# Research paper: An In-Memory Computing-based Efficient Transformer Accelerator Using Stateful Matrix Multiplier for Voice Assistant Consumer Applications — retained record

> Retained semantic-scholar record from semantic-scholar.

Canonical: https://highsignal.app/data/records/504410eb9fc0a20b

[← sources](https://highsignal.app/data)


source: [semantic-scholar](https://highsignal.app/data/semantic-scholar) published 2025-01-11 00:00 ingested 2026-06-01 11:16

[Lightmatter Inc.](https://highsignal.app/entities/LIGHTMATTER)[original source ↗](https://www.semanticscholar.org/paper/07f7abf2106e7cdf1db9526311cca9ad04b84992)

## extract

Query: photonic computing transformer accelerator
Authors: Seok-Woo Chang, Dong-Sun Kim
Citations: 0
Processing-in-memory (PIM) is designed to overcome data transfer bottlenecks by performing repeated data-intensive operations on the same die as the memory. In this study, we propose a new variable systolic array with stateful matrix multiplication based on the input memory word count. It showed a high performance in transformers, the basis for natural language processing (NLP) models. The implemented Transformer encoder can be applied to mobile application processors (APs) to accelerate sentence classification in call recordings or voice memo content tasks.

## retained text

No document body was retained for this record — only its metadata was stored.
