BKBehnam Khaleghi

Behnam Khaleghi

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Behnam Khaleghi has a total of 28 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    TRINE: A Token-Aware, Runtime-Adaptive FPGA Inference Engine for Multimodal AI

    by Hyunwoo Oh, Hanning Chen, Sanggeon Yun, Yang Ni, Suyeon Jang, Behnam Khaleghi, Fei Wen, Mohsen Imani

  • Private and Efficient Learning With Hyperdimensional Computing

    by Behnam Khaleghi, Xiaofan Yu, Jaeyoung Kang, Xuan Wang, Tajana Rosing

    Machine learning algorithms, especially deep neural networks (DNNs), are computationally expensive and vulnerable to privacy breaches. Hyperdimensional Computing (HDC) is a brain-inspired machine learning paradigm which offers a potential alternative for efficient and private learning. HDC encodes dense, sensory inputs into sparse, high-dimensional vectors of 10,000 bits or more, and learns using well-defined and lightweight operations on these vectors. However, naïve HDC is vulnerable to privacy attacks as the encoding process can be reversely determined. In this paper, we systematically study the fundamental privacy challenges of HDC due to reversibility and propose an end-to-end private training and inference framework for HDC with efficient hardware implementation. We first show that HDC is not naturally private as the original data and encoding parameters can be reconstructed from the encoded vector, regardless of the encoding methods applied. We term this attack as the feature extraction attack. We then propose a locally sparse encoding method to obfuscate the encoded information for protecting inference privacy, leading to computational and communication savings. We further propose Privé -HDnn which uses a hybrid architecture of Convolutional Neural Network (CNN) as feature extractors and private-preserved HDC classifier for complex tasks such as image classification. Thanks to the noise tolerance of HDC, Privé-HDnn enables private single- and multi-pass training of HDC by differential privacy-guaranteed noise injection. Results show that our new encoding method defends against decoding attacks with an RMSE deviation of 0.42, while also speeding up encoding by 2.6-5.3× (19-79×) compared to the FPGA (CPU) baseline. For privacy-preserved training, Privé - HDnn achieves comparable or up to 5.8% better accuracy than DNNs with up to 1231× faster training time on GPU as compared to the state-of-the-art NNs.


  • Open Access

    𝖧𝗒𝖣𝖱𝖤𝖠: Utilizing Hyperdimensional Computing for a More Robust and Efficient Machine Learning System

    by Justin Morris, Kazim Ergun, Behnam Khaleghi, Mohen Imani, Baris Aksanli, Tajana Simunic

    Today’s systems rely on sending all the data to the cloud and then using complex algorithms, such as Deep Neural Networks, which require billions of parameters and many hours to train a model. In contrast, the human brain can do much of this learning effortlessly. Hyperdimensional (HD) Computing aims to mimic the behavior of the human brain by utilizing high-dimensional representations. This leads to various desirable properties that other Machine Learning (ML) algorithms lack, such as robustness to noise in the system and simple, highly parallel operations. In this article, we propose 𝖧𝗒𝖣𝖱𝖤𝖠, a HyperDimensional Computing system that is Robust, Efficient, and Accurate. We propose a Processing-in-Memory (PIM) architecture that works in a federated learning environment with challenging communication scenarios that cause errors in the transmitted data. 𝖧𝗒𝖣𝖱𝖤𝖠 adaptively changes the bitwidth of the model based on the signal-to-noise ratio (SNR) of the incoming sample to maintain the accuracy of the HD model while achieving significant speedup and energy efficiency. Our PIM architecture is able to achieve a speedup of 28× and 255× better energy efficiency compared to the baseline PIM architecture for Classification and achieves 32 × speed up and 289 × higher energy efficiency than the baseline architecture for Clustering. 𝖧𝗒𝖣𝖱𝖤𝖠 is able to achieve this by relaxing hardware parameters to gain energy efficiency and speedup while introducing computational errors. We show experimentally, HD Computing is able to handle the errors without a significant drop in accuracy due to its unique robustness property. For wireless noise, we found that 𝖧𝗒𝖣𝖱𝖤𝖠 is 48 × more robust to noise than other comparable ML algorithms. Our results indicate that our proposed system loses less than 1% Classification accuracy, even in scenarios with an SNR of 6.64. We additionally test the robustness of using HD Computing for Clustering applications and found that our proposed system also looses less than 1% in the mutual information score, even in scenarios with an SNR under 7 dB, which is 57 × more robust to noise than K-means.


  • OpenHD: A GPU-Powered Framework for Hyperdimensional Computing

    by Jaeyoung Kang, Behnam Khaleghi, Tajana Rosing, Yeseong Kim

    Hyperdimensional computing (HDC) has emerged as an alternative lightweight learning solution to deep neural networks. A key characteristic of HDC is the great extent of parallelism that can facilitate hardware acceleration. However, previous hardware implementations of HDC seldom focus on GPU designs, which were also inefficient partly due to the complexity of accelerating HDC on GPUs. In this paper, we present OpenHD, a flexible and high-performance GPU-powered framework for automating the mapping of general HDC applications including classification and clustering to GPUs. OpenHD takes advantage of memory optimization strategies specialized for HDC, minimizing the access time to different memory subsystems, and removing redundant operations. We also propose a novel training method to enable data parallelism in HDC training. Our evaluation result shows that the proposed training rapidly achieves the target accuracy, reducing the required training epochs by 4×. With OpenHD, users can deploy GPU-accelerated HDC applications without domain expert knowledge. Compared to the state-of-the-art GPU-powered HDC implementation, our evaluation on NVIDIA Jetson TX2 shows that OpenHD is up to 10.5× and 314× faster for HDC-based classification and clustering, respectively. Compared with non-HDC classification and clustering on GPUs, OpenHD-based HDC is 11.7× and 53× faster at comparable accuracy. OpenHD is available at: https://github.com/UCSD-SEELab/openhd .


  • Voltage Adaptation Under Temperature Variation

    by Hussam Amrouch, Behnam Khaleghi, Jorg Henkel

    The maximum delay of a processor is subject to both temperature (T) and voltage (V) jointly. Therefore, there is the prospect of reducing V at runtime whenever T is below the worst-case temperature without violating the predetermined timing constraints, instead of employing throughout a conservative V that corresponds to Tworst. Hence, the efficiency can be maximized, while still accounting for temperature variation at runtime, without sacrificing reliability. In this work, we focus on the key challenge of how the small, yet sufficient voltage can be accurately, at design time, obtained in which timing constraints at runtime are assuredly fulfilled. To achieve that, we model the delay of processor under the joint effects of T and V through creating cell libraries that contain the detailed delay information of cells at a wide variety of T and V cases. Libraries are compatible with commercial EDA tools. Hence, standard tool flows, even though they were not designed for that purpose, can be employed to seamlessly obtain the correlation between T and V in any circuit regardless its complexity. After modeling of the correlation between T and V from the physical level to circuit level, we implement at the system level a temperature-guided voltage adaptation technique that tunes the voltage at runtime following temperature variations leading to a considerable reduction in power.


  • A Resistive RAM-Based FPGA Architecture Equipped With Efficient Programming Circuitry

    by Behnam Khaleghi, Hossein Asadi

    Despite the considerable effort has been put on the application of Non-Volatile Memories (NVMs) in Field-Programmable Gate Arrays FPGAs, previously suggested designs are not mature enough to substitute the state of-the-art SRAM-based counterparts mainly due to the inefficient building blocks and/or the overhead of programming structure which can impair their potential benefits. In this paper, we present a Resistive Random Access Memory RRAM-based FPGA architecture employing efficient Switch Box (SB) and Look-Up Table (LUT) designs with programming circuitry integrated in both SB and LUT designs that creates area and power efficient programmable components while precluding performance overhead to these blocks. In addition, we present an efficient scheme to load the configuration bitstream into the memory elements, which makes the configuration time comparable to that of SRAM-based FPGAs. Besides, we investigate the correct functionality and reliability of the programming structure subject to fluctuations in attributes of RRAM cells. Using Versatile Place and Route (VTR) tool with the obtained characteristics of the proposed blocks demonstrate that the average area and delay of the proposed FPGA architecture are 59.4% and 20.1% less than conventional SRAM-based FPGAs. Compared with a recent RRAM-based architecture, the proposed architecture improves the area and power by 49.7% and 33.8% while keeps the delay intact.


  • Designing Low Power and Durable Digital Blocks Using Shadow Nanoelectromechanical Relays

    by Sadegh Yazdanshenas, Behnam Khaleghi, Paolo Ienne, Hossein Asadi

    Nanoelectromechanical (NEM) relays are a promising emerging technology that has gained widespread research attention due to its zero leakage current, sharp ON-OFF transitions, and complementary metal-oxide-semiconductor compatibility. As a result, NEM relays have been significantly investigated as highly energy-efficient design solutions. A major shortcoming of NEMs preventing their widespread use is their limited switching endurance. Hence, in order to utilize the low-power advantages of NEM relays, further device, circuit, and architectural techniques are required. In this paper, we introduce the concept of shadow NEM relays, which is a circuit-level technique to leverage the energy efficiency of the NEM relays despite their low switching endurance. This technique creates two virtual ground nodes in a block to allow: 1) a low power mode with functional NEM relays and 2) a normal mode with failed NEM relays. To demonstrate the applicability of this concept, we have applied it to a six-transistor SRAM cell as an illustrative example. We also investigate the applicability of this SRAM cell in field-programmable gate arrays and on-chip caches. Experimental results reveal that shadow NEM relays can reduce the power consumption of SRAM cells by up to 80% while addressing the limited switching endurance of NEM relays.