Inverse-Designed Dot Product Engine

2025-07-18

AI Summary

To address the large area and high power consumption of dot-product engines in photonic AI accelerators, this work proposes an optical multiplier based on inverse-designed optical microcavities, enabling direct optical multiplication of two numbers within the [-1, 1] range. The method employs dual-wavelength optical encoding of inputs and leverages intracavity interference—specifically constructive and destructive interference—to map multiplication onto photocurrent intensity modulation; the output current exhibits a linear relationship with the product (R² = 0.88). Departing from conventional matrix-vector dot-product architectures, the design replaces multi-device arrays with a single cavity, reducing photonic core area by 88% and laser power consumption by 23.43%, while achieving a 0.88% energy reduction in DeiT model training. The key innovation lies in the first integrated fusion of differentiable inverse design, interference-based multiplication encoding, and a neural-network-inspired photonic computing architecture—establishing a new paradigm for energy-efficient on-chip photonic computing.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Machine Learning: Hardware-aware ML
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Economics, Online Markets and Human Computation: Economic aspects and information design of Web, GenAI, and cloud computing

Abstract

The work presents an inverse-designed optical cavity that can direct light from two sources such that if the sources were to represent any number in the range [-1,1] with magnitude encoded through the power emitted by the source and sign by switching the direction of source current, the photocurrent generated at the two output ports is proportional to the product of the two numbers. Let us say that the two sources encode x and y, which are two numbers $in$ [-1,1]. Multiplication is reduced to the form $(x+y)^2 - (x-y)^2 = 4xy propto xy$. The addition and subtraction operations of the numbers are supported by constructive and destructive interference, respectively. The work shows that replacing the DDOT dot product engine of the Lightening Transformer with the optical cavity proposed to calculate the dot product can lead to a reduction in the area occupied by the photonic core by 88 %, can reduce the power consumption by lasers by around 23.43 %, and bring down energy consumption while training DeiT models by 0.88 %. The cavities can generate photocurrents of the form $1.057 xy + 0.249$ with $R^2=0.88,$ thus showing a relationship of direct proportionality between the target product $xy$ and the output of the cavity in response to stimuli encoding $x$ and $y$.

Problem

• Design optical cavity for efficient dot product computation
• Reduce photonic core area and laser power consumption
• Achieve proportional photocurrent output for multiplication operations

Innovation

• Inverse-designed optical cavity for multiplication
• Uses interference for addition and subtraction
• Reduces photonic core area by 88%