1109/iscas56072.2025.11044312
[2025] Signed Approximate Adder Tree and Quantization- and Bit-Pruning-Aware Training for Digital Computation-in-Memory
DOI: https://doi.org/10.1587/transele.2024fup0001
[2025] Conductance Variation-Assisted Adversarial Attack Robustness on 40nm TaOx-based ReRAM CiM
DOI: https://doi.org/10.1109/irps48204.2025.10982818
[2025] Read Voltage Dependency of Random Telegraph Noise in the Intermediate State of TaOx-based ReRAM
DOI: https://doi.org/10.1109/irps48204.2025.10983826
[2025] Invertible Prediction Model for Si3N4 Wet Etching Using DHF
DOI: https://doi.org/10.1109/edtm61175.2025.11041419
[2025] Investigation of Effects of Non-Volatile Memory-Based Computation-in-Memory Non-Idealities and Model Size on Performance and Robustness of Small Language Model During Inference Phase
DOI: https://doi.org/10.1109/edtm61175.2025.11040348
[2025] Quantization tolerant network design and performance estimation of computation-in-memory for energy-efficient 3D object detection inference
DOI: https://doi.org/10.35848/1347-4065/ada9f5
[2025] Transformer Hetero-CiM: Heterogeneous Integration of ReRAM CiM and SRAM CiM for Vision Transformer at Edge Devices
DOI: https://doi.org/10.1145/3658617.3698494
[2025] Anatomical Characterization of the Motor Branch to the Fourth Lumbrical: A Cadaver Study
DOI: https://doi.org/10.1016/j.jhsa.2024.11.023
[2024] Assessment of inference accuracy and memory capacity of computation-in-memory enabled neural network due to quantized weights, gradients, input and output signals, and memory non-idealities
DOI: https://doi.org/10.35848/1347-4065/ad2e45
[2024] Co-design of SLC/MLC FeFET-based highly error-tolerant low-power CiM and strong lottery ticket hypothesis-based algorithm
DOI: https://doi.org/10.35848/1347-4065/ad2656
[2024] Stochastic Resonance Modeling of Floating Gate-Based Neurons in Summing Networks for Accurate and Energy-Efficient Operations
DOI: https://doi.org/10.1109/ted.2024.3354231
[2024] Boundary dissipative spin chains with partial solvability inherited from system Hamiltonians
DOI: https://doi.org/10.1103/physrevb.110.224306
[2024] Device and System Co-Design of Summing Network With Floating Gate-Based Stochastic Neurons
DOI: https://doi.org/10.1109/jeds.2024.3517655
[2024] Domain-Specific CiM Design for 3D Object Detection Network with Augmented Point Cloud, Multi-Stage Quantization Aware Training, and U-Quantization
DOI: https://doi.org/10.7567/ssdm.2024.b-5-03
[2024] Automated Recipe Creation & Verification of Single Wafer Wet Etching: Ensemble Learning with Backcasting & Forecasting AIs using Scarce Data
DOI: https://doi.org/10.1109/snw63608.2024.10639220
[2024] Cytokines released from human adipose tissue-derived stem cells by bFGF stimulation: Effects of IL-8 and CXCL-1 on wound healing
DOI: https://doi.org/10.1016/j.reth.2024.06.008
[2024] Embedded Transformer Hetero-CiM: SRAM CiM for 4b Read/Write-MAC Self-attention and MLC ReRAM CiM for 6b Read-MAC Linear&FC Layers
DOI: https://doi.org/10.1109/imw59701.2024.10536980
[2024] 3D Parallel ReRAM Computation-in-Memory for Hyperdimensional Computing
DOI: https://doi.org/10.1587/transele.2023cts0001
[2024] Comprehensive Analysis of Read Fluctuations in ReRAM CiM by Using Fluctuation Pattern Classifier
DOI: https://doi.org/10.1587/transele.2023ctp0002
[2024] REM-CiM: Attentional RGB-Event Fusion Multi-Modal Analog CiM for Area/Energy-Efficient Edge Object Detection during Both Day and Night
DOI: https://doi.org/10.1587/transele.2023ctp0001
[2024] Exact steady states of the impurity-doped XXZ spin chain coupled to dissipators
DOI: https://doi.org/10.1088/1742-5468/ad2b5c
[2024] Revealing Unique Scaling Effects of Random Telegraph Noise and Electron Injection Stochasticity in Stochastic Resonance with Floating Gate-based Neurons
DOI: https://doi.org/10.1109/edtm58488.2024.10511365
[2024] Dual integration design of approximate random weight generator and computation-in-memory for event-based neuromorphic computing
DOI: https://doi.org/10.35848/1347-4065/ad2415
[2024] Design methodology of compact edge vision transformer CiM considering non-volatile memory bit precision and memory error tolerance
DOI: https://doi.org/10.35848/1347-4065/ad1bbd
[2023] NN Algorithm Aware Alternate Layer Retraining on Computation-in-Memory for Write Variation Compensation of Non-volatile Memories at Edge AI
DOI: https://doi.org/10.1109/edtm55494.2023.10103064
[2023] ReRAM CiM Fluctuation Pattern Classification by CNN Trained on Artificially Created Dataset
DOI: https://doi.org/10.1109/irps48203.2023.10118305
[2023] Stochastic Resonance Effects of Floating Gate Technology-based Leaky Integrate-and-Fire (FG LIF) Neurons in Summing Network
DOI: https://doi.org/10.1109/edtm55494.2023.10103007
[2023] FeFET Local Multiply and Global Accumulate Voltage-Sensing Computation-In-Memory Circuit Design for Neuromorphic Computing
DOI: https://doi.org/10.1109/tvlsi.2023.3336379
[2023] LIORAT: NN Layer I/O Range Training for Area/Energy-Efficient Low-Bit A/D Conversion System Design in Error-Tolerant Computation-in-Memory
DOI: https://doi.org/10.1109/iccad57390.2023.10323664
[2023] Compact Edge Vision Transformer with 86% Non-volatile Memory Bit Reduction by Percentile Clipping, Per Layer Quantization, and Quantization Aware Training
DOI: https://doi.org/10.7567/ssdm.2023.k-8-03
[2023] Highly Error-tolerant Low Power CiM by Co-design of SLC Mask & MLC Weight Twin FeFETs and Strong Lottery Ticket Hypothesis Neural Network Algorithm
DOI: https://doi.org/10.7567/ssdm.2023.k-6-02
[2023] Dual Integration of Approximate Random Weight Generator & CiM for Event-based Reservoir Computing & Spiking Neural Networks
DOI: https://doi.org/10.7567/ssdm.2023.k-6-01
[2023] Analysis of Inference Accuracy of Convolutional Neural Networks due to Quantization of Weights, Gradients, Input and Output Signals Stemming from Computation-in-Memory
DOI: https://doi.org/10.7567/ssdm.2023.ps-2-01
[2023] Ferroelectric source follower for voltage-sensing nonvolatile memory and computing-in-memory
DOI: https://doi.org/10.1088/1361-6463/acee09
[2023] Machine Learning-based Accurate Single Wafer Wet Etching Amount Prediction
DOI: https://doi.org/10.23919/snw57900.2023.10183950
[2023] Stochastic Computing-based Computation-in-Memory (SC CiM) Architecture for DNNs and Hierarchical Evaluations of Non-volatile Memory Error and Defect Tolerance
DOI: https://doi.org/10.1109/imw56887.2023.10145982
[2023] 151. Decoding the Paracrine Effect of bFGF-Stimulated Adipose Tissue-derived Stem Cells on Angiogenesis and Lymphangiogenesis
DOI: https://doi.org/10.1097/01.gox.0000938188.72436.c3
[2023] TaOx ReRAM Cell-level Unidirectional Neural Network Weight Control for Non-linearity & Variation Robust Transfer Learning of Low Cost Digital eCiM
DOI: https://doi.org/10.1109/edtm55494.2023.10103092
[2023] Comprehensive analysis on error-robustness of FeFET computation-in-memory for hyperdimensional computing
DOI: https://doi.org/10.35848/1347-4065/acb1b8
[2022] A Stochastic Leaky-Integrate-and-Fire Neuron Model With Floating Gate-Based Technology for Fast and Accurate Population Coding
DOI: https://doi.org/10.1109/jeds.2022.3206317
[2022] Tiny and Error Torrent Convolutional LSTM for Event-based Vision Sensor with ReRAM Computation-in-Memory
DOI: https://doi.org/10.7567/ssdm.2022.f-7-02
[2022] Read-disturb & Data-retention Error-robust and Fast Training Methods of FeFET LM-GA CiM for Hyperdimensional Computing
DOI: https://doi.org/10.7567/ssdm.2022.f-3-06
[2022] Heterogeneous Integration of Precise and Approximate Storage for Error-Tolerant Workloads
DOI: https://doi.org/10.1587/transfun.2022vlp0001
[2022] Compact CiM Co-optimized by Heterogeneous Multi-level ReRAM & Random Weight SNN for Event-based Vision Sensor of Edge AI
DOI: https://doi.org/10.1109/snw56633.2022.9889065
[2022] QS28. Real-time Assessment of Capillary Circulation of Free Flap Using Video-capillaroscopy: A Technological Renaissance?
DOI: https://doi.org/10.1097/01.gox.0000842964.04781.e7
[2022] Versatile FeFET Voltage-sensing Analog CiM for Fast & Small-area Hyperdimensional Computing
DOI: https://doi.org/10.1109/iscas48785.2022.9937237
[2022] Domain Specific ReRAM Computation-in-Memory Design Considering Bit Precision and Memory Errors for Simulated Annealing
DOI: https://doi.org/10.1109/iscas48785.2022.9937947
[2022] Edge Computation-in-Memory for In-situ Class-incremental Learning with Knowledge Distillation
DOI: https://doi.org/10.1109/iscas48785.2022.9937792
[2022] Event data-based computation-in-memory (CiM) configuration by co-designing integrated in-sensor and CiM computing for extremely energy-efficient edge computing
DOI: https://doi.org/10.35848/1347-4065/ac5533
[2022] Inter Spike Interval and Stochasticity Engineering of Floating Gate Technology-based Neurons for Spiking Neural Network Hardware
DOI: https://doi.org/10.1109/edtm53872.2022.9798349
[2022] Computation-in-memory simulation platform to investigate inference accuracy degradation by device non-ideality interactions in deep neural network applications
DOI: https://doi.org/10.35848/1347-4065/ac44d0
[2022] Tiny and error-tolerant ConvLSTM for event-based vision sensor with optimized event representation and ReRAM computation-in-memory
DOI: https://doi.org/10.35848/1347-4065/acad88
[2022] 97.6% array area reduction, ReRAM computation-in-memory with hyperparameter optimization based on memory bit-error rate and bit precision of log-encoding simulated annealing
DOI: https://doi.org/10.35848/1347-4065/ac356f
[2022] Intraoperative Identification of Perforator Vessel Spasm and Decreased Pulsation During Free Flap Harvest Using Video-capillaroscopy
DOI: https://doi.org/10.1097/01.gox.0000898576.94777.76
[2021] Non-volatile memory system design of edge server and cloud centralized server for multiple-tier 5G network
DOI: https://doi.org/10.35848/1347-4065/abe7fc
[2021] BER Evaluation System Considering Device Characteristics of TLC and QLC NAND Flash Memories in Hybrid SSDs with Real Storage Workloads
DOI: https://doi.org/10.1109/iscas51556.2021.9401203
[2021] Cytokines Released from Human Adipose-tissue–Derived Stem Cells By b-FGF Stimulation: Effects on Angiogenesis and Lymphatic Vessels Formation By IL-8 and CXCL1
DOI: https://doi.org/10.1097/01.gox.0000799740.60269.17
[2021] Comprehensive Computation-in-Memory Simulation Platform with Non-volatile Memory Non-Ideality Consideration for Deep Learning Applications
DOI: https://doi.org/10.7567/ssdm.2021.b-5-03
[2021] Small Array Area, Memory Error Tolerant ReRAM Computation-in-Memory with Log-encoding Simulated Annealing for Combinational Optimization Problems
DOI: https://doi.org/10.7567/ssdm.2021.b-5-02
[2021] Event-Driven SRAM Computation-in-Memory with Partitioned WL Activation for 3D Heterogeneous Integration of Event-based Vision Sensor
DOI: https://doi.org/10.7567/ssdm.2021.l-1-04
[2021] Voltage-sensing FeFET CiM with MAC by Source-follower Read and Charge-sharing
[2021] Evaluation of ReRAM errors in Computation-in-Memory for simulated annealing applications
[2021] [Highlight]FeFET Computation-in-Memory (CiM) with MAC Operation by Source-follower Read and Charge-sharing
[2021] Influence of quantized bit precision and bit-error rate in Computation-in-Memory with ReRAM on optimal answers of combinatorial optimization problems
[2021] Energy-Efficient Reliable HZO FeFET Computation-in-Memory with Local Multiply & Global Accumulate Array for Source-Follower & Charge-Sharing Voltage Sensing
DOI: https://doi.org/10.23919/vlsicircuits52068.2021.9492448
[2021] Energy-Efficient Reliable HZO FeFET Computation-in-Memory with Local Multiply & Global Accumulate Array for Source-Follower & Charge-Sharing Voltage Sensing
[2021] Simulated Annealing Algorithm & ReRAM Device Co-optimization for Computation-in-Memory
DOI: https://doi.org/10.1109/imw51353.2021.9439610

科研費（0 件）

まだデータがありません（KAKEN 取り込み後に表示）。

所属学会・役職（0 件）

まだデータがありません（学会データ連携後に表示）。

AI 要約（直近 5 年の研究成果）

外部リンク

関連研究室(8 件)

研究成果（78 件）

科研費（0 件）

所属学会・役職（0 件）