| 赤部 知也 | M, 2回目発表 | コンピューティング・アーキテクチャ | 中島 康彦, | 林 優一, | TRAN THI HONG(客員), | 張 任遠 | |
|
title: Speeding Up of CGRAs by Reshaping and Stochastic FMA
abstract: Matrix multiplication in Coarse-grained reconfigurable arrays (CGRAs) is normally performed by vertically aligning input data and pipelining fused multiply-add (FMA) operations to achieve high efficiency. The final stage of deep convolutional neural networks (DCNNs) is equivalent to matrix multiplication, and there, input channels (ICs) are large and output channels (OCs) are small. However, there is a limit to the height of the pipeline in the CGRA, and thus it is necessary to divide the ICs. This causes an increase in the number of CGRA reboots and the execution time. The number of reboots can be reduced by reshaping the ICs horizontally and repeating FMA operations on the same unit, but SIMD is hard to use with ALUs with two 32-bit floating-point FMA units, and the throughput slows down to 4FMAs/4cycles. Recent studies show that high computational accuracy is not necessary for inference of DCNNs, and stochastic computing has been attracting attention. Stochastic computing can reduce the size of data and circuit resources and increase the frequency instead of reducing accuracy, and also facilitates horizontal accumulation. In this study, we introduced a stochastic fused multiply-add (SFMA) unit in the CGRA, which achieves a high throughput of 32FMAs/4cycles. In inference for handwritten character recognition, this method shows 94% accuracy, reduces the circuit area by 39%, improves the frequency by 63%, and the number of CGRA reboots to 0.2%, and improves the speed of inference execution by about 46 times compared to the CGRA with floating-point FMA units. language of the presentation: Japanese | |||||||
| YANG ZHENGCHANG | M, 2回目発表 | サイバネティクス・リアリティ工学 | 清川 清, | 加藤 博一, | 酒田 信親(客員), | 内山 英昭, | 磯山 直也 |
| 窪田 太一 | M, 1回目発表 | サイバネティクス・リアリティ工学 | 清川 清, | 荒牧 英治, | 内山 英昭, | 磯山 直也 | |
|
title: Development of a Video Calling Tool to Engage Speakers with Each Other through Nodding Exaggeration
abstract: In face-to-face conversations, people create a sense of unity by mutually synchronizing not only verbal information, but also non-verbal information such as nodding and physical movements. However, it is said to be difficult to induce these phenomena in non-face-to-face communication such as video calls, where it is difficult to convey each other's physical movements. In this study, we aim to develop a system that can exaggerate the nodding of the speaker in a video call, and provide a trigger for the speaker to engage. In addition, we will examine whether this system can promote the phenomenon of attraction in non-face-to-face communication and create a conversation with a sense of unity. Translated with www.DeepL.com/Translator (free version) language of the presentation: Japanese 発表題目: 頷き誇張により話者を相互に引き込むビデオ通話ツールの開発 発表概要: 人は対面での会話において言葉によるバーバルな情報だけでなく、頷きや身体動作などのノンバーバルな情報を相互に同調させる「引き込み現象」を引き起こすことで一体感のあるコミュニケーションを構築している。しかし、これらの現象をビデオ通話など、互いの身体的な動作が伝わりづらい非対面コミュニケーションで引き起こすことは困難であるとされている。そこで、本研究ではビデオ通話において話者の頷きを誇張し、引き込みのきっかけを与えるシステムの実現を目指す。また、本システムを用いることで、非対面での引き込み現象を促進し、一体感のある会話を作ることができるかを検証する。 | |||||||
| 淺田 樹生 | M, 1回目発表 | インタラクティブメディア設計学 | 加藤 博一, | 清川 清, | 神原 誠之, | 藤本 雄一郎, | 澤邊 太志 |
|
title:Study of a display method to change driver's speed perception using AR
abstract:Accidents caused by excessive speed and traffic jams caused by slowing down are some of the problems that arise when driving. These problems are mainly caused by misperception of speed. Changing the perception of speed is one way to solve this problem. In the past, drivers were encouraged to change their speed perception by adding various things to the road environment.However, this method cannot take into account the differences in speed perception of each individual. Therefore, in this research, I will examine the patterns that promote changes in speed perception using AR and how to display them. In this presentation, I will focus on the related works that I investigated at the beginning of this research. language of the presentation:Japanese 発表題目:ARを用いた運転者の速度認知を変化させる表示方法の検討 発表概要:運転をする際に生じる問題として速度超過による事故や速度低下による渋滞があげられる。これは主に速度を誤認することによって発生する問題である。速度認知を変化させることはこの問題を解決する1つの方法である。従来は道路環境に様々なものを付加することにより、運転者の速度認知の変化を促していた。しかし、この手法では個々人の速度認知の差を考慮することができない。そこで本研究ではARを用いて速度認知の変化を促すパターンとその表示方法について検証する。本発表では、研究を始めるに際して調査した関連研究を中心に紹介する。 | |||||||