Clarifying user needs is essential for existing task-oriented dialogue systems.However, in real-world applications, developers can never guarantee that all possible user demands are taken into account in the design phase.Consequently, existing systems will break down when encountering unconsidered user needs.To address this problem, we propose a novel incremental learning framework to design task-oriented dialogue systems, or for short Incremental Dialogue System (IDS), without pre-defining the exhaustive list of user needs.Specifically, we introduce an uncertainty estimation module to evaluate the confidence of giving correct responses.If there is high confidence, IDS will provide responses to users.Otherwise, humans will be involved in the dialogue process, and IDS can learn from human intervention through an online learning module.To evaluate our method, we propose a new dataset which simulates unanticipated user needs in the deployment stage.Experiments show that IDS is robust to unconsidered user actions, and can update itself online by smartly selecting only the most effective training data, and hence attains better performance with less annotation cost. 1

Weikang Wang, Jiajun Zhang, Qian Li, Chengqing Zong, Zhifei Li. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019.

Reinforcement learning (RL) is an attractive solution for task-oriented dialog systems.However, extending RL-based systems to handle new intents and slots requires a system redesign.The high maintenance cost makes it difficult to apply RL methods to practical systems on a large scale.To address this issue, we propose a practical teacherstudent framework to extend RL-based dialog systems without retraining from scratch.Specifically, the "student" is an extended dialog manager based on a new ontology, and the "teacher" is existing resources used for guiding the learning process of the "student".By specifying constraints held in the new dialog manager, we transfer knowledge of the "teacher" to the "student" without additional resources.Experiments show that the performance of the extended system is comparable to the system trained from scratch.More importantly, the proposed framework makes no assumption about the unsupported intents and slots, which makes it possible to improve RL-based systems incrementally.U: I'm looking for a Sichuan restaurant.S: "Spicy Little Girl"

Achieving MAC layer fairness in wireless ad-hoc networks is a very challenging issue. We first show that IEEE 802.11 exhibits substantial short-term unfairness due to the randomness in the binary exponential back-off (BEB), the freezing mechanism of the back-off timer, and the concealed information problem (introduced in this paper). To achieve short-term fairness, we propose a novel fairness scheme, called adaptive transmission control (ATC). The basic idea of ATC is to make the senders to contend for the shared medium in a cooperative and adaptive manner by exploiting the overheard information. The ATC includes three mechanisms: CW-tuning, early-reset, and receiver-coordination, which are able to tackle the underlying problems that cause short-term unfairness in IEEE 802.11. Specifically, the CW-tuning mechanism reflects the past usage of the medium among the nodes, and thus greatly reduces the randomness in BEB, while the early-reset mechanism largely prevents the short-term unfairness caused by the freezing mechanism. At last, the receiver-coordination mechanism copes effectively with the concealed information problem by using information available at the receiver side. Extensive simulation results show that the proposed ATC substantially improves the short-term fairness without unduly degrading the throughput.