DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Moon, Jaekyun | - |
dc.contributor.advisor | 문재균 | - |
dc.contributor.author | Hong, Seokchan | - |
dc.date.accessioned | 2022-04-27T19:31:14Z | - |
dc.date.available | 2022-04-27T19:31:14Z | - |
dc.date.issued | 2021 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=948995&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/295984 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 전기및전자공학부, 2021.2,[III, 23p :] | - |
dc.description.abstract | Deep Neural Network(DNN) has been widely used to build A.I. in various fields such as computer vision and natural language process and outperformed other classical techniques. However, its massive computation and memory requirement hinder the implementation of DNNs on edge devices that have low computing power and smaller memory. Quantization of DNN is a network compression technique that reduces the precision of data involved only in the inference, or in both the inference and training. The quantized inference has been studied by lots of researchers, but the quantized training has had little focus and nothing has dealt with the utilization of only 4 or fewer-bit for all data objects in the training. Here I propose a novel quantized training algorithm GSNet (Gradient Stabilizer Network), that utilizes only 4 or fewer-bit data using a 4-bit additional object `stabilizer'. The stabilizer securely accumulates 2-bit gradients which are ternarized using a novel gradient scaling method. GSNet was evaluated in the image classification task and showed competitive performance compared with other quantization techniques that utilize even higher bit-precision. | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | Deep learning▼aOn-device machine learninig▼aNetwork compression▼aQuantized DNN▼aLow-precision training | - |
dc.subject | 심층학습 | - |
dc.subject | 온디바이스 기계학습▼a신경망 압축▼a신경망 양자화▼a저정밀도 학습 | - |
dc.title | 4-bit quantized training of deep neural networks | - |
dc.title.alternative | 심층 신경망의 4-bit 양자화 학습 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :전기및전자공학부, | - |
dc.contributor.alternativeauthor | 홍석찬 | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.