BHBenyamin Haghi

Benyamin Haghi

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Benyamin Haghi has a total of 23 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Multi-Feature Fusion and Compressed Bi-LSTM for Memory-Efficient Heartbeat Classification on Wearable Devices

    by Reza Nikandish, Jiayu He, Benyamin Haghi

    In this paper, we present a memory-efficient ECG based heartbeat classification for wearable devices enabled by multi-feature fusion and compressed bidirectional long short term memory (Bi-LSTM). A multi-feature fusion technique based on time intervals and under-the-curve areas is proposed to extract the main characteristic points of the ECG waveform with high accuracy and robustness against noise and artifacts. A Bi-LSTM network is developed to process the input sequence in both forward and backward directions, resulting in higher accuracy and a 28% smaller network size compared with a conventional LSTM. Multiple neural networks with varying sizes, including tiny (84k), small (150k), medium (478k), and large (1.25M) models, are developed to achieve high accuracy across all classes. The overall accuracy is 96.4% for the large model and 94.6% for the tiny model, while the F1 score across all classes exceeds 89.1% and 85.1%, respectively. The proposed models, compressed using post-training quantization techniques, achieve state-of-the art performance. The compressed large model with 8-bit integer quantization (INT8) achieves an accuracy of 88.4% with 1.3MB of memory, while the compressed tiny model with dynamic range quantization achieves 94.6% accuracy with only 139kB of memory.


  • A 192-Channel 1D CNN-Based Neural Feature Extractor in 65nm CMOS for Brain-Machine Interfaces

    by Steven P. Bulfer, Jorge Gámez, Albert Yan-Huang, Benyamin Haghi, Volnei A. Pedroni, Richard A. Andersen, Azita Emami

    We present a 192-channel 1D convolutional neural network (1D CNN) based neural feature extractor for Brain-Machine Interfaces (BMI) that achieves state-of-the-art decoding stability at $\mathbf{1.8\ \mu W}$ and 12801 $\mathbf{\mu m^{2}}$ per channel in 65nm CMOS technology. Our device is a fully configurable, scalable, area and power efficient solution that supports models with 2-8 feature layers and a total kernel length of up to 256. This architecture reduces caching requirements by $\mathbf{5}{\boldsymbol\times}$ over conventional computation schemes. Channels and layers are individually power-switchable to further optimize power efficiency for a given neural application. We introduce an on-chip model, FENet-66, that achieves the highest cross-validated decoding performance compared to all previously reported feature sets. We show that this model maintains superior stability over time using recorded data from tetraplegic human participants with spinal cord injury. Our features have 18% higher overall average cross-validated R2 decoding performance compared to Spiking Band Power (SBP), with 28% better performance during the 4th year. Our proposed architecture can also extract mean wavelet power features at low power and latency. We show that custom 1D-CNN kernels achieve 10% better performance compared to wavelet features while compressing the neural data stream by $\mathbf{38}{\boldsymbol\times}$. The models and hardware were validated in real time with a human subject in online closed-loop center-out cursor control experiments with micro-electrode arrays that were implanted for 6 years. Decoders using features generated with this work substantially improve the viability of long-term neural implants compared to other feature extraction methods currently present in low power BMI hardware.

  • Open Access

    Author Correction: Enhanced control of a brain–computer interface by tetraplegic participants via neural-network-mediated feature extraction

    by Benyamin Haghi, Tyson Aflalo, Spencer Kellis, Charles Guan, Jorge A. Gamez de Leon, Albert Yan Huang, Nader Pouratian, Richard A. Andersen, Azita Emami

  • Enhanced control of a brain–computer interface by tetraplegic participants via neural-network-mediated feature extraction

    by Benyamin Haghi, Tyson Aflalo, Spencer Kellis, Charles Guan, Jorge A. Gamez de Leon, Albert Yan Huang, Nader Pouratian, Richard A. Andersen, Azita Emami

    To infer intent, brain–computer interfaces must extract features that accurately estimate neural activity. However, the degradation of signal quality over time hinders the use of feature-engineering techniques to recover functional information. By using neural data recorded from electrode arrays implanted in the cortices of three human participants, here we show that a convolutional neural network can be used to map electrical signals to neural features by jointly optimizing feature extraction and decoding under the constraint that all the electrodes must use the same neural-network parameters. In all three participants, the neural network led to offline and online performance improvements in a cursor-control task across all metrics, outperforming the rate of threshold crossings and wavelet decomposition of the broadband neural data (among other feature-extraction techniques). We also show that the trained neural network can be used without modification for new datasets, brain areas and participants.


  • EKGNet: A 10.96μW Fully Analog Neural Network for Intra-Patient Arrhythmia Classification

    by Benyamin Haghi, Lin Ma, Sahin Lale, Anima Anandkumar, Azita Emami

    We present an integrated approach by combining analog computing and deep learning for electrocardiogram (ECG) arrythmia classification. We propose EKGNet, a hardware-efficient and fully analog arrythmia classification architecture that achieves high accuracy with low power consumption. The proposed architecture leverages the energy efficiency of transistors operating in the subthreshold region, eliminating the need for analog-to-digital converters (ADC) and static random-access memory (SRAM). The system design includes a novel analog sequential Multiply-Accumulate (MAC) circuit that mitigates process, supply voltage, and temperature variations. Experimental evaluations on PhysionNet’s MIT-BIH and PTB Diagnostics datasets demonstrate the effectiveness of the proposed method, achieving an average balanced accuracies of 95% and 94.25% for intra-patient arrhythmia classification and myocardial infarction (MI) classification, respectively. This innovative approach presents a promising avenue for developing low power arrythmia classification systems with enhanced accuracy and transferability in biomedical applications.


  • Unsupervised ECG Analysis: A Review

    by Kasra Nezamabadi, Neda Sardaripour, Benyamin Haghi, Mohamad Forouzanfar

    Electrocardiography is the gold standard technique for detecting abnormal heart conditions. Automatic analysis of electrocardiogram (ECG) can help physicians in the interoperation of the large amount of data produced daily by cardiac monitors. As the successful application of supervised machine learning algorithms relies on unprecedented amounts of labeled training data, there is a growing need for unsupervised algorithms for ECG analysis. Unsupervised learning aims to partition ECG into distinct abnormality classes without cardiologist-supplied labelsa process referred to as ECG clustering. In addition to abnormality detection, ECG clustering can discover inter and intra-individual patterns that carry valuable information about the whole body and mind, such as emotions and mental disorders. ECG clustering can also resolve specific challenges facing supervised learning systems, such as the imbalanced data problem, and can enhance biometric systems. While several reviews exist on supervised ECG analysis, a comprehensive review of unsupervised ECG analysis techniques is still lacking. This study reviews recent ECG clustering techniques with the focus on machine learning and deep learning algorithms. We critically review and compare these techniques, discuss their applications and limitations, and provide future research directions. This review provides further insights into ECG clustering and presents the necessary information required to adopt the appropriate algorithm for a specific application.


  • Energy-Efficient Classification for Resource-Constrained Biomedical Applications

    by Mahsa Shoaran, Benyamin Allahgholizadeh Haghi, Milad Taghavi, Masoud Farivar, Azita Emami-Neyestanak

    Biomedical applications often require classifiers that are both accurate and cheap to implement. Today, deep neural networks achieve state-of-the-art accuracy in most learning tasks that involve large datasets of unstructured data. However, the application of deep learning techniques may not be beneficial in problems with limited training sets and computational resources, or under domain-specific test time constraints. Among other algorithms, ensembles of decision trees, particularly the Gradient Boosted models have recently been very successful in machine learning competitions. Here, we propose an efficient hardware architecture to implement gradient boosted trees in applications under stringent power, area, and delay constraints, such as medical devices. Specifically, we introduce the concepts of asynchronous tree operation and sequential feature extraction to achieve an unprecedented energy and area efficiency. The proposed architecture is evaluated in automated seizure detection for epilepsy, using 3074 hours of intracranial EEG data from 26 patients with 393 seizures. Average F1 scores of 99.23% and 87.86% are achieved for random and block-wise splitting of data into train/test sets, respectively, with an average detection latency of 1.1s. The proposed classifier is fabricated in a 65nm TSMC process, consuming 41.2 nJ/class in a total area of 540×1850 μm2. This design improves the state-of-the-art by 27× reduction in energy-area-latency product. Moreover, the proposed gradient-boosting architecture offers the flexibility to accommodate variable tree counts specific to each patient, to trade the predictive accuracy with energy. This patient-specific and energy-quality scalable classifier holds great promise for low-power sensor data classification in biomedical applications.