Generating Persona Consistent Dialogues by Exploiting Natural Language Inference
Authors: Haoyu Song, Wei-Nan Zhang, Jingwen Hu, Ting Liu8878-8885
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experimental results on both human and automatic metrics, including the model-based consistency evaluation, demonstrate that the proposed approach outperforms strong generative baselines, especially in the personaconsistency of generated responses. Experiments Datasets Persona-Chat We perform persona-based dialogue generation experiments on the Persona-Chat dataset (Zhang et al. 2018). |
| Researcher Affiliation | Academia | Haoyu Song, Wei-Nan Zhang, Jingwen Hu, Ting Liu Research Center for Social Computing and Information Retrieval Harbin Institute of Technology, Heilongjiang Province, China {hysong, wnzhang, jwhu, tliu}@ir.hit.edu.cn |
| Pseudocode | Yes | Algorithm 1 Sketch of the training procedure |
| Open Source Code | No | The paper does not contain an explicit statement about releasing its source code or a direct link to a code repository for the methodology described. |
| Open Datasets | Yes | Persona-Chat We perform persona-based dialogue generation experiments on the Persona-Chat dataset (Zhang et al. 2018). DNLI The recently released Dialogue Natural Language Inference dataset (Welleck et al. 2019) offers a new domain for NLI models. |
| Dataset Splits | Yes | The testing set contains 1,000 dialogues (15,119 utterances) and 200 never seen before personas. We set aside 968 dialogues (15,705 utterances) together with its personas from the training set for validation. The final data has 10,000/968/1,000 dialogues for train/validate/test2. This dataset has 310,110/16,500/16,500 pairs for train/validate/test. |
| Hardware Specification | No | The paper does not specify any hardware details (e.g., GPU models, CPU types, memory) used for running the experiments. |
| Software Dependencies | No | The paper states 'We implement the model in Open NMT-py.' However, it does not provide a specific version number for Open NMT-py or any other ancillary software. |
| Experiment Setup | Yes | For the generator, both encoder and decoder are two-layer GRU with a hidden size 500. Embeddings of size 300 are randomly initialized and updated during training. Vocabulary size is 18,300, and other tokens are replaced with the UNK token. Encoder and decoder share the same vocabularies and embeddings. The model parameters are optimized using Adam with an initial learning rate of 0.0003. Learning rate decay is 0.98. Training minibatch size is 32. We set λ to 0.4 and N to 5. |