Establishing shared goals is a fundamental step in human-AI communication.However, ambiguities can lead to outputs that seem correct but fail to reflect the speaker's intent.In this paper, we explore this issue with a focus on the data visualization domain, where ambiguities in natural language impact the generation of code that visualizes data.The availability of multiple views on the contextual (e.g. the intended plot and the code rendering the plot) allows for a unique and comprehensive analysis of diverse ambiguity types.We develop a taxonomy of types of ambiguity that arise in this task and propose metrics to quantify them.Using Matplotlib problems from the DS-1000 dataset, we demonstrate that our ambiguity metrics better correlate with human annotations than uncertainty baselines.Our work also explores how multi-turn dialogue can reduce ambiguity, and therefore, improve code accuracy by better matching user goals.We evaluate three pragmatic models to inform our dialogue strategies: Gricean Cooperativity, Discourse Representation Theory, and Questions under Discussion.A simulated user study reveals how pragmatic dialogues reduce ambiguity and enhance code accuracy, highlighting the value of multi-turn exchanges in code generation.

Recent advances in deep neural networks (DNNs) have greatly benefited mobile applications that perform real-time video analytics. However, mobile devices' computing power usually limits them from inferring complex DNN models timely. Edge intelligence has emerged to help mobile apps offload DNN inference tasks to powerful edge servers for accelerated inference services. One major challenge that edge intelligence faces is maintaining a satisfactory quality of service when users move across edge servers. To address this issue, we propose a novel solution to help improve the quality of inference services for real-time video analytics applications that use chained DNN models. This solution includes two schemes: one maximizes the use of mobile devices to improve inference quality during the handover between edge servers, and the other provides offloading decisions to minimize the end-to-end inference delay when edge servers are available. We evaluate the proposed scheme using a DNN-based realtime traffic monitoring application through testbed and simulation experiments. The results show that our solution can improve inference quality by 52% during handover compared to the greedy algorithm-based solution.