Mr. Alexander Sludds Profile

Alexander Sludds

at Massachusetts Institute of Technology

SPIE Involvement:

Author

Publications (8)

Proceedings Article | 12 March 2024 Paper

Optimized encoding for coherent communication using photonic crystal cavity IQ modulators in thin film lithium niobate

Dashiell L. Vitullo, Sanjaya Lohani, Hugo Larocque, Alexander Sludds, Hamed Sattari, Ian Christen, Gregory Choong, Ivan Prieto, Jacopo Leo, Homa Zarebidaki, Moe Soltani, Thomas Searles, Amir Ghadimi, Brian Kirby, Mikkel Heuck, Dirk Englund

Proceedings Volume 12871, 1287107 (2024) https://doi.org/10.1117/12.3011172

KEYWORDS: Modulators, Electrooptic modulators, Thin films, Photonic crystals, Neurons, Lithium niobate, Telecommunications, Optical modulators, Artificial intelligence, Neural networks, Machine learning

Read Abstract +

Proceedings Article | 15 March 2023 Presentation + Paper

Multiplexing methods for scaling up photonic logic

Ryan Hamerly, Saumil Bandyopadhyay, Alexander Sludds, Zaijun Chen, Liane Bernstein, Sri Krishna Vadlamani, Jasvith Basani, Ronald Davis, Dirk Englund

Proceedings Volume 12438, 1243802 (2023) https://doi.org/10.1117/12.2650902

KEYWORDS: Vertical cavity surface emitting lasers, Wavelength division multiplexing, Multiplexing, Modulation, Matrices, Integrated optics, Detector arrays, Design and modelling, Free space optics, Signal detection

Read Abstract +

Proceedings Article | 15 March 2023 Presentation + Paper

Coherent VCSEL homodyne neural networks

Zaijun Chen, Alexander Sludds, Ronald Davis, Ian Christen, Liane Bernstein, Tobias Heuser, Niels Heermeier, James Lott, Stephan Reitzenstein, Ryan Hamerly, Dirk Englund

Proceedings Volume 12438, 1243809 (2023) https://doi.org/10.1117/12.2648628

KEYWORDS: Vertical cavity surface emitting lasers, Neural networks, Homodyne detection, Optical computing, Matrices, Energy efficiency, Electro optics, Quantum processing, Quantum machine learning, Parallel computing

Read Abstract +

Proceedings Article | 24 May 2022 Presentation

Design of efficient modally phase-matched SHG waveguides with automatic differentiation

Ryan Hamerly, Alexander Sludds, Saumil Bandyopadhyay, David Heydari, Mircea Catuneanu, Dirk Englund, Kambiz Jamshidi

Proceedings Volume PC12130, PC1213002 (2022) https://doi.org/10.1117/12.2624485

KEYWORDS: Waveguides, Second-harmonic generation, Structural design, Metamaterials, Supercontinuum generation, Slow light, Silicon, Resonators, Quantum optics, Optical parametric oscillators

Read Abstract +

Proceedings Article | 3 March 2022 Presentation + Paper

Design of asymptotically perfect linear photonic circuits

Ryan Hamerly, Saumil Bandyopadhyay, Alexander Sludds, Dirk Englund

Proceedings Volume 12019, 120190D (2022) https://doi.org/10.1117/12.2610494

KEYWORDS: Photonic integrated circuits, Interferometers, Error analysis, Numerical simulations, Integrated optics, Phase shifts, Optical calibration, Matrices, Mach-Zehnder interferometers

Read Abstract +

Proceedings Article | 2 September 2021 Presentation + Paper

Edge computing with optical neural networks via WDM weight broadcasting

Ryan Hamerly, Alexander Sludds, Saumil Bandyopadhyay, Liane Bernstein, Zaijun Chen, Manya Ghobadi, Dirk Englund

Proceedings Volume 11804, 118041R (2021) https://doi.org/10.1117/12.2594886

KEYWORDS: Wavelength division multiplexing, Neural networks, Integrated optics, Modulators, Computer programming, Modulation, Analog electronics, Time-frequency analysis, Signal to noise ratio

Read Abstract +

Proceedings Article | 5 March 2021 Presentation + Paper

Progress on photonic tensor processors based on time multiplexing and photoelectric multiplication

Ryan Hamerly, Alexander Sludds, Liane Bernstein, Vivienne Sze, Joel Emer, Marin Soljacic, Dirk Englund

Proceedings Volume 11680, 116800E (2021) https://doi.org/10.1117/12.2576990

KEYWORDS: Multiplexing, Neural networks, Free space optics, Optical components, Logic, Yield improvement, Transmitters, Transistors, Silicon photonics, Receivers

Read Abstract +

Proceedings Article | 24 February 2020 Presentation + Paper

A scalable optical neural network architecture using coherent detection

Alexander Sludds, Liane Bernstein, Ryan Hamerly, Marin Soljacic, Dirk Englund

Proceedings Volume 11299, 112990H (2020) https://doi.org/10.1117/12.2546940

KEYWORDS: Neural networks, Computer architecture, Convolutional neural networks, Homodyne detection, Electronics, Photonics, Photodetectors, Matrix multiplication

Read Abstract +

Storing, processing, and learning from data is a central task in both industrial practice and modern science. Recent advances in modern statistical learning, particularly Deep Neural Networks (DNNs), have given record breaking performance on tasks in game playing,^{1, 2} natural language processing,³ computer vision,⁴ computational biology,^{5, 6} and many others. The rapid growth of the field has been driven by an increase in the amount of public datasets,⁷ improvements to algorithms,⁸ and a substantial growth in computing power.⁹ In order to perform well on these tasks networks have had to grow in size, learning more complicated statistical features. The training and deployment of these large neural networks has spurred the creation of many neural network accelerators to aid in the computation of these networks.^10-12

Existing general purpose computing devices such as CPUs and GPUs are limited both by thermal dissipation per unit area and yield associated with large chips.^{13, 14} The design of Application Specific Integrated circuits (ASICs) has aided in decreasing the energy consumption per workload substantially by limiting the supported operations on chip. An example of this is the first generation tensor processing unit (TPU)¹⁵ which is able to perform the inference of large convolutional neural networks in datacenter in <10ms with an idle power of 28W and an workload power of 40W. It may seen counterintuitive then that the limiting factor for the implementation of DNNs is not computation, but rather the energy and bandwidth associated with reading and writing data from memory as well as the energy cost of moving data inside of the ASIC.^{15, 16} Several emerging technologies, such as in-memory computing,¹⁷ memristive crossbar arrays¹⁸ promise increased performance, but these emerging architectures suffer from calibration issues and limited accuracy.¹⁹

Photonics as a field has had tremendous success in improving the energy efficiency of data interconnects.²⁰ This has motivated the creation of optical neural networks (ONNs) based on 3D-printed diffractive elements,²¹ spiking neural networks utilizing ring-resonators,²² reservoir computing²³ and nanophotonic circuits.²⁴ However, these architectures have several issues. 3D-printed diffractive networks and schemes requiring spatial light modulators are non-programmable, meaning that they are unable to perform the task of training. Nanophotonic circuits allow for an O(N²) array of interferometers to be programmed, providing passive matrix-vector multiplication. However, the large (≈1mm²) size of on chip electro-optic interferometers means that scaling to an array of 100x100 would require 10; 000mm² of silicon, demonstrating the limitations of scaling this architecture. To date no architecture has demonstrated high-speed (GHz) speed computation with more than N ≥ 10; 000 neurons.

Here we present an architecture that is scalable to N ≥ 10⁶ neurons. The key mechanism of this architecture is balanced homodyne detection. By scaling the architecture to such a large size we show that we can decimate energy costs per operation associated with the optical component of this architecture, reaching a bound set by shot noise on the receiving photodetectors which leads to classification error. We call this bound a standard quantum limit (SQL) which reaches 100zJ/MAC on problems such as MNIST. We also analyze the energy consumption using existing technologies and show that sub-fJ/MAC energy consumption should be possible.

This paper is organized as follows: In section 1 we will discuss the function of this architecture as a matrixmatrix processor. In section 2 we will analyze the energy consumption of the architecture. In section 3 we will discuss methods for training and extending the accelerator to a broader scope of problems, namely convolutionally neural networks (CNNs).

Showing 5 of 8 publications

View contact details

UPDATE YOUR PROFILE

Is this your profile? Update it now.

Sign into your SPIE.org account

Don’t have a profile and want one?

Create an account on SPIE.org

Keywords/Phrases

Search In:

Publication Years