CPM-2: Large-scale cost-effective pre-trained language models.
Zhengyan ZhangYuxian GuXu HanShengqi ChenChaojun XiaoZhenbo SunYuan YaoFanchao QiJian GuanPei KeYanzheng CaiGuoyang ZengZhixing TanZhiyuan LiuMinlie HuangWentao HanYang LiuXiaoyan ZhuMaosong SunPublished in: AI Open (2021)
Keyphrases
- cost effective
- language model
- pre trained
- language modeling
- n gram
- document retrieval
- speech recognition
- probabilistic model
- information retrieval
- low cost
- retrieval model
- cost effectiveness
- language modelling
- training data
- query expansion
- statistical language models
- language models for information retrieval
- smoothing methods
- test collection
- document ranking
- relevance model
- training examples
- machine learning
- neural network
- multi modal
- learning algorithm