Tohru Ishihara 研究室

主宰者：Tohru Ishihara

名古屋大学

AI 要約（直近 5 年の研究成果）

石原研究室は、エネルギー効率を重視したコンピュータシステムの設計と最適化に取り組んでいます。人工知能の推論処理やニューラルネットワークなどの計算負荷が大きい処理において、完全な精度ではなく許容範囲内の精度で動作させることで、消費電力を削減する「近似計算」という手法を活用しています。また、光を利用した高速・低消費電力の光ニューラルネットワークなども研究対象としており、様々な角度からエネルギー効率の向上を追求しています。これらの研究を実現するため、同研究室では複数のアプローチを並行して展開しています。ハードウェアレベルでは、近似乗算回路やメモリの設計最適化、電圧制御技術（DVFS）の活用などを検討しており、ソフトウェアレベルでは不要な計算の削減やビット幅の最適化などを実施しています。さらに、設計した回路やシステムが仕様を満たしているか確認するための検証技術の開発にも注力しており、実用的なエネルギー効率化設計の実現を目指しています。

※ 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（33 件）

[2026] Message from the Editor-in-Chief
DOI: https://doi.org/10.2197/ipsjtsldm.19.1
[2026] A Method of Improving Area-Efficiency for Vision Transformer Inference Accelerator Based on Accuracy-Aware Piecewise-Linear Approximation
DOI: https://doi.org/10.1109/access.2026.3692644
[2025] Utilizing Wide Range Energy/Delay Tradeoffs in Logarithmic Multipliers to Reduce Energy Dissipated for AI Inference Workloads
DOI: https://doi.org/10.1109/mcsoc67473.2025.00076
[2025] Cross-Layer Approximate Computing and Critical Path Isolation for Energy-Efficient and Variation-Tolerant Design
DOI: https://doi.org/10.1109/mcsoc67473.2025.00075
[2025] Piecewise-Linear Approximation of Self-Attention and Its Accuracy-Aware Training for Area-Efficient Vision Transformer Inference Accelerator
DOI: https://doi.org/10.1109/isqed65160.2025.11014475
[2025] Panel Discussions: “Sustainable AI: Emerging Architectures, Devices, and Quantum Computing Towards Future Computing”
DOI: https://doi.org/10.1109/coolchips65488.2025.11018567
[2025] Message from the Editor-in-Chief
DOI: https://doi.org/10.2197/ipsjtsldm.18.1
[2024] Optoelectronic Pipeline Architecture of Convolutional RNN for Energy Efficient Inference at the Speed of Light
DOI: https://doi.org/10.1587/transele.2024lhp0004
[2024] Message from the Editor-in-Chief
DOI: https://doi.org/10.2197/ipsjtsldm.17.1
[2023] A Standard Cell Memory Based on 2T Gain Cell DRAM for Memory-Centric Accelerator Design
DOI: https://doi.org/10.1109/socc58585.2023.10257009

続きを表示（残り 23 件）

[2023] Identification of Redundant Flip-Flops Using Fault Injection for Low-Power Approximate Computing Circuits
DOI: https://doi.org/10.1587/transfun.2023vlp0008
[2023] Virtualizing DVFS for Energy Minimization of Embedded Dual-OS Platform
DOI: https://doi.org/10.1587/transfun.2023kep0002
[2023] Feedback-Tuned Fuzzing for Accelerating Quality Verification of Approximate Computing Design
DOI: https://doi.org/10.1109/iolts59296.2023.10224891
[2023] An Efficient Fault Injection Algorithm for Identifying Unimportant FFs in Approximate Computing Circuits
DOI: https://doi.org/10.23919/date56975.2023.10137272
[2022] Analysis of Demagnetizing Effects on Microstrip Line Type Probe for the Permeability Measurement of Thick Specimen
DOI: https://doi.org/10.3379/msjmag.2211r002
[2022] Zero-Aware Fine-Grained Power Gating for Standard-Cell Memories in Voltage-Scaled Circuits
DOI: https://doi.org/10.1109/socc56010.2022.9908116
[2022] Yb-doped Y–Al–O thin films with a self-organized columnar structure and their anti-Stokes photoluminescence properties
DOI: https://doi.org/10.1063/5.0079632
[2022] An Accuracy Reconfigurable Vector Accelerator based on Approximate Logarithmic Multipliers for Energy-Efficient Computing
DOI: https://doi.org/10.1587/transfun.2022vlp0005
[2022] Dynamic Verification Framework of Approximate Computing Circuits using Quality-Aware Coverage-Based Grey-Box Fuzzing
DOI: https://doi.org/10.1587/transfun.2022vlp0002
[2022] DVFS Virtualization for Energy Minimization of Mixed-Criticality Dual-OS Platforms
DOI: https://doi.org/10.1109/rtcsa55878.2022.00020
[2022] Power-aware pruning for ultrafast, energy-efficient, and accurate optical neural network design
DOI: https://doi.org/10.1145/3489517.3530405
[2022] Optoelectronic Implementation of Compact and Power-efficient Recurrent Neural Networks
DOI: https://doi.org/10.1109/isvlsi54635.2022.00087
[2022] A Correction of Measurement for Complex Magnetic Permeability by Direct Contact Probe
DOI: https://doi.org/10.3379/msjmag.2207r002
[2022] An Accuracy Reconfigurable Vector Accelerator Based on Approximate Logarithmic Multipliers
DOI: https://doi.org/10.1109/asp-dac52403.2022.9712504
[2022] Energy efficient OEO conversion and its applications to photonic integrated systems
DOI: https://doi.org/10.1364/ofc.2022.th3a.4
[2021] Approximate Minimum Energy Point Tracking and Task Scheduling for Energy-Efficient Real-Time Computing
DOI: https://doi.org/10.1587/transfun.2021vlp0007
[2021] Low-Power Design Methodology of Voltage Over-Scalable Circuit with Critical Path Isolation and Bit-Width Scaling
DOI: https://doi.org/10.1587/transfun.2021vlp0002
[2021] Dynamic Verification of Approximate Computing Circuits using Coverage-based Grey-box Fuzzing
DOI: https://doi.org/10.1109/iolts52814.2021.9486690
[2021] A Synthesis Method Based on Multi-Stage Optimization for Power-Efficient Integrated Optical Logic Circuits
DOI: https://doi.org/10.1587/transfun.2020kep0018
[2021] Neural Network Calculations at the Speed of Light Using Optical Vector-Matrix Multiplication and Optoelectronic Activation
DOI: https://doi.org/10.1587/transfun.2020kep0016
[2021] Integration of Minimum Energy Point Tracking and Soft Real-Time Scheduling for Edge Computing
DOI: https://doi.org/10.1109/isqed51717.2021.9424343
[2021] Optical-electronic implementation of artificial neural network for ultrafast and accurate inference processing
DOI: https://doi.org/10.1117/12.2577966
[2021] Critical Path Isolation and Bit-Width Scaling Are Highly Compatible for Voltage Over-Scalable Design
DOI: https://doi.org/10.23919/date51398.2021.9473946

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（33 件）

科研費（0 件）

所属学会・役職（0 件）