Kentaro Yoshioka
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Kentaro Yoshioka has a total of 165 co-authors. Here are the top co-authors by number of shared publications:
- Akihide Sai
Toshiba,Kawasaki,Japan
6 shared publications - Wenlun Zhang
Keio University, Tokyo, Japan
6 shared publications - Xinyu Li
School of Electronic Science and Engineering, Nanjing University, Nanjing 210023, China
5 shared publications Hirotomo Ishii
Toshiba Corporation, Kawasaki 212-8582, Japan
4 shared publicationsKohei Onizuka
4 shared publicationsKoichiro Ban
4 shared publications
Latest Publications
- Open Access
- 14 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 4462-4476
Illusion of Depth: Revealing Hidden Stereo Vision Vulnerabilities in Depth Estimation
- 14 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 4462-4476
Open AccessIllusion of Depth: Revealing Hidden Stereo Vision Vulnerabilities in Depth Estimation
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 73 (11) , 8057-8070
A 65-nm Compute-in-Read-Only-Memory Macro With Bidirectional Ternary Weight Storage and Sparsity-Aware Local Accumulation for 1.58-bit LLM Inference
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 73 (11) , 8057-8070
A 65-nm Compute-in-Read-Only-Memory Macro With Bidirectional Ternary Weight Storage and Sparsity-Aware Local Accumulation for 1.58-bit LLM Inference
- 4 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
- Vol. 34 (5) , 1556-1566
MCRA: Multicolumn Residue Accumulation Analog Compute-in-Memory Architecture With Time-Domain M-Input ΣΔ ADC
Analog compute-in-memory (ACIM) architectures offer significant throughput and energy benefits by performing multiplication-and-accumulation (MAC) operations directly within memory arrays. However, their overall efficiency is fundamentally constrained by the high power consumption of the per-column high-resolution analog-to-digital converters (ADCs) required to support modern DNNs (e.g., transformers), many of which demand both high computational precision and large-column throughput. In conventional ADC designs, energy in the noise-limited regime scales near-exponentially, typically by $4\times $ per additional bit, making high-resolution ADCs on every column power-prohibitive. This article proposes a high-precision and power-efficient multicolumn residue accumulation (MCRA) ACIM architecture to efficiently support precision-demanding modern DNNs. Each column uses a low-resolution coarse ADC (cADC), while the per-column residuals are accumulated and further quantized by an energy-efficient time-domain multi-input incremental sigma–delta (Mi- $\Sigma \Delta $ ) fine ADC (fADC). This approach amortizes the near-exponential energy growth across columns, while exploiting the more favorable power-resolution scaling of the time-domain Mi- $\Sigma \Delta $ quantization. Postlayout simulations demonstrate a 66.2-dB signal-to-noise-and-distortion ratio (SNDR) per column at only 1/21 the energy of the baseline with per-column high-resolution ADCs, and $0.405\times $ (1/2.47) the energy of an energy-saving ADC per column, which achieves a similar SNDR. Circuit- and system-level simulations demonstrate that our MCRA CIM architecture achieves negligible accuracy degradation in precision-demanding ViT tasks while delivering high energy and area efficiency.
- 4 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
- Vol. 34 (5) , 1556-1566
MCRA: Multicolumn Residue Accumulation Analog Compute-in-Memory Architecture With Time-Domain M-Input ΣΔ ADC
Analog compute-in-memory (ACIM) architectures offer significant throughput and energy benefits by performing multiplication-and-accumulation (MAC) operations directly within memory arrays. However, their overall efficiency is fundamentally constrained by the high power consumption of the per-column high-resolution analog-to-digital converters (ADCs) required to support modern DNNs (e.g., transformers), many of which demand both high computational precision and large-column throughput. In conventional ADC designs, energy in the noise-limited regime scales near-exponentially, typically by $4\times $ per additional bit, making high-resolution ADCs on every column power-prohibitive. This article proposes a high-precision and power-efficient multicolumn residue accumulation (MCRA) ACIM architecture to efficiently support precision-demanding modern DNNs. Each column uses a low-resolution coarse ADC (cADC), while the per-column residuals are accumulated and further quantized by an energy-efficient time-domain multi-input incremental sigma–delta (Mi- $\Sigma \Delta $ ) fine ADC (fADC). This approach amortizes the near-exponential energy growth across columns, while exploiting the more favorable power-resolution scaling of the time-domain Mi- $\Sigma \Delta $ quantization. Postlayout simulations demonstrate a 66.2-dB signal-to-noise-and-distortion ratio (SNDR) per column at only 1/21 the energy of the baseline with per-column high-resolution ADCs, and $0.405\times $ (1/2.47) the energy of an energy-saving ADC per column, which achieves a similar SNDR. Circuit- and system-level simulations demonstrate that our MCRA CIM architecture achieves negligible accuracy degradation in precision-demanding ViT tasks while delivering high energy and area efficiency.
- Open Access
- 24 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 26 (7) , 11233-11242
LiDAR Beacon: 300-m-Range Fiducial Marker for Mobile Robot Localization With Dual Temporal Range–Reflectance Modulation
Autonomous mobile robots require accurate localization in large-scale environments where Global Navigation Satellite System (GNSS) signals are unavailable. Mobile LiDAR-detectable fiducial markers can provide absolute position references, but existing approaches encode identifiers (IDs) using static geometric/reflectance patterns, which limits detection range due to LiDAR resolution. We introduce LiDAR Beacon, a novel mobile LiDAR-based fiducial marker that achieves reliable long-range detection and identification. LiDAR Beacon encodes its ID by introducing controlled modulations in both the range and reflectance of LiDAR returns via an electrically switchable polymer-dispersed liquid crystal (PDLC) layer. This approach enables robust ID extraction using only a single-point LiDAR return tracked over several scan frames. To address different deployment needs, we introduce two operating modes prioritizing either decoding speed or installation flexibility. Experimental validation demonstrates exceptional performance, achieving a maximum outdoor detection range of ≈ 309m with a precision of 100.0% and a recall of 29.7% under targeted conditions. Furthermore, we demonstrate the system’s practical utility by integrating infrastructure-mounted LiDAR Beacons as absolute references into a factor-graph-based multi-sensor localization framework. Field experiments in large-scale outdoor and indoor environments show that our system significantly reduces marker density requirements while maintaining high localization accuracy, even when conventional LiDAR scan matching fails or only sparse markers are available.
- 24 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 26 (7) , 11233-11242
Open AccessLiDAR Beacon: 300-m-Range Fiducial Marker for Mobile Robot Localization With Dual Temporal Range–Reflectance Modulation
Autonomous mobile robots require accurate localization in large-scale environments where Global Navigation Satellite System (GNSS) signals are unavailable. Mobile LiDAR-detectable fiducial markers can provide absolute position references, but existing approaches encode identifiers (IDs) using static geometric/reflectance patterns, which limits detection range due to LiDAR resolution. We introduce LiDAR Beacon, a novel mobile LiDAR-based fiducial marker that achieves reliable long-range detection and identification. LiDAR Beacon encodes its ID by introducing controlled modulations in both the range and reflectance of LiDAR returns via an electrically switchable polymer-dispersed liquid crystal (PDLC) layer. This approach enables robust ID extraction using only a single-point LiDAR return tracked over several scan frames. To address different deployment needs, we introduce two operating modes prioritizing either decoding speed or installation flexibility. Experimental validation demonstrates exceptional performance, achieving a maximum outdoor detection range of ≈ 309m with a precision of 100.0% and a recall of 29.7% under targeted conditions. Furthermore, we demonstrate the system’s practical utility by integrating infrastructure-mounted LiDAR Beacons as absolute references into a factor-graph-based multi-sensor localization framework. Field experiments in large-scale outdoor and indoor environments show that our system significantly reduces marker density requirements while maintaining high localization accuracy, even when conventional LiDAR scan matching fails or only sparse markers are available.
- 17 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Solid-State Circuits
- Vol. 61 (3) , 977-989
A 4.2–373 K Functional 800-MS/s 12-b Buffer-Then-Amplify Charge-Pump-Based Pipelined TI-SAR ADC With Integrated Active-Hold Technique
This article presents a source-follower (SF) and charge-pump (CP)-based amplification that is linear and bias-calibration-free across a wide temperature range from 4.2 to 373 K, targeting pipelined analog-to-digital converters (ADCs). While prior SF and CP-based amplifiers suffer from poor linearity because the SF must drive large amplified signals, we propose a buffer-then-amplify (BTA) CP architecture that relocates the CP to the SF output. As a result, the SF only buffers unamplified small-swing residues, fully exploiting its intrinsic linearity and eliminating the need for high-order gain calibrations even in advanced process nodes. To further enhance robustness and efficiency, we introduce an adaptively biased class-AB SF. An integrated active-hold technique is also proposed to eliminate the timing and bandwidth mismatch between the first-pipeline-stage sampling capacitor and the sub-ADC by repurposing the SF as a hold buffer for the sub-ADC. The prototype 12-bit 800-MS/s pipelined time-interleaved successive approximation register (TI-SAR) ADC is fabricated in 12-nm FinFET technology. It achieves SNDR higher than 55.0 dB across the 4.2–373 K temperature range with 40-MHz input, and peak Walden figure-of-merit (FoMW) of 9.2 fJ/c.-step at 4.2 K. To the best of the authors’ knowledge, the achieved functional temperature range is the widest among >10-bit >300-MS/s ADCs.
- 17 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Solid-State Circuits
- Vol. 61 (3) , 977-989
A 4.2–373 K Functional 800-MS/s 12-b Buffer-Then-Amplify Charge-Pump-Based Pipelined TI-SAR ADC With Integrated Active-Hold Technique
This article presents a source-follower (SF) and charge-pump (CP)-based amplification that is linear and bias-calibration-free across a wide temperature range from 4.2 to 373 K, targeting pipelined analog-to-digital converters (ADCs). While prior SF and CP-based amplifiers suffer from poor linearity because the SF must drive large amplified signals, we propose a buffer-then-amplify (BTA) CP architecture that relocates the CP to the SF output. As a result, the SF only buffers unamplified small-swing residues, fully exploiting its intrinsic linearity and eliminating the need for high-order gain calibrations even in advanced process nodes. To further enhance robustness and efficiency, we introduce an adaptively biased class-AB SF. An integrated active-hold technique is also proposed to eliminate the timing and bandwidth mismatch between the first-pipeline-stage sampling capacitor and the sub-ADC by repurposing the SF as a hold buffer for the sub-ADC. The prototype 12-bit 800-MS/s pipelined time-interleaved successive approximation register (TI-SAR) ADC is fabricated in 12-nm FinFET technology. It achieves SNDR higher than 55.0 dB across the 4.2–373 K temperature range with 40-MHz input, and peak Walden figure-of-merit (FoMW) of 9.2 fJ/c.-step at 4.2 K. To the best of the authors’ knowledge, the achieved functional temperature range is the widest among >10-bit >300-MS/s ADCs.
- Open Access
- 27 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 78-86
BasketLiDAR: The First LiDAR-Camera Multimodal Dataset for Professional Basketball MOT
- 27 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 78-86
Open AccessBasketLiDAR: The First LiDAR-Camera Multimodal Dataset for Professional Basketball MOT
- Open Access
BitROM: Weight Reload-Free CiROM Architecture Towards Billion-Parameter 1.58-bit LLM Inference
Compute-in-Read-Only-Memory (CiROM) accelerators offer outstanding energy efficiency for CNNs by eliminating runtime weight updates. However, their scalability to Large Language Models (LLMs) is fundamentally constrained by their vast parameter sizes. Notably, LLaMA-7B - the smallest model in LLaMA series - demands more than 1,000 cm2 of silicon area even in advanced CMOS nodes. This paper presents BitROM, the first CiROM-based accelerator that overcomes this limitation through co-design with BitNet's 1.58-bit quantization model, enabling practical and efficient LLM inference at the edge. BitROM introduces three key innovations: 1) a novel Bidirectional ROM Array that stores two ternary weights per transistor; 2) a Tri-Mode Local Accumulator optimized for ternary-weight computations; and 3) an integrated Decode-Refresh (DR) eDRAM that supports on-die KV-cache management, significantly reducing external memory access during decoding. In addition, BitROM integrates LoRA-based adapters to enable efficient transfer learning across various downstream tasks. Evaluated in 65nm CMOS, BitROM achieves 20.8 TOPS/W and a bit density of 4,967 kB/mm2 - offering a 10x improvement in area efficiency over prior digital CiROM designs. Moreover, the DR eDRAM contributes to a 43.6% reduction in external DRAM access, further enhancing deployment efficiency for LLMs in edge applications.
Open AccessBitROM: Weight Reload-Free CiROM Architecture Towards Billion-Parameter 1.58-bit LLM Inference
Compute-in-Read-Only-Memory (CiROM) accelerators offer outstanding energy efficiency for CNNs by eliminating runtime weight updates. However, their scalability to Large Language Models (LLMs) is fundamentally constrained by their vast parameter sizes. Notably, LLaMA-7B - the smallest model in LLaMA series - demands more than 1,000 cm2 of silicon area even in advanced CMOS nodes. This paper presents BitROM, the first CiROM-based accelerator that overcomes this limitation through co-design with BitNet's 1.58-bit quantization model, enabling practical and efficient LLM inference at the edge. BitROM introduces three key innovations: 1) a novel Bidirectional ROM Array that stores two ternary weights per transistor; 2) a Tri-Mode Local Accumulator optimized for ternary-weight computations; and 3) an integrated Decode-Refresh (DR) eDRAM that supports on-die KV-cache management, significantly reducing external memory access during decoding. In addition, BitROM integrates LoRA-based adapters to enable efficient transfer learning across various downstream tasks. Evaluated in 65nm CMOS, BitROM achieves 20.8 TOPS/W and a bit density of 4,967 kB/mm2 - offering a 10x improvement in area efficiency over prior digital CiROM designs. Moreover, the DR eDRAM contributes to a 43.6% reduction in external DRAM access, further enhancing deployment efficiency for LLMs in edge applications.
- 5 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
- Vol. 33 (10) , 2838-2851
ASiM: Modeling and Analyzing Inference Accuracy of SRAM-Based Analog CiM Circuits
Static random-access memory (SRAM)-based analog compute-in-memory (ACiM) demonstrates promising energy efficiency for deep neural network (DNN) processing. Nevertheless, efforts to optimize efficiency frequently compromise accuracy, and this trade-off remains insufficiently studied due to the difficulty of performing full-system validation. Specifically, existing simulation tools rarely target SRAM-based ACiM and exhibit inconsistent accuracy predictions, highlighting the need for a standardized, SRAM compute-in-memory (CiM) circuit-aware evaluation methodology. This article presents ASiM, a simulation framework for evaluating inference accuracy in SRAM-based ACiM systems. ASiM captures critical effects in SRAM-based analog compute in memory systems, such as analog-to-digital converter (ADC) quantization, bit-parallel encoding, and analog noise, which must be modeled with high fidelity due to their distinct behavior in charge-domain architectures compared to other memory technologies. ASiM supports a wide range of modern DNN workloads, including CNN and Transformer-based models such as ViT, and scales to large-scale tasks like ImageNet classification. Our results indicate that bit-parallel encoding can improve energy efficiency with only modest accuracy degradation; however, even 1 LSB of analog noise can significantly impair inference performance, particularly in complex tasks such as ImageNet. To address this, we explore hybrid analog-digital execution and majority voting schemes, both of which enhance robustness without negating energy savings. ASiM bridges the gap between hardware design and inference performance, offering actionable insights for energy-efficient, high-accuracy ACiM deployment. The code is available at https://github.com/Keio-CSG/ASiM
- 5 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
- Vol. 33 (10) , 2838-2851
ASiM: Modeling and Analyzing Inference Accuracy of SRAM-Based Analog CiM Circuits
Static random-access memory (SRAM)-based analog compute-in-memory (ACiM) demonstrates promising energy efficiency for deep neural network (DNN) processing. Nevertheless, efforts to optimize efficiency frequently compromise accuracy, and this trade-off remains insufficiently studied due to the difficulty of performing full-system validation. Specifically, existing simulation tools rarely target SRAM-based ACiM and exhibit inconsistent accuracy predictions, highlighting the need for a standardized, SRAM compute-in-memory (CiM) circuit-aware evaluation methodology. This article presents ASiM, a simulation framework for evaluating inference accuracy in SRAM-based ACiM systems. ASiM captures critical effects in SRAM-based analog compute in memory systems, such as analog-to-digital converter (ADC) quantization, bit-parallel encoding, and analog noise, which must be modeled with high fidelity due to their distinct behavior in charge-domain architectures compared to other memory technologies. ASiM supports a wide range of modern DNN workloads, including CNN and Transformer-based models such as ViT, and scales to large-scale tasks like ImageNet classification. Our results indicate that bit-parallel encoding can improve energy efficiency with only modest accuracy degradation; however, even 1 LSB of analog noise can significantly impair inference performance, particularly in complex tasks such as ImageNet. To address this, we explore hybrid analog-digital execution and majority voting schemes, both of which enhance robustness without negating energy savings. ASiM bridges the gap between hardware design and inference performance, offering actionable insights for energy-efficient, high-accuracy ACiM deployment. The code is available at https://github.com/Keio-CSG/ASiM
- Open Access
BasketLiDAR: The First LiDAR-Camera Multimodal Dataset for Professional Basketball MOT
Real-time 3D trajectory player tracking in sports plays a crucial role in tactical analysis, performance evaluation, and enhancing spectator experience. Traditional systems rely on multi-camera setups, but are constrained by the inherently two-dimensional nature of video data and the need for complex 3D reconstruction processing, making real-time analysis challenging. Basketball, in particular, represents one of the most difficult scenarios in the MOT field, as ten players move rapidly and complexly within a confined court space, with frequent occlusions caused by intense physical contact. To address these challenges, this paper constructs BasketLiDAR, the first multimodal dataset in the sports MOT field that combines LiDAR point clouds with synchronized multi-view camera footage in a professional basketball environment, and proposes a novel MOT framework that simultaneously achieves improved tracking accuracy and reduced computational cost. The BasketLiDAR dataset contains a total of 4,445 frames and 3,105 player IDs, with fully synchronized IDs between three LiDAR sensors and three multi-view cameras. We recorded 5-on-5 and 3-on-3 game data from actual professional basketball players, providing complete 3D positional information and ID annotations for each player. Based on this dataset, we developed a novel MOT algorithm that leverages LiDAR's high-precision 3D spatial information. The proposed method consists of a real-time tracking pipeline using LiDAR alone and a multimodal tracking pipeline that fuses LiDAR and camera data. Experimental results demonstrate that our approach achieves real-time operation, which was difficult with conventional camera-only methods, while achieving superior tracking performance even under occlusion conditions. The dataset is available upon request at: https://sites.google.com/keio.jp/keio-csg/projects/basket-lidar
Open AccessBasketLiDAR: The First LiDAR-Camera Multimodal Dataset for Professional Basketball MOT
Real-time 3D trajectory player tracking in sports plays a crucial role in tactical analysis, performance evaluation, and enhancing spectator experience. Traditional systems rely on multi-camera setups, but are constrained by the inherently two-dimensional nature of video data and the need for complex 3D reconstruction processing, making real-time analysis challenging. Basketball, in particular, represents one of the most difficult scenarios in the MOT field, as ten players move rapidly and complexly within a confined court space, with frequent occlusions caused by intense physical contact. To address these challenges, this paper constructs BasketLiDAR, the first multimodal dataset in the sports MOT field that combines LiDAR point clouds with synchronized multi-view camera footage in a professional basketball environment, and proposes a novel MOT framework that simultaneously achieves improved tracking accuracy and reduced computational cost. The BasketLiDAR dataset contains a total of 4,445 frames and 3,105 player IDs, with fully synchronized IDs between three LiDAR sensors and three multi-view cameras. We recorded 5-on-5 and 3-on-3 game data from actual professional basketball players, providing complete 3D positional information and ID annotations for each player. Based on this dataset, we developed a novel MOT algorithm that leverages LiDAR's high-precision 3D spatial information. The proposed method consists of a real-time tracking pipeline using LiDAR alone and a multimodal tracking pipeline that fuses LiDAR and camera data. Experimental results demonstrate that our approach achieves real-time operation, which was difficult with conventional camera-only methods, while achieving superior tracking performance even under occlusion conditions. The dataset is available upon request at: https://sites.google.com/keio.jp/keio-csg/projects/basket-lidar