Medical online consultation is important to healthcare worldwide, with hundreds of millions of participants each year. However, expert-level online consultations are expensive due to the shortage of medical professionals, while AI models are unreliable because they have unpredictable risks. Therefore, we introduce human-machine collaboration to medical online consultation and focus on symptom inquiry, as the basis for disease diagnosis. There are two key issues: 1) how to design an intelligent assignment strategy that can determine whether doctors or models participate in each turn? 2) how to design an effective execution strategy that can improve the machine's inquiry ability among considerable symptoms? To address the above issues, we propose the Human-AI Diagnostic Team (HADT) framework based on Hierarchical Reinforcement Learning (HRL), which aims to achieve high accuracy with low manpower. Specifically, HADT has two layers. The upper one is responsible for assignment, in which we propose a module called master that enables intelligent human-machine assignments through the masked RL with reward shaping. The lower one is responsible for execution, consisting of a doctor and a proposed module called machine. This module can effectively ask about symptoms through the masked HRL with bottom-up training. Experiments on the public datasets show that HADT can achieve up to 89.4% accuracy with only 10.9% human effort, as confirmed by real clinical doctors using our online interface.

Knowledge-enhanced dialogue systems aim at generating factually correct and coherent responses by reasoning over knowledge sources, which is a promising research trend. The truly harmonious human-agent dialogue systems need to conduct engaging conversations from three aspects as humans, namely (1) stating factual contents (e.g., records in Wikipedia), (2) conveying subjective and informative opinions about objects (e.g., user discussions on Twitter), and (3) impressing interlocutors with diverse expression styles (e.g., personalized expression habits). The existing knowledge base is a standardized and unified coding for factual knowledge, which could not portray the other two kinds of knowledge to make responses more informative and expressive diverse. To address this, we present CrowdDialog , a crowd intelligence knowledge-enhanced dialogue system, which takes advantage of “crowd intelligence knowledge” extracted from social media (with rich subjective descriptions and diversified expression styles) to promote the performance of dialogue systems. Firstly, to thoroughly mine and organize the crowd intelligence knowledge underlying large-scale and unstructured online contents, we elaborately design the C rowd I ntelligence K nowledge G raph ( CIKG ) structure, including the domain commonsense subgraph, descriptive subgraph, and expressive subgraph. Secondly, to reasonably integrate heterogeneous crowd intelligence knowledge into responses while ensuring logicality and fluency, we propose the G ated F usion with D ynamic Knowledge- D ependent ( GFDD ) model, which generates responses from the semantic and syntactic perspective with the context-aware knowledge gate and dynamic knowledge decoding. Finally, extensive experiments over both Chinese and English dialogue datasets demonstrate that our approach GFDD outperforms competitive baselines in terms of both automatic evaluation and human judgments. Besides, ablation studies indicate that the proposed CIKG has the potential to promote dialogue systems to generate fluent, informative, and diverse dialogue responses.

Recently,fake newshas been readily spread by massive amounts of users in social media, and automatic fake news detection has become necessary. The existing works need to prepare the overall data to perform detection, losing important information about the dynamic evolution of crowd opinions, and usually neglect the issue of uneven arrival of data in the real world. To address these issues, in this article, we focus on a kind of approach for fake news detection, namelyprogressive detection, which can be achieved by thedynamic Probabilistic Graphical Model. Based on the observation on real-world datasets, we adaptively improve the Kalman Filter to theLabeled Variable Dimension Kalman Filter(LVDKF) that learns two universal patterns from true and fake news, respectively, which can capture the temporal information of time-series data that arrive unevenly. It can take sequential data as input, distill the dynamic evolution knowledge regarding a post, and utilize crowd wisdom from users’ responses to achieve progressive detection. Then we derive the formulas using the Forward, Backward, and EM Algorithm, and we design a dynamic detection algorithm using Bayes’ theorem. Finally, we design experimental scenarios simulating progressive detection and evaluate LVDKF on two public datasets. It outperforms the baseline methods in these experimental scenarios, which indicates that it is adequate for progressive detection.

The prediction of express delivery sequence, i.e., modeling and estimating the volumes of daily incoming and outgoing parcels for delivery, is critical for online business, logistics, and positive customer experience, and specifically for resource allocation optimization and promotional activity arrangement. A precise estimate of consumer delivery requests has to involve sequential factors such as shopping behaviors, weather conditions, events, business campaigns, and their couplings. Despite that various methods have integrated external features to enhance the effects, extant works fail to address complex feature-sequence couplings in the following aspects: weaken the inter-dependencies when processing heterogeneous data and ignore the cumulative and evolving situation of coupling relationships. To address these issues, we propose DeepExpress—a deep-learning-based express delivery sequence prediction model, which extends the classic seq2seq framework to learn feature-sequence couplings. DeepExpress leverages an express delivery seq2seq learning, a carefully designed heterogeneous feature representation, and a novel joint training attention mechanism to adaptively handle heterogeneity issues and capture feature-sequence couplings for accurate prediction. Experimental results on real-world data demonstrate that the proposed method outperforms both shallow and deep baseline models.