The paper introduces the Neural Responding Machine (NRM), a neural network-based response generator designed for Short-Text Conversation (STC). NRM employs an encoder-decoder framework, where the encoder summarizes the input text into a latent representation, and the decoder generates responses based on this representation. The model is trained using a large dataset of one-round conversations from Weibo, a popular Chinese microblogging platform. Empirical results show that NRM can generate grammatically correct and contextually appropriate responses for over 75% of the input texts, outperforming both retrieval-based and statistical machine translation (SMT)-based models. The paper also discusses the limitations of traditional methods and highlights the advantages of NRM, particularly in handling the complex nature of natural language conversations. The main contributions of the paper are the proposal of NRM and its superior performance compared to existing methods.The paper introduces the Neural Responding Machine (NRM), a neural network-based response generator designed for Short-Text Conversation (STC). NRM employs an encoder-decoder framework, where the encoder summarizes the input text into a latent representation, and the decoder generates responses based on this representation. The model is trained using a large dataset of one-round conversations from Weibo, a popular Chinese microblogging platform. Empirical results show that NRM can generate grammatically correct and contextually appropriate responses for over 75% of the input texts, outperforming both retrieval-based and statistical machine translation (SMT)-based models. The paper also discusses the limitations of traditional methods and highlights the advantages of NRM, particularly in handling the complex nature of natural language conversations. The main contributions of the paper are the proposal of NRM and its superior performance compared to existing methods.