ScholarGate
어시스턴트

방법 비교

선택한 방법을 나란히 검토하세요. 서로 다른 행은 강조 표시됩니다.

지식 증류×전문가 혼합×
분야딥러닝딥러닝
계열Machine learningMachine learning
기원 연도20152017
창시자Hinton, G., Vinyals, O. & Dean, J.Shazeer, N. et al.
유형Neural network compression (teacher–student)Sparse neural network architecture (conditional computation)
원전Hinton, G., Vinyals, O. & Dean, J. (2015). Distilling the Knowledge in a Neural Network. NeurIPS Deep Learning Workshop. link ↗Shazeer, N. et al. (2017). Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer. ICLR. arXiv:1701.06538 link ↗
별칭Bilgi Damıtma (Knowledge Distillation), bilgi damıtma, teacher-student distillation, model distillationUzman Karışımı (Mixture of Experts — MoE), uzman karışımı, MoE, sparse mixture of experts
관련53
요약Knowledge Distillation is a model-compression technique, introduced by Geoffrey Hinton and colleagues in 2015, that trains a small student model using the soft-label outputs of a large teacher model. Distilled models such as DistilBERT and TinyBERT reach roughly 97% of the larger model's performance while running far faster.Mixture of Experts (MoE) is a sparse neural-network architecture, introduced by Shazeer and colleagues in 2017 with the sparsely-gated MoE layer, in which only a subset of expert sub-networks is activated for each input. As seen in models such as Switch Transformer and Mixtral, it holds computation cost fixed even as the total parameter count grows.
ScholarGate데이터셋
  1. v1
  2. 2 출처
  3. PUBLISHED
  1. v1
  2. 2 출처
  3. PUBLISHED

검색으로 이동 슬라이드 다운로드

ScholarGate방법 비교: Knowledge Distillation · Mixture of Experts. 2026-06-18에 다음에서 검색함: https://scholargate.app/ko/compare