Dialogue policy learning (DPL) is a crucial component of dialogue modelling. Its primary role is to determine the appropriate abstract response, commonly referred to as the "dialogue action". Traditional DPL methodologies have treated this as a sequential decision problem, using pre-defined action candidates extracted from a corpus. However, these incomplete candidates can significantly limit the diversity of responses and pose challenges when dealing with edge cases, which are scenarios that occur only at extreme operating parameters. To address these limitations, we introduce a novel framework, JoTR. This framework is unique as it leverages a text-to-text Transformer-based model to generate flexible dialogue actions. Unlike traditional methods, JoTR formulates a word-level policy that allows for a more dynamic and adaptable dialogue action generation, without the need for any action templates. This setting enhances the diversity of responses and improves the system's ability to handle edge cases effectively. In addition, JoTR employs reinforcement learning with a reward-shaping mechanism to efficiently finetune the word-level dialogue policy, which allows the model to learn from its interactions, improving its performance over time. We conducted an extensive evaluation of JoTR to assess its effectiveness. Our extensive evaluation shows that JoTR achieves state-of-the-art performance on two benchmark dialogue modelling tasks, as assessed by both user simulators and human evaluators.

Generating persona consistent dialogue response is important for developing an intelligent conversational agent. Recent works typically fine-tune large-scale pre-trained models on this task by concatenating persona texts and dialogue history as a single input sequence to generate the target response. While simple and effective, our analysis shows that this popular practice is seriously affected by order sensitivity where different input orders of persona sentences significantly impact the quality and consistency of generated response, resulting in severe performance fluctuations (i.e., 29.4% on GPT2 and 83.2% on BART). To mitigate the order sensitivity problem, we propose a model-agnostic framework, ORder Insensitive Generation (ORIG), which enables dialogue models to learn robust representation under different persona orders and improve the consistency of response generation. Experiments on the Persona-Chat dataset justify the effectiveness and superiority of our method with two dominant pre-trained models (GPT2 and BART).

Meta-learning is widely used for few-shot slot tagging in the task of few-shot learning. The performance of existing methods is, however, seriously affected by catastrophic forgetting. This phenomenon is common in deep learning as the training and testing modules fail to take into account historical information, i.e. previously trained episodes in the metric-based meta-learning. To overcome this predicament, we propose the Memory-based Contrastive Meta-learning (MCML) method. Specifically, we propose a learn-from-memory mechanism that use explicit memory to keep track of the label representations of previously trained episodes and propose a contrastive learning method to compare the current label embedded in the few shot episode with the historic ones stored in the memory, and an adaption-from memory mechanism to determine the output label based on the contrast between the input labels embedded in the test episode and the label clusters in the memory. Experimental results show that MCML is scalable and outperforms metric-based meta-learning and optimization-based meta-learning on all 1shot, 5-shot, 10-shot, and 20-shot scenarios of the SNIPS dataset.