Kleber Stangherlin
Research Subjects
Kleber Stangherlin has a total of 16 co-authors. Here are the top co-authors by number of shared publications:
- Barbara De Salvo
Meta Reality Labs Research
3 shared publications - Manoj Sachdev
Department of Electrical and Computer Engineering, University of Waterloo, Waterloo, ON, Canada
3 shared publications - Syed Shakib Sarwar
Meta Reality Labs Research
3 shared publications - Chiao Liu
Meta Reality Labs
2 shared publications - Jae-Sun Seo
Cornell Tech, New York, NY, USA
2 shared publications - Jinhui Chen
Meta Reality Labs
2 shared publications
Latest Publications
- Open Access
- 28 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 28-30
H4H: Hybrid Convolution-Transformer Architecture Search for NPU-CIM Heterogeneous Systems for AR/VR Applications (Abstract)
- 28 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 28-30
Open AccessH4H: Hybrid Convolution-Transformer Architecture Search for NPU-CIM Heterogeneous Systems for AR/VR Applications (Abstract)
- Open Access
- 20 January 2025
- Published by Association for Computing Machinery (ACM)
- p. 1133-1141
H4H: Hybrid Convolution-Transformer Architecture Search for NPU-CIM Heterogeneous Systems for AR/VR Applications
Low-latency and low-power edge AI is crucial for Augmented/Virtual Reality applications. Recent advances demonstrate that hybrid models, combining convolution layers (CNN) and transformers (ViT), often achieve a superior accuracy/performance tradeoff on various computer vision and machine learning (ML) tasks. However, hybrid ML models can present system challenges for latency and energy efficiency due to their diverse nature in dataflow and memory access patterns. In this work, we leverage architecture heterogeneity from Neural Processing Units (NPU) and Compute-In-Memory (CIM) and explore diverse execution schemas for efficient hybrid model executions. We introduce H4H-NAS, a two-stage Neural Architecture Search (NAS) framework to automate the design of hybrid CNN/ViT models for heterogeneous edge systems featuring both NPU and CIM. We propose a two-phase incremental supernet training in our NAS to resolve gradient conflicts between sampled subnets caused by different block types in a hybrid model search space. Our H4H-NAS approach is also powered by a performance estimator built with NPU performance results measured on real silicon, and CIM performance based on industry IPs. H4H-NAS searches hybrid CNN-ViT models with fine granularity and achieves significant (up to 1.34%) top-1 accuracy improvement on ImageNet-1k. Moreover, results from our algorithm/hardware co-design reveal up to 56.08% overall latency and 41.72% energy improvements by introducing heterogeneous computing over baseline solutions. Overall, our framework guides the design of hybrid network architectures and system architectures for NPU+CIM heterogeneous systems.
- 20 January 2025
- Published by Association for Computing Machinery (ACM)
- p. 1133-1141
Open AccessH4H: Hybrid Convolution-Transformer Architecture Search for NPU-CIM Heterogeneous Systems for AR/VR Applications
Low-latency and low-power edge AI is crucial for Augmented/Virtual Reality applications. Recent advances demonstrate that hybrid models, combining convolution layers (CNN) and transformers (ViT), often achieve a superior accuracy/performance tradeoff on various computer vision and machine learning (ML) tasks. However, hybrid ML models can present system challenges for latency and energy efficiency due to their diverse nature in dataflow and memory access patterns. In this work, we leverage architecture heterogeneity from Neural Processing Units (NPU) and Compute-In-Memory (CIM) and explore diverse execution schemas for efficient hybrid model executions. We introduce H4H-NAS, a two-stage Neural Architecture Search (NAS) framework to automate the design of hybrid CNN/ViT models for heterogeneous edge systems featuring both NPU and CIM. We propose a two-phase incremental supernet training in our NAS to resolve gradient conflicts between sampled subnets caused by different block types in a hybrid model search space. Our H4H-NAS approach is also powered by a performance estimator built with NPU performance results measured on real silicon, and CIM performance based on industry IPs. H4H-NAS searches hybrid CNN-ViT models with fine granularity and achieves significant (up to 1.34%) top-1 accuracy improvement on ImageNet-1k. Moreover, results from our algorithm/hardware co-design reveal up to 56.08% overall latency and 41.72% energy improvements by introducing heterogeneous computing over baseline solutions. Overall, our framework guides the design of hybrid network architectures and system architectures for NPU+CIM heterogeneous systems.
- Open Access
x-RAGE: eXtended Reality -- Action & Gesture Events Dataset
With the emergence of the Metaverse and focus on wearable devices in the recent years gesture based human-computer interaction has gained significance. To enable gesture recognition for VR/AR headsets and glasses several datasets focusing on egocentric i.e. first-person view have emerged in recent years. However, standard frame-based vision suffers from limitations in data bandwidth requirements as well as ability to capture fast motions. To overcome these limitation bio-inspired approaches such as event-based cameras present an attractive alternative. In this work, we present the first event-camera based egocentric gesture dataset for enabling neuromorphic, low-power solutions for XR-centric gesture recognition. The dataset has been made available publicly at the following URL: https://gitlab.com/NVM_IITD_Research/xrage.
Open Accessx-RAGE: eXtended Reality -- Action & Gesture Events Dataset
With the emergence of the Metaverse and focus on wearable devices in the recent years gesture based human-computer interaction has gained significance. To enable gesture recognition for VR/AR headsets and glasses several datasets focusing on egocentric i.e. first-person view have emerged in recent years. However, standard frame-based vision suffers from limitations in data bandwidth requirements as well as ability to capture fast motions. To overcome these limitation bio-inspired approaches such as event-based cameras present an attractive alternative. In this work, we present the first event-camera based egocentric gesture dataset for enabling neuromorphic, low-power solutions for XR-centric gesture recognition. The dataset has been made available publicly at the following URL: https://gitlab.com/NVM_IITD_Research/xrage.
- 25 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
- Vol. 31 (1) , 55-64
Enhancing Strong PUF Security With Nonmonotonic Response Quantization
Strong physical unclonable functions (PUFs) provide a low-cost authentication primitive for resource-constrained devices. However, most strong PUF architectures can be modeled through learning algorithms with a limited number of CRPs. In this article, we introduce the concept of nonmonotonic response quantization for strong PUFs. Responses depend not only on which path is faster but also on the distance between the arriving signals. Our experiments show that the resulting PUF has increased security against learning attacks. To demonstrate, we designed and implemented a nonmonotonically quantized ring oscillator-based PUF in 65-nm technology. Measurement results show nearly ideal uniformity and uniqueness with a bit error rate of 13.4% over the temperature range from 0 °C to 50 °C.
- 25 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
- Vol. 31 (1) , 55-64
Enhancing Strong PUF Security With Nonmonotonic Response Quantization
Strong physical unclonable functions (PUFs) provide a low-cost authentication primitive for resource-constrained devices. However, most strong PUF architectures can be modeled through learning algorithms with a limited number of CRPs. In this article, we introduce the concept of nonmonotonic response quantization for strong PUFs. Responses depend not only on which path is faster but also on the distance between the arriving signals. Our experiments show that the resulting PUF has increased security against learning attacks. To demonstrate, we designed and implemented a nonmonotonically quantized ring oscillator-based PUF in 65-nm technology. Measurement results show nearly ideal uniformity and uniqueness with a bit error rate of 13.4% over the temperature range from 0 °C to 50 °C.
- Open Access
Design and Implementation of a Secure RISC-V Microprocessor
Secret keys can be extracted from the power consumption or electromagnetic emanations of unprotected devices. Traditional counter-measures have limited scope of protection, and impose several restrictions on how sensitive data must be manipulated. We demonstrate a bit-serial RISC-V microprocessor implementation with no plain-text data. All values are protected using Boolean masking. Software can run with little to no counter-measures, reducing code size and performance overheads. Unlike previous literature, our methodology is fully automated and can be applied to designs of arbitrary size or complexity. We also provide details on other key components such as clock randomizer, memory protection, and random number generator. The microprocessor was implemented in 65 nm CMOS technology. Its implementation was evaluated using NIST tests as well as side channel attacks. Random numbers generated with our RNG pass on all NIST tests. Side-channel analysis on the baseline implementation extracted the AES key using only 375 traces, while our secure microprocessor was able to withstand attacks using 20 M traces.
Open AccessDesign and Implementation of a Secure RISC-V Microprocessor
Secret keys can be extracted from the power consumption or electromagnetic emanations of unprotected devices. Traditional counter-measures have limited scope of protection, and impose several restrictions on how sensitive data must be manipulated. We demonstrate a bit-serial RISC-V microprocessor implementation with no plain-text data. All values are protected using Boolean masking. Software can run with little to no counter-measures, reducing code size and performance overheads. Unlike previous literature, our methodology is fully automated and can be applied to designs of arbitrary size or complexity. We also provide details on other key components such as clock randomizer, memory protection, and random number generator. The microprocessor was implemented in 65 nm CMOS technology. Its implementation was evaluated using NIST tests as well as side channel attacks. Random numbers generated with our RNG pass on all NIST tests. Side-channel analysis on the baseline implementation extracted the AES key using only 375 traces, while our secure microprocessor was able to withstand attacks using 20 M traces.
- 19 September 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
- Vol. 30 (11) , 1705-1715
Design and Implementation of a Secure RISC-V Microprocessor
Secret keys can be extracted from the power consumption or electromagnetic emanations of unprotected devices. Traditional countermeasures have a limited scope of protection and impose several restrictions on how sensitive data must be manipulated. We demonstrate a bit-serial RISC-V microprocessor implementation with no plain-text data. All values are protected using Boolean masking. Software can run with little to no countermeasures, reducing code size and performance overheads. Unlike previous literature, our methodology is fully automated and can be applied to designs of arbitrary size or complexity. We also provide details on other key components, such as clock randomizer, memory protection, and random number generator (RNG). The microprocessor was implemented in 65-nm CMOS technology. Its implementation was evaluated using NIST tests and side-channel attacks. Random numbers generated with our RNG pass on all NIST tests. The side-channel analysis on the baseline implementation extracted the advanced encryption system (AES) key using only 375 traces, while our secure microprocessor was able to withstand attacks using 20M traces.
- 19 September 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
- Vol. 30 (11) , 1705-1715
Design and Implementation of a Secure RISC-V Microprocessor
Secret keys can be extracted from the power consumption or electromagnetic emanations of unprotected devices. Traditional countermeasures have a limited scope of protection and impose several restrictions on how sensitive data must be manipulated. We demonstrate a bit-serial RISC-V microprocessor implementation with no plain-text data. All values are protected using Boolean masking. Software can run with little to no countermeasures, reducing code size and performance overheads. Unlike previous literature, our methodology is fully automated and can be applied to designs of arbitrary size or complexity. We also provide details on other key components, such as clock randomizer, memory protection, and random number generator (RNG). The microprocessor was implemented in 65-nm CMOS technology. Its implementation was evaluated using NIST tests and side-channel attacks. Random numbers generated with our RNG pass on all NIST tests. The side-channel analysis on the baseline implementation extracted the advanced encryption system (AES) key using only 375 traces, while our secure microprocessor was able to withstand attacks using 20M traces.