상세정보
학술대회
Selective Inference for Accelerating Deep Learning-based Image Classification
- 발행일
- 201610
- 출처
- International Conference on Information and Communication Technology Convergence (ICTC) 2016, pp.135-137
- 협약과제
- 16ZI2200, 지역광부품 고도화를 위한 광융합 기술개발, 이병탁
- 초록
- Reducing the computational budget of inference in deep neural network while achieving high accuracy is important for time-sensitive applications. In this paper, unlike other approaches that try to compress a large neural network to a neural network with a smaller number of parameters, we try to complete the image classification as early as possible. Adding a middle output layer, we try to complete the image classification at the middle output layer when Top1 confidence exceeds a predefined confidence threshold. We prove the feasibility of proposed approach based on experiment using Inception-v3 in TensorFlow.
- KSP 제안 키워드
- Confidence threshold, Deep neural network(DNN), High accuracy, Image classification, Inception-v3, Learning-based, Output layer, Selective inference, Time-sensitive, deep learning(DL)