Hangbo Yang

Hangbo Yang

Auto-generated profile, based on publicly available information

Hangbo Yang

Univ. of Florida (United States)

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 September 2025 in Advanced Photonics
Verified
Journal Article
Open access
HYHangbo Yang
Hangbo Yang
NPNicola Peserico
Nicola Peserico
SLShurui Li
Shurui Li
XMXiaoxuan Ma
Xiaoxuan Ma
RSRussell L. T. Schwartz
Russell L. T. Schwartz
VSVolker J. Sorger
Volker J. Sorger
Convolutional operations are computationally intensive in artificial intelligence (AI) services, and their overhead in electronic hardware limits machine learning scaling. Here, we introduce a photonic joint transform correlator (pJTC) using a near-energy-free on-chip Fourier transformation to accelerate convolution operations. The pJTC reduces computational complexity for both convolution and cross-correlation from O(N4) to O(N2), where N2 is the input data size. Demonstrating functional Fourier transforms and convolution, this pJTC achieves 98.0% accuracy on an exemplary Modified National Institute of Standards and Technology inference task. Furthermore, a wavelength-multiplexed pJTC architecture shows potential for high throughput and energy efficiency, reaching 305 TOPS/W and 40.2 TOPS/mm2, based on currently available foundry processes. An efficient, compact, and low-latency convolution accelerator promises to advance next-generation AI capabilities across edge demands, high-performance computing, and cloud services.

9 May 2022 in IEEE Journal of Solid-State Circuits
Verified
Journal Article
Closed access
LCLi-Yang Chen
Li-Yang Chen
AVAbhinav Kumar Vinod
Abhinav Kumar Vinod
JMJames F. McMillan
James F. McMillan
HYHangbo Yang
Hangbo Yang
CWChee Wei Wong
Chee Wei Wong
CYC.-K. Ken Yang
C.-K. Ken Yang
This article presents a pulsed-coherent lidar that uses a 19-GHz pulse-modulated optical carrier and segmented time-of-flight (ToF) detection. The receiver analog frontend (AFE) employs a down-conversion chain that consists of a phase-invariant programmable-gain low-noise amplifier (PI-PGLNA), a sub-harmonic mixer (SHM), and an inverter-based phase-invariant programmable-gain amplifier (PI-PGA). The local oscillator (LO) is generated from a phase-locked loop (PLL) with a ring-type voltage-controlled oscillator (VCO). An analog-to-digital converter (ADC) subsamples the down-converted signal, and a digital signal processor (DSP) calculates the coarse, intermediate, and fine ToF. At the receiver input, a narrowband matching is implemented between the p-i-n photodiode and the PI-PGLNA with direct wire bonding. The system achieves 6- $\mu \text{m}$ precision with a 5-MSa/s sampling rate at a 2.5-m distance. The 2-D scanning is achieved with this lidar with the use of an MEMS mirror.

28 October 2023
Verified
Conference Paper
Open access
SLShurui Li
Shurui Li
HYHangbo Yang
Hangbo Yang
CWChee Wei Wong
Chee Wei Wong
VSVolker Sorger
Volker Sorger
PGPuneet Gupta
Puneet Gupta
In recent years, there has been a significant focus on achieving low-latency and high-throughput convolutional neural network (CNN) inference. Integrated photonics offers the potential to substantially expedite neural networks due to its inherent low-latency properties. Recently, on-chip Fourier optics-based neural network accelerators have been demonstrated and achieved superior energy efficiency for CNN acceleration. By incorporating Fourier optics, computationally intensive convolution operations can be performed instantaneously through on-chip lenses at a significantly lower cost compared to other on-chip photonic neural network accelerators. This is thanks to the complexity reduction offered by the convolution theorem and the passive Fourier transforms computed by on-chip lenses. However, conversion overhead between optical and digital domains and memory access energy still hinder overall efficiency. We introduce ReFOCUS, a Joint Transform Correlator (JTC) based on-chip neural network accelerator that efficiently reuses light through optical buffers. By incorporating optical delay lines, wavelength-division multiplexing, dataflow, and memory hierarchy optimization, ReFOCUS minimizes both conversion overhead and memory access energy. As a result, ReFOCUS achieves 2 × throughput, 2.2 × energy efficiency, and 1.36 × area efficiency compared to state-of-the-art photonic neural network accelerators.

1 January 2021
Verified
Conference Paper
Closed access
PDF file
HBHyunpil Boo
Hyunpil Boo
YLYoo Seung Lee
Yoo Seung Lee
HYHangbo Yang
Hangbo Yang
BMBrian Matthews
Brian Matthews
TLTom G. Lee
Tom G. Lee
CWChee Wei Wong
Chee Wei Wong
We report state-of-the-art metasurface optical elements enabling our high-resolution full color prototype, via various combination of analytical and numerical simulations. Our prototype device has achieved FoV (>40o), high output efficiency (>1%), and 1080-pixel resolution.

1 January 2020
Verified
Conference Paper
Closed access
PDF file
HBHyunpil Boo
Hyunpil Boo
YLYoo Seung Lee
Yoo Seung Lee
HYHangbo Yang
Hangbo Yang
KCKai-Chi Chang
Kai-Chi Chang
CWChee Wei Wong
Chee Wei Wong
We demonstrate metasurfaces consisting of meta-atoms that generate spatial-mode entangled photons with different polarization, designed through careful finite-difference time-domain simulations. Using our biphoton comb setup, these metasurfaces will enable for testing spatial-energy-time hyperentanglement.

21 March 2025
Verified
Conference Paper
Closed access
NPNicola Peserico
Nicola Peserico
HYHangbo Yang
Hangbo Yang
RSRussell L. T. Schwartz
Russell L. T. Schwartz
SASalem Altaleb
Salem Altaleb
BMBenyamin F. Motlagh
Benyamin F. Motlagh
VSVolker J. Sorger
Volker J. Sorger
The exponential grow of neural network algorithms has push for novel hardware solutions that can keep the pace of the demanding in terms of speed, size, latency, and power consumption. Here, we present the experimental realization and test of a high-speed silicon photonics system implementing a 1D Joint Transform Correlator for Convolution Neural Network. By designing, combining, and testing the photonic chip, the high-speed PCB with integrated TIAs, and the FPGA, we show the implementation of 16-inputs and 16-output channels JTC with 1 GHz bandwidth, for a total of 0.52 TOPS. Moreover, we present the next generation of the chip using a monolithic electro-photonic approach and with a higher number of channels for further reductions of sizes and power