Considering the inherent limitations of parametric knowledge in large language models (LLMs), retrieval-augmented generation (RAG) is widely employed to expand their knowledge scope. Since RAG has shown promise in knowledge-intensive tasks like open-domain question answering, its broader application to complex tasks and intelligent assistants has further advanced its utility. Despite this progress, the underlying knowledge utilization mechanisms of LLM-based RAG remain underexplored. In this paper, we present a systematic investigation of the intrinsic mechanisms by which LLMs integrate internal (parametric) and external (retrieved) knowledge in RAG scenarios. Specially, we employ knowledge stream analysis at the macroscopic level, and investigate the function of individual modules at the microscopic level. Drawing on knowledge streaming analyses, we decompose the knowledge utilization process into four distinct stages within LLM layers: knowledge refinement, knowledge elicitation, knowledge expression, and knowledge contestation. We further demonstrate that the relevance of passages guides the streaming of knowledge through these stages. At the module level, we introduce a new method, knowledge activation probability entropy (KAPE) for neuron identification associated with either internal or external knowledge. By selectively deactivating these neurons, we achieve targeted shifts in the LLM's reliance on one knowledge source over the other. Moreover, we discern complementary roles for multi-head attention and multi-layer perceptron layers during knowledge formation. These insights offer a foundation for improving interpretability and reliability in retrieval-augmented LLMs, paving the way for more robust and transparent generative solutions in knowledge-intensive domains.

Conversational recommendation casts the recommendation problem as a dialog-based interactive task, which could acquire user interest more efficiently and effectively by allowing users to express what they like. In this work, we move a step towards a new conversational recommendation task that is more suitable for real-world applications. In this task, the recommender proactively and naturally lead a dialog from non-recommendation content to approach an item being of interest to users, and allow users to ask questions for better support of user decisions. The challenge of this task lies in how to effectively control the dialog flow to complete the recommendation while appropriately responding to user utterances. To address this challenge, we first construct a Chinese recommendation dialog dataset DuRecDial. We then propose a two-stage Multi-Goal driven Conversation Generation framework, MGCG. In particular, the goal planning module leverages the global graph structure information and local goal-sequence information to effectively control the dialog flow step by step. The goal-guided responding module can produce an in-depth dialog about each goal by fully exploiting hierarchical goal information for response retrieval or generation. Results on DuRecDial demonstrate that MGCG can lead the dialog more proactively and naturally, and complete the recommendation task more effectively.

With the increased popularity of mobile devices, Web mapping services have become an indispensable tool in our daily lives. To provide user-satisfied services, such as location searches, the point of interest (POI) database is the fundamental infrastructure, as it archives multimodal information on billions of geographic locations closely related to people's lives, such as a shop or a bank. Therefore, verifying the correctness of a large-scale POI database is vital. To achieve this goal, many industrial companies adopt volunteered geographic information (VGI) platforms that enable thousands of crowdworkers and expert mappers to verify POIs seamlessly; but to do so, they have to spend millions of dollars every year. To save the tremendous labor costs, we devised DuMapper, an automatic system for large-scale POI verification with the multimodal street-view data at Baidu Maps. This paper presents not only DuMapper I, which imitates the process of POI verification conducted by expert mappers, but also proposes DuMapper II, a highly efficient framework to accelerate POI verification by means of deep multimodal embedding and approximate nearest neighbor (ANN) search. DuMapper II takes the signboard image and the coordinates of a real-world place as input to generate a low-dimensional vector, which can be leveraged by ANN algorithms to conduct a more accurate search through billions of archived POIs in the database for verification within milliseconds. Compared with DuMapper I, experimental results demonstrate that DuMapper II can significantly increase the throughput of POI verification by 50 times. DuMapper has already been deployed in production since June 2018, which dramatically improves the productivity and efficiency of POI verification at Baidu Maps. As of December 31, 2021, it has enacted over 405 million iterations of POI verification within a 3.5-year period, representing an approximate workload of 800 high-performance expert mappers.

Estimated time of arrival (ETA) prediction, also known as travel time estimation, is a fundamental task for a wide range of intelligent transportation applications, such as navigation, route planning, and ride-hailing services. To accurately predict the travel time of a route, it is essential to take into account both contextual and predictive factors, such as spatial-temporal interaction, driving behavior, and traffic congestion propagation inference. The ETA prediction models previously deployed at Baidu Maps have addressed the factors of spatial-temporal interaction (ConSTGAT) and driving behavior (SSML). In this work, we believe that modeling traffic congestion propagation patterns is of great importance toward accurately performing ETA prediction, and we focus on this factor to improve ETA performance. Traffic congestion propagation pattern modeling is challenging, and it requires accounting for impact regions over time and cumulative effect of delay variations over time caused by traffic events on the road network. In this paper, we present a practical industrial-grade ETA prediction framework named DuETA. Specifically, we construct a congestion-sensitive graph based on the correlations of traffic patterns, and we develop a route-aware graph transformer to directly learn the long-distance correlations of the road segments. This design enables DuETA to capture the interactions between the road segment pairs that are spatially distant but highly correlated with traffic conditions. Extensive experiments are conducted on large-scale, real-world datasets collected from Baidu Maps. Experimental results show that ETA prediction can significantly benefit from the learned traffic congestion propagation patterns, which demonstrates the effectiveness and practical applicability of DuETA. In addition, DuETA has already been deployed in production at Baidu Maps, serving billions of requests every day. This demonstrates that DuETA is an industrial-grade and robust solution for large-scale ETA prediction services.

The task of POI attribute acquisition, which aims at completing missing attributes (e.g., POI name, address, status, phone, and open/close time) for a point of interest (POI) or updating existing attribute values of a POI, plays an essential role in enabling users to entertain location-based services using commercial map applications, such as Baidu Maps. Existing solutions have adopted street views or web documents to acquire POI attributes, which have a major limitation in applying for large-scale production due to the labor-intensive and time-consuming nature of collecting data, error accumulation in processing textual/visual data in unstructured or free format, and necessitating post-processing steps with manual efforts. In this paper, we present our efforts and findings from a 3-year longitudinal study on designing and implementing DuIVRS, which is an alternative, fully automatic, and production-proven solution for large-scale POI attribute acquisition via completely machine-directed dialogues. Specifically, DuIVRS is designed to proactively acquire POI attributes via a telephonic interactive voice response system, whose tasks are to generate machine-initiative directed dialogues, make scripted telephone calls to businesses, and interact with people who answered the phone to achieve predefined goals through multi-turn dialogues. DuIVRS has already been deployed in production at Baidu Maps since December 2018, which greatly improves productivity and reduces production cost of POI attribute acquisition. As of December 31, 2021, DuIVRS has made 140 million calls and 42 million POI attribute updates within a 3-year period, which represents an approximately 3-year workload for a high-performance team of 1,000 call center workers. This demonstrates that DuIVRS is an industrial-grade and robust solution for cost-effective, large-scale acquisition of POI attributes.

As the core foundation of web mapping, each geographic object (geo-object), such as a traffic sign, plays a vital role in navigation and intelligent driving. Determining how to obtain the latest high-precision geo-object information is a classic topic in updating road databases. Benefiting from the cost-effective attribute and availability of the positioning equipment and camera, the vision-based update pattern is becoming increasingly popular in the industry. Generally speaking, the road database update mainly includes three phases: geo-object recognition, localization, and change detection. Previous change detection strategies are mainly performed by comparing the historical road information (i.e., geo-object type and position) with the new geographic data of geo-objects collected from the street-view imagery. However, limited by the localization precision of the positioning equipment and the discriminative power of the vanilla differential-based method, the accuracy, recall, and efficiency of previous systems for geo-object change detection are greatly impaired. In addition, the artificially prescribed production standards make the geo-object position in the map data deviate from its position in the real world, as well as some geo-objects do not need to be updated (e.g., temporary speed limit), which further yields many false-positive detections and significantly increases the labor costs of existing systems. To address these challenges, we propose a novel framework called DuARUS for automatic geo-object change detection with street-view imagery. In this paper, we mainly focus on automatic geo-object localization and change detection. Specifically, for geo-object localization, we propose a two-stage, integrated localization algorithm based on image matching and monocular depth estimation. Furthermore, to achieve automatic change detection, vision-based representation learning and scene understanding strategies are introduced to build a large-scale geo-object semantic map, which can provide sufficient multimodal information support for change detection. Based on such artful modeling, we recast the complicated, labor-based change detection problem as a vanilla binary classification task, which is a robust and efficient strategy that contributes to resolving this problem. By combining these operations, we construct an industrial-grade, fully automatic production system for road database updates. Extensive experiments conducted on large-scale, real-world datasets from Baidu Maps demonstrate the superiority and effectiveness of the system. Moreover, this system has already been deployed in production at Baidu Maps since July 2020, handling 96% of automatic road database updates. DuARUS improves the annual update mileage from millions to tens of millions, and it achieves weekly updates.

The task of live traffic condition prediction, which aims at predicting live traffic conditions (i.e., fast, slow, and congested) based on traffic information on roads, plays a vital role in intelligent transportation systems, such as navigation, route planning, and ride-hailing services. Existing solutions have adopted aggregated trajectory data to generate traffic estimates, which inevitably suffer from GPS drift caused by cluttered urban road scenarios. In addition, the trajectory information alone is insufficient to provide evidence for sudden traffic situations and perception of street-wise elements. To alleviate these problems, in this paper, we present DuTraffic, which is a robust and production-ready solution for live traffic condition prediction by taking both trajectory data and street views into account. Specifically, the vision-based detection and segmentation modules are developed to forecast traffic flow by using street views. Then, we propose a spatial-temporal-based module, TRST-Net, to learn the latent trajectory representation. Finally, a bilinear model is introduced to mix these two representations and then predicts live traffic conditions with trajectory data and street views in a mutually complementary manner. The task is recast as a multi-task learning problem, which could benefit from the strong representation of latent space manifold modeling. Extensive experiments conducted on large-scale, real-world datasets from Baidu Maps demonstrate the superiority and effectiveness of DuTraffic. In addition, DuTraffic has already been deployed in production at Baidu Maps since December 2020, handling tens of millions of requests every day. This demonstrates that DuTraffic is a practical and robust industrial solution for live traffic condition prediction.

Pre-trained models (PTMs) have become a fundamental backbone for downstream tasks in natural language processing and computer vision. Despite initial gains that were obtained by applying generic PTMs to geo-related tasks at Baidu Maps, a clear performance plateau over time was observed. One of the main reasons for this plateau is the lack of readily available geographic knowledge in generic PTMs. To address this problem, in this paper, we present ERNIE-GeoL, which is a geography-and-language pre-trained model designed and developed for improving the geo-related tasks at Baidu Maps. ERNIE-GeoL is elaborately designed to learn a universal representation of geography-language by pre-training on large-scale data generated from a heterogeneous graph that contains abundant geographic knowledge. Extensive quantitative and qualitative experiments conducted on large-scale real-world datasets demonstrate the superiority and effectiveness of ERNIE-GeoL. ERNIE-GeoL has already been deployed in production at Baidu Maps since April 2021, which significantly benefits the performance of various downstream tasks. This demonstrates that ERNIE-GeoL can serve as a fundamental backbone for a wide range of geo-related tasks.

Mobile map apps such as the Baidu Maps app have become a ubiquitous and essential tool for users to find optimal routes and get turn-by-turn navigation services while driving. However, interacting with such apps while driving through visual-manual interaction modality inevitably causes driver distraction, due to the highly conspicuous nature of the time-sharing, multi-tasking behavior of the driver. In this paper, we present our efforts and findings of a 4-year longitudinal study on designing and implementing DuIVA, which is an intelligent voice assistant (IVA) embedded in the Baidu Maps app for hands-free, eyes-free human-to-app interaction in a fully voice-controlled manner. Specifically, DuIVA is designed to enable users to control the functionalities of Baidu Maps (e.g., navigation and location search) through voice interaction, rather than visual-manual interaction, which minimizes driver distraction and promotes safe driving by allowing the driver to keep "eyes on the road and hands on the wheel'' while interacting with the Baidu Maps app. DuIVA has already been deployed in production at Baidu Maps since November 2017, which facilitates a better interaction modality with the Baidu Maps app and improves the accessibility and usability of the app by providing users with in-app voice activation, natural language queries, and multi-round dialogue. As of December 31, 2021, over 530 million users have used DuIVA, which demonstrates that DuIVA is an industrial-grade and production-proven solution for in-app intelligent voice assistants.

Providing timely accessibility reminders (such as closed and relocated) of a point-of-interest (POI) plays a vital role in improving user satisfaction of finding places and making visiting decisions. However, it is difficult to keep the POI database in sync with the real-world counterparts due to the dynamic nature of business changes and innovations. To alleviate this problem, we formulate and present a practical solution that jointly extracts POI mentions and identifies their coupled accessibility labels from unstructured text (hereafter referred to as joint POI and accessibility extraction). We approach this task as a sequence tagging problem, where the goal is to produce (POI name, accessibility label) pairs from unstructured text. This task is challenging because of two main issues: (1) POI names are often newly-coined words so as to successfully register new entities or brands and (2) there may exist multiple pairs in the text, which necessitates dealing with one-to-many or many-to-one mapping to make each POI coupled with its matching accessibility label. To this end, we propose a Geographic-Enhanced and Dependency-guIded sequence Tagging (GEDIT) model to concurrently address the two challenges. First, to alleviate challenge #1, we develop a geographic-enhanced pre-trained model to learn the text representations, which is able to significantly relieve the problem of newly-coined words. Second, to mitigate challenge #2, we apply a relational graph convolutional network to learn the tree node representations from the parsed dependency tree, which enables us to establish a correlation between a POI and its accessibility label. Finally, we construct a neural sequence tagging model by integrating and feeding the previously pre-learned representations into a CRF layer. Extensive experiments conducted on a real-world dataset demonstrate the superiority and effectiveness of GEDIT. In addition, it has already been deployed in production at Baidu Maps, and it successfully keeps processing hundreds of thousands of Web documents every week. Statistics show that the proposed solution can save significant human effort and labor costs to deal with the same amount of documents, which confirms that it is a practical way for POI accessibility maintenance.

Point Of Interest Auto-Completion (abbr. as POI-AC) is one of the featured functions for the search engine at Baidu Maps. It can dynamically suggest a list of POI candidates within milliseconds as a user enters each character (e.g., English, Chinese, or Pinyin character) into the search box. Ideally, a user may need to provide only one character and immediately obtain the desired POI at the top of the POI list suggested by POI-AC. In this way, the user's keystrokes can be dramatically saved, which significantly reduces the time and effort of typing, especially on mobile devices that have limited space for display and user interfaces. Despite using a user's profile and input prefixes for personalized POI suggestions, however, the state-of-the-art approach, i.e., P^3AC, still has a long way to go so as to generate not only personalized but, more importantly, time- and geography-aware suggestions. In this paper, we find that 17.9% of users tend to look for diverse POIs at different times or locations using the same prefix. This insight drives us to establish an end-to-end spatial-temporal POI-AC (abbr. as ST-PAC) module to replace P3AC at Baidu Maps. To alleviate the problem of the long-tail distribution of time- and location-specific data on POI-AC, we further propose a meta-learned ST-PAC (abbr. as MST-PAC) updated by an efficient MapReduce algorithm. MST-PAC can significantly overcome the "long-tail" issue and rapidly adapt to the cold-start POI-AC tasks with fewer examples. We sample several benchmark datasets from the large-scale search logs at Baidu Maps to assess the offline performance of MST-PAC in line with multiple metrics, including Mean Reciprocal Rank (MRR), Success Rate (SR) and normalized Discounted Cumulative Gain (nDCG). The consistent improvements on these metrics give us more confidence to launch this meta-learned POI-AC module online. As a result, the critical indicator on user satisfaction online, i.e., the average number of keystrokes in a POI-AC session, significantly decreases as well. For now, MST-PAC has already been deployed in production at Baidu Maps, handling billions of POI-AC requests every day. It confirms that MST-PAC is a practical and robust industrial solution for large-scale POI Search.

Travel time estimation (TTE) is one of the most critical modules at Baidu Maps, which plays a vital role in intelligent transportation services such as route planning and navigation. During the driving en route, the navigation system of Baidu Maps can provide real-time estimations on when a user will arrive at the destination. It automatically recalculates and updates the remaining travel time from the driver's current position to the destination (hereafter referred to as remaining route) every few minutes. The previously deployed TTE model at Baidu Maps, i.e., ConSTGAT, takes the remaining route as well as the current time as input and provides the corresponding estimated time of arrival. However, it ignores the route that has been already traveled from the origin to the driver's current position (hereafter referred to as traveled route), which could contribute to improving the accuracy of time estimation. In this work, we believe that the traveled route conveys valuable evidence that could facilitate the modeling of driving preference and take that into consideration for the task of en route travel time estimation (ER-TTE). This task is non-trivial because it requires adapting fast to a user's driving preference using a few observed behaviors in the traveled route. To this end, we frame ER-TTE as a few-shot learning problem and consider the observed behaviors in the traveled route as training examples while the future behaviors in the remaining route as test examples. To tackle the few-shot learning problem, we propose a novel model-based meta-learning approach, called SSML, to learn the meta-knowledge so as to fast adapt to a user's driving preference and improve the time estimation of the remaining route. SSML leverages the technique of self-supervised learning, which is equivalent to generating a significant number of synthetic learning tasks, to further improve the performance. Extensive offline tests conducted on large-scale real-world datasets collected from Baidu Maps demonstrate the superiority of SSML. The online tests before deploying in production were successfully performed, which confirms the practical applicability of SSML.

The increasing interest in international travel has raised the demand of retrieving point of interests (POIs) in multiple languages. This is even superior to find local venues such as restaurants and scenic spots in unfamiliar languages when traveling abroad. Multilingual POI retrieval, enabling users to find desired POIs in a demanded language using queries in numerous languages, has become an indispensable feature of today's global map applications such as Baidu Maps. This task is non-trivial because of two key challenges: (1) visiting sparsity and (2) multilingual query-POI matching. To this end, we propose a Heterogeneous Graph Attention Matching Network (HGAMN) to concurrently address both challenges. Specifically, we construct a heterogeneous graph that contains two types of nodes: POI node and query node using the search logs of Baidu Maps. First, to alleviate challenge #1, we construct edges between different POI nodes to link the low-frequency POIs with the high-frequency ones, which enables the transfer of knowledge from the latter to the former. Second, to mitigate challenge #2, we construct edges between POI and query nodes based on the co-occurrences between queries and POIs, where queries in different languages and formulations can be aggregated for individual POIs. Moreover, we develop an attention-based network to jointly learn node representations of the heterogeneous graph and further design a cross-attention module to fuse the representations of both types of nodes for query-POI relevance scoring. In this way, the relevance ranking between multilingual queries and POIs with different popularity can be better handled. Extensive experiments conducted on large-scale real-world datasets from Baidu Maps demonstrate the superiority and effectiveness of HGAMN. In addition, HGAMN has already been deployed in production at Baidu Maps, and it successfully keeps serving hundreds of millions of requests every day. Compared with the previously deployed model, HGAMN achieves significant performance improvement, which confirms that HGAMN is a practical and robust solution for large-scale real-world multilingual POI retrieval service.

As the recent COVID-19 outbreak rapidly expands all over the world, various containment measures have been carried out to fight against the COVID-19 pandemic. In Mainland China, the containment measures consist of three types, i.e., Wuhan travel ban, intra-city quarantine and isolation, and intercity travel restriction. In order to carry out the measures, local economy and information acquisition play an important role. In this paper, we investigate the correlation of local economy and the information acquisition on the execution of containment measures to fight against the COVID-19 pandemic in Mainland China. First, we use a parsimonious model, i.e., SIR-X model to estimate the parameters, which represent the execution of intra-city quarantine and isolation in major cities of Mainland China. In order to understand the execution of intra-city quarantine and isolation, we analyze the correlation between the representative parameters including local economy, mobility, and information acquisition. To this end, we collect the data of Gross Domestic Product (GDP), the inflows from Wuhan and outflows, and the COVID-19 related search frequency from a widely-used Web mapping service, i.e., Baidu Maps, and Web search engine, i.e., Baidu Search Engine, in Mainland China. Based on the analysis, we confirm the strong correlation between the local economy and the execution of information acquisition in major cities of Mainland China. We further evidence that, although the cities with high GDP per capita attract more inflows from Wuhan, people are more likely to conduct the quarantine measure and to reduce travelling to other cities. Finally, the correlation analysis using search data shows that well-informed individuals are likely to carry out containment measures.

The task of travel time estimation (TTE), which estimates the travel time for a given route and departure time, plays an important role in intelligent transportation systems such as navigation, route planning, and ride-hailing services. This task is challenging because of many essential aspects, such as traffic prediction and contextual information. First, the accuracy of traffic prediction is strongly correlated with the traffic speed of the road segments in a route. Existing work mainly adopts spatial-temporal graph neural networks to improve the accuracy of traffic prediction, where spatial and temporal information is used separately. However, one drawback is that the spatial and temporal correlations are not fully exploited to obtain better accuracy. Second, contextual information of a route, i.e., the connections of adjacent road segments in the route, is an essential factor that impacts the driving speed. Previous work mainly uses sequential encoding models to address this issue. However, it is difficult to scale up sequential models to large-scale real-world services. In this paper, we propose an end-to-end neural framework named ConSTGAT, which integrates traffic prediction and contextual information to address these two problems. Specifically, we first propose a spatial-temporal graph neural network that adopts a novel graph attention mechanism, which is designed to fully exploit the joint relations of spatial and temporal information. Then, in order to efficiently take advantage of the contextual information, we design a computationally efficient model that applies convolutions over local windows to capture a route's contextual information and further employs multi-task learning to improve the performance. In this way, the travel time of each road segment can be computed in parallel and in advance. Extensive experiments conducted on large-scale real-world datasets demonstrate the superiority of ConSTGAT. In addition, ConSTGAT has already been deployed in production at Baidu Maps, and it successfully keeps serving tens of billions of requests every day. This confirms that ConSTGAT is a practical and robust solution for large-scale real-world TTE services.

Point of interest auto-completion (POI-AC) is a featured function in the search engine of many Web mapping services. This function keeps suggesting a dynamic list of POIs as a user types each character, and it can dramatically save the effort of typing, which is quite useful on mobile devices. Existing approaches on POI-AC for industrial use mainly adopt various learning to rank (LTR) models with handcrafted features and even historically clicked POIs are taken into account for personalization. However, these prior arts tend to reach performance bottlenecks as both heuristic features and search history of users cannot directly model personal input habits. In this paper, we present an end-to-end neural-based framework for POI-AC, which has been recently deployed in the search engine of Baidu Maps, one of the largest Web mapping applications with hundreds of millions monthly active users worldwide. In order to establish connections among users, their personal input habits, and correspondingly interested POIs, the proposed framework (abbr. P3AC) is composed of three components, i.e., a multi-layer Bi-LSTM network to adapt to personalized prefixes, a CNN-based network to model multi-sourced information on POIs, and a triplet ranking loss function to optimize both personalized prefix embeddings and distributed representations of POIs. We first use large-scale real-world search logs of Baidu Maps to assess the performance of P3AC offline measured by multiple metrics, including Mean Reciprocal Rank (MRR), Success Rate (SR), and normalized Discounted Cumulative Gain (nDCG). Extensive experimental results demonstrate that it can achieve substantial improvements. Then we decide to launch it online and observe that some other critical indicators on user satisfaction, such as the average number of keystrokes and the average typing speed at keystrokes in a POI-AC session, which significantly decrease as well. In addition, we have released both the source codes of P3AC and the experimental data to the public for reproducibility tests.

The constrained outbreak of COVID-19 in Mainland China has recently been regarded as a successful example of fighting this highly contagious virus. Both the short period (in about three months) of transmission and the sub-exponential increase of confirmed cases in Mainland China have proved that the Chinese authorities took effective epidemic prevention measures, such as case isolation, travel restrictions, closing recreational venues, and banning public gatherings. These measures can, of course, effectively control the spread of the COVID-19 pandemic. Meanwhile, they may dramatically change the human mobility patterns, such as the daily transportation-related behaviors of the public. To better understand the impact of COVID-19 on transportation-related behaviors and to provide more targeted anti-epidemic measures, we use the huge amount of human mobility data collected from Baidu Maps, a widely-used Web mapping service in China, to look into the detail reaction of the people there during the pandemic. To be specific, we conduct data-driven analysis on transportation-related behaviors during the pandemic from the perspectives of 1) means of transportation, 2) type of visited venues, 3) check-in time of venues, 4) preference on "origin-destination'' distance, and 5) "origin-transportation-destination'' patterns. For each topic, we also give our specific insights and policy-making suggestions. Given that the COVID-19 pandemic is still spreading in more than 200 overseas countries, infecting millions of people worldwide, the insights and suggestions provided here may help fight COVID-19.

Entity recommendation, providing users with an improved search experience by proactively recommending related entities to a given query, has become an indispensable feature of today’s Web search engine. Existing studies typically only consider the query issued at the current timestep while ignoring the in-session user search behavior (short-term search history) or historical user search behavior across all sessions (long-term search history) when generating entity recommendations. As a consequence, they may fail to recommend entities of interest relevant to a user’s actual information need. In this work, we believe that both short-term and long-term search history convey valuable evidence that could help understand the user’s search intent behind a query, and take both of them into consideration for entity recommendation. Furthermore, there has been little work on exploring whether the use of other companion tasks in Web search such as document ranking as auxiliary tasks could improve the performance of entity recommendation. To this end, we propose a multi-task learning framework with deep neural networks (DNNs) to jointly learn and optimize two companion tasks in Web search engines: entity recommendation and document ranking, which can be easily trained in an end-to-end manner. Specifically, we regard document ranking as an auxiliary task to improve the main task of entity recommendation, where the representations of queries, sessions, and users are shared across all tasks and optimized by the multi-task objective during training. We evaluate our approach using large-scale, real-world search logs of a widely-used commercial Web search engine. We also performed extensive ablation experiments over a number of facets of the proposed multi-task DNN model to figure out their relative importance. The experimental results show that both short-term and long-term search history can bring significant improvements in recommendation effectiveness, and the combination of both outperforms using either of them individually. In addition, the experiments show that the performance of both entity recommendation and document ranking can be significantly improved, which demonstrates the effectiveness of using multi-task learning to jointly optimize the two companion tasks in Web search.

The value assessment of private properties is an attractive but challenging task which is widely concerned by a majority of people around the world. A prolonged topic among us is "how much is my house worth?". To answer this question, most experienced agencies would like to price a property given the factors of its attributes as well as the demographics and the public facilities around it. However, no one knows the exact prices of these factors, especially the values of public facilities which may help assess private properties. In this paper, we introduce our newly launched project "Monopoly" (named after a classic board game) in which we propose a distributed approach for revaluing private properties by learning to price public facilities (such as hospitals, schools, and metros) with the large-scale urban data we have accumulated via Baidu Maps. To be specific, our method organizes many points of interest (POIs) into an undirected weighted graph and formulates multiple factors including the virtual prices of surrounding public facilities as adaptive variables to parallelly estimate the housing prices we know. Then the prices of both public facilities and private properties can be iteratively updated according to the loss of prediction until convergence. We have conducted extensive experiments with the large-scale urban data of several metropolises in China. Results show that our approach outperforms several mainstream methods with significant margins. Further insights from more in-depth discussions demonstrate that the "Monopoly" is an innovative application in the interdisciplinary field of business intelligence and urban computing, and it will be beneficial to tens of millions of our users for investments and to the governments for urban planning as well as taxation.

The e-commerce websites are ready to build the community question answering (CQA) service, as it can facilitate questioners (potential buyers) to obtain satisfying answers from experienced customers and furthermore stimulate consumption. Given that more than 50% product-related questions only anticipate a binary response (i.e., “Yes” or “No”), the research on product-related question answering (PQA), which aims to automatically provide instant and correct replies to questioners, emerges rapidly. The mainstream approaches on PQA generally employ customer reviews as the evidence to help predict answers to the questions which are product-specific and concerned more about subjective personal experiences. However, the supportive features either extracted by heuristic rules or acquired from unsupervised manners are not able to perform well on PQA. In this paper, we contribute an end-to-end neural architecture directly fed by the raw text of product-related questions and customer reviews to predict the answers. Concretely, it teaches machines to generate and to synthesize multiple question-aware review representations in a reading comprehension fashion to make the final decision. We also extract a real-world dataset crawled from 9 categories in Amazon.com for PQA to assess the performance of our neural reading architecture (NRA) and other mainstream approaches such as COR-L [12], MOQA [12], and AAP [21]. Experimental results show that our NRA sets up a new state-of-the-art performance on this dataset, significantly outperforming existing algorithms.

Helpful reviews play a pivotal role in recommending desirable goods and accelerating purchase decisions of customers in e-commercial services. Given a large proportion of product reviews with unknown helpfulness/unhelpfulness, the research on automatic identification of helpful reviews has drawn much attention in recent years. However, state-of-the-art approaches still rely heavily on extracting heuristic text features from reviews with domain-specific knowledge. In this paper, we first introduce a multi-task neural learning (MTNL) architecture for identifying helpful reviews. The end-to-end neural architecture can learn to reconstruct effective features upon the raw input of words and even characters, and the multi-task learning paradigm helps to make more accurate predictions of helpful reviews based on a secondary task which fits the star ratings of reviews. We also build two datasets containing helpful/unhelpful reviews from different product categories in Amazon, and compare the performance of MTNL with several mainstream methods on both datasets. Experimental results confirm that MTNL outperforms the state-of-the-art approaches by a significant margin.

Intelligent personal assistant is widely recognized as a more natural and efficient way of human-computer interaction, which has attracted extensive interests from both academia and industry. In this talk, I describe Duer, Baidu's intelligent personal assistant. In particular, I would like to focus on the following three features. Firstly, Duer comprehensively understands people's requirements via multiple channels, including not only explicit utterances, but also user models and rich contexts. Duer's user models are learnt from users' interaction history, and the rich contexts consist of temporal and geographical information, as well as the foregoing dialogues. Secondly, Duer meets diverse requirements with a range of instruments, such as chatting, information provision, reminder service, etc. These instruments are implemented based on mining the big data of web pages, applications, and user logs, which are then seamlessly integrated in the dialogue flow. Thirdly, Duer features multi-modal interaction, which allows people to interact with it by means of texts, speech, and images. We believe the above features will enable Duer to become a better and distinguished intelligent assistant for each of you.

We present a novel method for open domain named entity extraction by exploiting the collective hidden structures in webpage titles. Our method uncovers the hidden textual structures shared by sets of webpage titles based on generalized URL patterns and a multiple sequence alignment technique. The highlights of our method include: 1) The boundaries of entities can be identified automatically in a collective way without any manually designed pattern, seed or class name. 2) The connections between entities are also discovered naturally based on the hidden structures, which makes it easy to incorporate distant or weak supervision. The experiments show that our method can harvest large scale of open domain entities with high precision. A large ratio of the extracted entities are long-tailed and complex and cover diverse topics. Given the extracted entities and their connections, we further show the effectiveness of our method in a weakly supervised setting. Our method can produce better domain specific entities in both precision and recall compared with the state-of-the-art approaches.

We propose a framework which can perform Web page segmentation with a structured prediction approach. It formulates the segmentation task as a structured labeling problem on a transformed Web page segmentation graph (WPS-graph). WPS-graph models the candidate segmentation boundaries of a page and the dependency relation among the adjacent segmentation boundaries. Each labeling scheme on the WPS-graph corresponds to a possible segmentation of the page. The task of finding the optimal labeling of the WPS-graph is transformed into a binary Integer Linear Programming problem, which considers the entire WPS-graph as a whole to conduct structured prediction. A learning algorithm based on the structured output Support Vector Machine framework is developed to determine the feature weights, which is capable to consider the inter-dependency among candidate segmentation boundaries. Furthermore, we investigate its efficacy in supporting the development of automatic Web page classification.

No abstract available.

Statistical bilingual word alignment has been well studied in the field of machine translation. This article adapts the bilingual word alignment algorithm into a monolingual scenario to extract collocations from monolingual corpus, based on the fact that the words in a collocation tend to co-occur in similar contexts as in bilingual word alignment. First, the monolingual corpus is replicated to generate a parallel corpus, in which each sentence pair consists of two identical sentences. Next, the monolingual word alignment algorithm is employed to align potentially collocated words. Finally, the aligned word pairs are ranked according to the alignment scores and candidates with higher scores are extracted as collocations. We conducted experiments on Chinese and English corpora respectively. Compared to previous approaches that use association measures to extract collocations from co-occurrence word pairs within a given window, our method achieves higher precision and recall. According to human evaluation, our method achieves precisions of 62% on a Chinese corpus and 64% on an English corpus. In particular, we can extract collocations with longer spans, achieving a higher precision of 83% on the long-span (> 6 words) Chinese collocations.