Accurate Travel Time Estimation (TTE) is critical for ride-hailing platforms, where errors directly impact user experience and operational efficiency. While existing production systems excel at holistic route-level dependency modeling, they struggle to capture city-scale traffic dynamics and long-tail scenarios, leading to unreliable predictions in large urban networks. In this paper, we propose \model, a scalable and adaptive framework that synergistically integrates link-level modeling with industrial route-level TTE systems. Specifically, we propose a spatio-temporal external attention module to capture global traffic dynamic dependencies across million-scale road networks efficiently. Moreover, we construct a stabilized graph mixture-of-experts network to handle heterogeneous traffic patterns while maintaining inference efficiency. Furthermore, an asynchronous incremental learning strategy is tailored to enable real-time and stable adaptation to dynamic traffic distribution shifts. Experiments on real-world datasets validate MixTTE significantly reduces prediction errors compared to seven baselines. MixTTE has been deployed in DiDi, substantially improving the accuracy and stability of the TTE service.

Subseasonal-to-seasonal (S2S) forecasting, which predicts climate conditions from several weeks to months in advance, represents a critical frontier for agricultural planning, energy management, and disaster preparedness. However, it remains one of the most challenging problems in atmospheric science, due to the chaotic dynamics of atmospheric systems and complex interactions across multiple scales. Current approaches often fail to explicitly model underlying physical processes and teleconnections that are crucial at S2S timescales. We introduce TelePiT, a novel deep learning architecture that enhances global S2S forecasting through integrated multi-scale physics and teleconnection awareness. Our approach consists of three key components: (1) Spherical Harmonic Embedding, which accurately encodes global atmospheric variables onto spherical geometry; (2) Multi-Scale Physics-Informed Neural ODE, which explicitly captures atmospheric physical processes across multiple learnable frequency bands; (3) Teleconnection-Aware Transformer, which models critical global climate interactions through explicitly modeling teleconnection patterns into the self-attention. Extensive experiments demonstrate that TelePiT significantly outperforms state-of-the-art data-driven baselines and operational numerical weather prediction systems across all forecast horizons, marking a significant advance toward reliable S2S forecasting.

Recent advancements in deep learning have led to the development of Foundation Models (FMs) for weather forecasting, yet their ability to predict extreme weather events remains limited. Existing approaches either focus on general weather conditions or specialize in specific-type extremes, neglecting the real-world atmospheric patterns of diversified extreme events. In this work, we identify two key characteristics of extreme events: (1) the spectral disparity against normal weather regimes, and (2) the hierarchical drivers and geographic blending of diverse extremes. Along this line, we propose UniExtreme, a universal extreme weather forecasting foundation model that integrates (1) an Adaptive Frequency Modulation (AFM) module that captures region-wise spectral differences between normal and extreme weather, through learnable Beta-distribution filters and multi-granularity spectral aggregation, and (2) an Event Prior Augmentation (EPA) module which incorporates region-specific extreme event priors to resolve hierarchical extreme diversity and composite extreme schema, via a dual-level memory fusion network. Extensive experiments demonstrate that UniExtreme outperforms state-of-the-art baselines in both extreme and general weather forecasting, showcasing superior adaptability across diverse extreme scenarios.

Modern information retrieval (IR) is no longer consumed primarily by humans but increasingly by large language models (LLMs) via retrieval-augmented generation (RAG) and agentic search. Unlike human users, LLMs are constrained by limited attention budgets and are uniquely vulnerable to noise; misleading or irrelevant information is no longer just a nuisance, but a direct cause of hallucinations and reasoning failures. In this perspective paper, we argue that denoising-maximizing usable evidence density and verifiability within a context window-is becoming the primary bottleneck across the full information access pipeline. We conceptualize this paradigm shift through a four-stage framework of IR challenges: from inaccessible to undiscoverable, to misaligned, and finally to unverifiable. Furthermore, we provide a pipeline-organized taxonomy of signal-to-noise optimization techniques, spanning indexing, retrieval, context engineering, verification, and agentic workflow. We also present research works on information denoising in domains that rely heavily on retrieval such as lifelong assistant, coding agent, deep research, and multimodal understanding.

Web recommendations provide personalized items from massive catalogs for users, which rely heavily on retrieval stages to trade off the effectiveness and efficiency of selecting a small relevant set from billion-scale candidates in online digital platforms. As one of the largest Chinese search engine and news feed providers, Baidu resorts to Deep Neural Network (DNN) and graph-based Approximate Nearest Neighbor Search (ANNS) algorithms for accurate relevance estimation and efficient search for relevant items. However, current retrieval at Baidu fails in comprehensive user-item relational understanding due to dissected interaction modeling, and performs inefficiently in large-scale graph-based ANNS because of suboptimal traversal navigation and the GPU computational bottleneck under high concurrency. To this end, we propose a GPU-accelerated Multi-relational Parallel Graph Retrieval (GMP-GR) framework to achieve effective yet efficient retrieval in web-scale recommendations. First, we propose a multi-relational user-item relevance metric learning method that unifies diverse user behaviors through multi-objective optimization and employs a self-covariant loss to enhance pathfinding performance. Second, we develop a hierarchical parallel graph-based ANNS to boost graph retrieval throughput, which conducts breadth-depth-balanced searches on a large-scale item graph and cost-effectively handles irregular neural computation via adaptive aggregation on GPUs. In addition, we integrate system optimization strategies in the deployment of GMP-GR in Baidu. Extensive experiments demonstrate the superiority of GMP-GR in retrieval accuracy and efficiency. Deployed across more than twenty applications at Baidu, GMP-GR serves hundreds of millions of users with a throughput exceeding one hundred million requests per second.

The rise of Internet of Vehicles (IoV) technologies is transforming traffic management from isolated control to a collective, multi-vehicle process. At the heart of this shift is multi-vehicle dynamic navigation, which requires simultaneously routing large fleets under evolving traffic conditions. Existing path search algorithms and reinforcement learning methods struggle to scale to city-wide networks, often failing to capture the nonlinear, stochastic, and coupled dynamics of urban traffic. To address these challenges, we propose CityNav, a hierarchical, LLM-powered framework for large-scale multi-vehicle navigation. CityNav integrates a global traffic allocation agent, which coordinates strategic traffic flow distribution across regions, with local navigation agents that generate locally adaptive routes aligned with global directives. To enable effective cooperation, we introduce a cooperative reasoning optimization mechanism, in which agents are jointly trained with a dual-reward structure: individual rewards promote per-vehicle efficiency, while shared rewards encourage network-wide coordination and congestion reduction. Extensive experiments on four real-world road networks of varying scales (up to 1.6 million roads and 430,000 intersections) and traffic datasets demonstrate that CityNav consistently outperforms nine classical path search and RL-based baselines in city-scale travel efficiency and congestion mitigation. Our results highlight the potential of LLMs to enable scalable, adaptive, and cooperative city-wide traffic navigation, providing a foundation for intelligent, large-scale vehicle routing in complex urban environments. Our project is available at https://github.com/usail-hkust/CityNav.

Graph Anomaly Detection (GAD) plays a vital role in various data mining applications such as e-commerce fraud prevention and malicious user detection. Recently, Graph Neural Network (GNN) -based approach has demonstrated great effectiveness in GAD by first encoding graph data into low-dimensional representations and then identifying anomalies under the guidance of supervised or unsupervised signals. However, existing GNN-based approaches implicitly follow the homophily principle (i.e., the “like attracts like” phenomenon) and fail to learn discriminative embedding for anomalies that connect vast normal nodes. Moreover, such approaches identify anomalies in a unified global perspective but overlook diversified abnormal patterns conditioned on local graph context, leading to suboptimal performance. To overcome the aforementioned limitations, in this article, we propose a Multi-hypersphere Heterophilic Graph Learning (MHetGL) framework for unsupervised GAD. Specifically, we first devise a Heterophilic Graph Encoding (HGE) module to learn distinguishable representations for potential anomalies by purifying and augmenting their neighborhood in a fully unsupervised manner. Then, we propose a Multi-Hypersphere Learning module to enhance the detection capability for context-dependent anomalies by jointly incorporating critical patterns from both global and local perspectives. Extensive experiments on 11 real-world datasets show that MHetGL outperforms 26 baselines. Our code is publicly available at https://github.com/KennyNH/MHetGL .

Spatial crowdsourcing (SC) is a new form of crowdsourcing that utilizes users (i.e., workers) equipped with smart devices to complete tasks at specific locations. Previous studies usually focus on single task relationships (e.g., dependencies or conflicts) without considering task allocation under multiple relationships. To address this limitation, we jointly consider task dependency and conflict while also considering collaboration among workers for task allocation. In this paper, we define and formulate a new problem, called Joint Dependency and Conflicting Task Allocation in Collaboration-aware Spatial Crowdsourcing (JDCTA), which is proved to be NP-hard. To tackle the JDCTA problem, we first design an approximation algorithm, JDCTA-Greedy, which constructs a set of associated task groups based on task relationships and then greedily allocates these groups, in which we can obtain results with a theoretical bound on the approximate ratio. We then propose JDCTA-Game, a both dependency and conflict aware game approach. JDCTA -Game reduces the strategy space by defining dependency and conflict trees, combined with a dynamic payoff function based on the multiple relationships between tasks, to achieve high-quality solutions. Theoretical analysis demonstrates that this method guarantees the existence of at least one Nash equilibrium, and the solution quality is bounded. Experimental results on both synthetic and real datasets show that our proposed approach outperforms the representative approaches in terms of overall utility.

Pre-trained Language Models (PLMs), such as ChatGPT, have significantly advanced the field of natural language processing. This progress has inspired a series of innovative studies that explore the adaptation of PLMs to time series analysis, intending to create a unified foundation model that addresses various time series analytical tasks. However, these efforts predominantly focus on Regularly Sampled Time Series (RSTS), neglecting the unique challenges posed by Irregularly Sampled Time Series (ISTS), which are characterized by uneven sampling intervals and prevalent missing data. To bridge this gap, this work takes the first step in exploring the potential of PLMs for ISTS analysis. We begin by investigating the effect of various methods for representing ISTS, aiming to maximize the efficacy of PLMs in the analysis. Furthermore, we propose a unified PLM-based framework, named ISTS-PLM, to address diverse ISTS analytical tasks. It integrates novel time-aware and variable-aware PLMs tailored to tackle the intractable intra- and inter-time series modeling in ISTS. Finally, extensive experiments on a comprehensive benchmark demonstrate that the ISTS-PLM, utilizing a structured and effective series-based representation for ISTS, consistently achieves state-of-the-art performance across various analytical tasks, such as classification, interpolation, extrapolation, few-shot and zero-shot learning scenarios, spanning scientific domains like healthcare, biomechanics, and climate science.

Scientific researchers need intensive information about datasets to effectively evaluate and develop theories and methodologies. The information needs regarding datasets are implicitly embedded in particular research tasks, rather than explicitly expressed in search queries. However, existing scientific retrieval and question-answering (QA) datasets typically address straightforward questions, which do not align with the distribution of real-world research inquiries. To bridge this gap, we developed ScIRGen, a dataset generation framework for scientific QA & retrieval that more accurately reflects the information needs of professional science researchers, and uses it to create a large-scale scientific retrieval-augmented generation (RAG) dataset with realistic queries, datasets and papers. Technically, we designed a dataset-oriented information extraction method that leverages academic papers to augment the dataset representation. We then proposed a question generation framework by employing cognitive taxonomy to ensure the quality of synthesized questions. We also design a method to automatically filter synthetic answers based on the perplexity shift of LLMs, which is highly aligned with human judgment of answers' validity. Collectively, these methodologies culminated in the creation of the 61k QA dataset, ScIRGen-Geo. We benchmarked representative methods on the ScIRGen-Geo dataset for their question-answering and retrieval capabilities, finding out that current methods still suffer from reasoning from complex questions. This work advances the development of more sophisticated tools to support the intricate information needs of the scientific community.

Nuclear radiation, which refers to the energy emitted from atomic nuclei during decay, poses significant risks to human health and environmental safety. Recently, advancements in monitoring technology have facilitated the effective recording of nuclear radiation levels and related factors, such as weather conditions. The abundance of monitoring data enables the development of accurate and reliable nuclear radiation forecasting models, which play a crucial role in informing decision-making for individuals and governments. However, this task is challenging due to the imbalanced distribution of monitoring stations over a wide spatial range and the non-stationary radiation variation patterns. In this study, we introduce NRFormer, a novel framework tailored for the nationwide prediction of nuclear radiation variations. By integrating a non-stationary temporal attention module, an imbalance-aware spatial attention module, and a radiation propagation prompting module, NRFormer collectively captures complex spatio-temporal dynamics of nuclear radiation. Extensive experiments on two real-world datasets demonstrate the superiority of our proposed framework against 11 baselines. NRFormer has been deployed online to provide 1-24-day nuclear radiation forecasts, empowering individuals and governments with timely, data-driven decisions for emergency response and public safety. Our framework is designed for general applicability and can be readily adapted for deployment in other regions. The deployed system is available at https://NRFormer.github.io and the dataset and code of the predictive model are available at https://github.com/usail-hkust/NRFormer.

Spatio-temporal forecasting is a critical component of various smart city applications, such as transportation optimization, energy management, and socio-economic analysis. Recently, several automated spatio-temporal forecasting methods have been proposed to automatically search the optimal neural network architecture for capturing complex spatio-temporal dependencies. However, the existing automated approaches suffer from expensive neural architecture search overhead, which hinders their practical use and the further exploration of diverse spatio-temporal operators in a finer granularity. In this paper, we propose AutoSTF, a decoupled automatic neural architecture search framework for cost-effective automated spatio-temporal forecasting. From the efficiency perspective, we first decouple the mixed search space into temporal space and spatial space and respectively devise representation compression and parameter-sharing schemes to mitigate the parameter explosion. The decoupled spatio-temporal search not only expedites the model optimization process but also leaves new room for more effective spatio-temporal dependency modeling. From the effectiveness perspective, we propose a multi-patch transfer module to jointly capture multi-granularity temporal dependencies and extend the spatial search space to enable finer-grained layer-wise spatial dependency search. Extensive experiments on eight datasets demonstrate the superiority of AutoSTF in terms of both accuracy and efficiency. Specifically, our proposed method achieves up to 13.48x speed-up compared to state-of-the-art automatic spatio-temporal forecasting methods while maintaining the best forecasting accuracy. The source code and data are available at https://github.com/usail-hkust/AutoSTF.

On-demand ride-hailing services like DiDi, Uber, and Lyft have transformed urban transportation, offering unmatched convenience and flexibility. In this paper, we introduce DiMA, an LLM-powered ride-hailing assistant deployed in DiDi Chuxing. Its goal is to provide seamless ride-hailing services and beyond through a natural and efficient conversational interface under dynamic and complex spatiotemporal urban contexts. To achieve this, we propose a spatiotemporal-aware order planning module that leverages external tools for precise spatiotemporal reasoning and progressive order planning. Additionally, we develop a cost-effective dialogue system that integrates multi-type dialog repliers with cost-aware LLM configurations to handle diverse conversation goals and trade-off response quality and latency. Furthermore, we introduce a continual fine-tuning scheme that utilizes real-world interactions and simulated dialogues to align the assistant's behavior with human prefered decision-making processes. Since its deployment in the DiDi application, DiMA has demonstrated exceptional performance, achieving 93% accuracy in order planning and 92% in response generation during real-world interactions. Offline experiments further validate DiMA's capabilities, showing improvements of up to 70.23% in order planning and 321.27% in response generation compared to three state-of-the-art agent frameworks, while reducing latency by 0.72x to 5.47x. These results establish DiMA as an effective, efficient, and intelligent mobile assistant for ride-hailing services.

In recent years, advancements in deep learning have spurred the development of numerous models for Long-term Time Series Forecasting (LTSF). However, most existing approaches struggle to fully capture the complex and structured dependencies inherent in time series data. In this work, we identify and formally define three critical dependencies that are fundamental to forecasting accuracy: order dependency and semantic dependency along the temporal dimension, as well as cross-variate dependency across the feature dimension. These dependencies are often treated in isolation, and improper handling can introduce noise and degrade forecasting performance. To bridge this gap, we investigate the potential of State Space Models (SSMs) for LTSF and emphasize their inherent advantages in capturing these essential dependencies. Additionally, we empirically observe that excessive nonlinearity in conventional SSMs introduce redundancy when applied to semantically sparse time series data. Motivated by this insight, we propose SDE (Simplified and Disentangled Dependency Encoding), a novel framework designed to enhance the capability of SSMs for LTSF. Specifically, we first eliminate unnecessary nonlinearities in vanilla SSMs, thereby improving the suitability for time series forecasting. Building on this foundation, we introduce a disentangled encoding strategy, which empowers SSMs to efficiently model cross-variate dependencies while mitigating interference between the temporal and feature dimensions. Furthermore, we provide rigorous theoretical justifications to substantiate our design choices. Extensive experiments on nine real-world benchmark datasets demonstrate that SDE-enhanced SSMs consistently outperform state-of-the-art time series forecasting models. Our code is available at https://github.com/YukinoAsuna/SAMBA.

Graph Neural Networks (GNNs) have demonstrated remarkable proficiency in handling a range of graph analytical tasks across various domains, such as e-commerce and social networks. Despite their versatility, GNNs face significant challenges in transferability, limiting their utility in real-world applications. Existing research in GNN transfer learning overlooks discrepancies in distribution among various graph datasets, facing challenges when transferring across different distributions. How to effectively adopt a well-trained GNN to new graphs with varying feature and structural distributions remains an under-explored problem. Taking inspiration from the success of Low-Rank Adaptation (LoRA) in adapting large language models to various domains, we propose GraphLoRA, an effective and parameter-efficient method for transferring well-trained GNNs to diverse graph domains. Specifically, we first propose a Structure-aware Maximum Mean Discrepancy (SMMD) to align divergent node feature distributions across source and target graphs. Moreover, we introduce low-rank adaptation by injecting a small trainable GNN alongside the pre-trained one, effectively bridging structural distribution gaps while mitigating the catastrophic forgetting. Additionally, a structure-aware regularization objective is proposed to enhance the adaptability of the pre-trained GNN to target graph with scarce supervision labels. Extensive experiments on eight real-world datasets demonstrate the effectiveness of GraphLoRA against fourteen baselines by tuning only 20% of parameters, even across disparate graph domains. The code is available at https://github.com/AllminerLab/GraphLoRA.

Recent years have witnessed the prevalent integration of Large Language Models (LLMs) in various Web applications, such as search engines and recommender systems. As an emerging technique, instruction tuning aims to align pre-trained LLMs as capable chatbots that excel at following human instructions. Previous research indicates that selecting an appropriate subset of a large instruction dataset can enhance the capabilities of LLMs and reduce training costs. However, existing works tend to overlook external correlations between instruction examples during data selection process, which can introduce potential bias and lead to sub-optimal performance. To bridge this gap, we formalize this problem from graph influence maximization perspective and propose Uncertainty-aware influence Maximization (UniMax), a data selection framework that explicitly incorporates the complex inter-dependencies within instruction data. Specifically, we first define a latent instruction graph, treating each instruction example as a graph node and representing their implicit relations as graph edges. Instead of solely relying on heuristic metrics for graph construction, we develop a self-supervised graph learner to uncover the latent structure beyond surface-level feature correlations. After that, we propose an uncertainty-aware influence function to score each example on the instruction graph, allowing a simple greedy algorithm to select a valuable subset that embodies both high influence and uncertainty with an approximation guarantee. Extensive experiments on public datasets show that the proposed approach can significantly enhance model capabilities, underscoring the importance of exploiting data dependencies in instruction data selection.

Subgraph federated learning aims to collaboratively train a global model over distributed subgraphs stored in multiple local clients with strict privacy constraints, which is crucial to a wide range of applications such as healthcare, recommendation systems, and financial crime detection. With the increasing emphasis on the "right to be forgotten," the issue of machine unlearning of subgraph federated models has gained significant importance. However, existing federated unlearning approaches largely focused on unstructured data, overlooking the impact of structural dependency and cross-client interferences in graph-based data. To this end, in this paper, we propose ReGEnUnlearn, a subgraph federated unlearning framework for efficient and comprehensive unlearning of multiple target clients. Specifically, we first propose the Reinforced Federated Policy Sampler (RFPS) to learn optimal sampling strategies that minimize the interference among cross-client subgraphs. By interacting with the federated graph sampling environment, the agent learns to selectively forget an optimal subgraph of target clients, thus preserving the global model utility. Moreover, we propose the Parameter-free Graph Prompt Knowledge Distillation (PGPKD) module, which retains the unique graph knowledge contributed by the target clients, thereby facilitating comprehensive unlearning via a tailored gradient ascent objective. Extensive experiments in various federated settings demonstrate ReGEnUnlearn's superiority over existing federated unlearning methods, offering a speedup of 3.6× to 9× compared to traditional retraining while maintaining model utility within a range of 100%-102%.

Spatio-Temporal Prediction (STP) is crucial for various smart city applications, such as traffic management and resource allocation. However, training samples can be scarce in data-constrained scenarios, which often degrades the predictive capability of existing deep STP models. Although recent STP foundation models excel in few-shot and zero-shot learning through extensive pre-training on large-scale, multi-domain spatio-temporal data, they often rely on large parameter scale to achieve enhanced performance, resulting in high computational demands that hinder practical deployment. In response, we develop CompactST, an efficient, compact, and versatile pre-trained model for STP in data-scarce settings. Recognizing the complexities posed by large-scale, heterogeneous pre-training datasets, CompactST integrates three specialized components: (1) a mixture-of-normalizers module to address domain and spatial heterogeneity, (2) a multi-scale spatio-temporal mixer that captures diverse patterns from datasets with varying spatio-temporal resolutions, and (3) an adaptive dataset-oriented tuning module that transfers the handling of dataset-specific parameters from pre-training to fine-tuning stage. These tailored designs enable CompactST to maximize generalizability across diverse datasets while maintaining a compact model size ( i.e. , only 300K parameters). To validate its effectiveness, we pre-train CompactST on a substantial corpus of public spatio-temporal datasets spanning over 10 domains and encompassing 300 million data points. Extensive experimental results on ten real-world datasets demonstrate CompactST's significantly improved prediction accuracy and efficiency in data-scarce scenarios.

Urban land use, intrinsically linked to people’s daily activities, undergoes continuous evolution, presenting a complex interplay that remains partially understood. To bridge this gap, our study leverages fine-grained human mobility data to predict these changes, adopting a novel approach that conceptualizes “community-level” land use shifts as a regression problem and represents citywide changes through dynamic graphs. We harness recent advancements in graph neural networks (GNNs), which, despite their success in various applications, face challenges in directly predicting land use changes due to the temporal mismatch between the slow evolution of urban land and the immediacy of human mobility data. Our research stands out by introducing a temporal skeleton for dynamic GNNs to synchronize human activity graphs with urban land use changes, a dynamic heterogeneous GNN approach for integrating diverse human activity data to capture essential temporal dependencies, and a novel algorithm powered by causal inference to elucidate the primary factors influencing land use predictions at the community level, all of which contribute to a training process informed by the generated causal graph. Empirically validated on three real-world datasets, our model demonstrates a performance leap over state-of-the-art baselines, marking a pivotal step toward understanding and predicting the dynamics of urban land use.

Graph recommendation methods, representing a connected interaction perspective, reformulate user–item interactions as graphs to leverage graph structure and topology to recommend and have proved practical effectiveness at scale. Large language models (LLMs), representing a textual generative perspective, excel at modeling user languages, understanding behavioral contexts, capturing user–item semantic relationships, analyzing textual sentiments, and generating coherent and contextually relevant texts as recommendations. However, there is a gap between the connected graph perspective and the text generation perspective as the task formulations are different. A research question arises: how can we effectively integrate the two perspectives for more personalized RecSys? To fill this gap, we propose to incorporate graph-edge information into LLMs via prompt and attention innovations. We reformulate recommendations as a probabilistic generative problem using prompts. We develop a framework to incorporate graph edge information from the prompt and attention mechanisms for graph-structured LLM recommendations. We develop a new prompt design that brings in both first-order and second-order graph relationships; we devise an improved LLM attention mechanism to embed direct the spatial and connectivity information of edges. Our evaluation of real-world datasets demonstrates the framework’s ability to understand connectivity information in graph data and to improve the relevance and quality of recommendation results. Our code is released at: https://github.com/anord-wang/LLM4REC.git .

Machine learning techniques are now integral to the advancement of intelligent urban services, playing a crucial role in elevating the efficiency, sustainability, and livability of urban environments. The recent emergence of foundation models such as ChatGPT marks a revolutionary shift in the fields of machine learning and artificial intelligence. Their unparalleled capabilities in contextual understanding, problem solving, and adaptability across a wide range of tasks suggest that integrating these models into urban domains could have a transformative impact on the development of smart cities. Despite growing interest in Urban Foundation Models (UFMs), this burgeoning field faces challenges such as a lack of clear definitions and systematic reviews. To this end, this paper first introduces the concept of UFMs and discusses the unique challenges involved in building them. We then propose a data-centric taxonomy that categorizes and clarifies current UFM-related works, based on urban data modalities and types. Furthermore, we explore the application landscape of UFMs, detailing their potential impact in various urban contexts. Relevant papers and open-source resources have been collated and are continuously updated at: https://github.com/usail-hkust/Awesome-Urban-Foundation-Models.

Accurate traffic forecasting is crucial for the development of Intelligent Transportation Systems (ITS), playing a pivotal role in modern urban traffic management. Traditional forecasting methods, however, struggle with the irregular traffic time series resulting from adaptive traffic signal controls, presenting challenges in asynchronous spatial dependency, irregular temporal dependency, and predicting variable-length sequences. To this end, we propose an Asynchronous Spatio-tEmporal graph convolutional nEtwoRk (ASeer) tailored for irregular traffic time series forecasting. Specifically, we first propose an Asynchronous Graph Diffusion Network to capture the spatial dependency between asynchronously measured traffic states regulated by adaptive traffic signals. After that, to capture the temporal dependency within irregular traffic state sequences, a personalized time encoding is devised to embed the continuous time signals. Then, we propose a Transformable Time-aware Convolution Network, which adapts meta-filters for time-aware convolution on the sequences with inconsistent temporal flow. Additionally, a Semi-Autoregressive Prediction Network, comprising a state evolution unit and a semiautoregressive predictor, is designed to predict variable-length traffic sequences effectively and efficiently. Extensive experiments on a newly established benchmark demonstrate the superiority of ASeer compared with twelve competitive baselines across six metrics.

Federated Graph Learning (FGL) has emerged as a promising way to learn high-quality representations from distributed graph data with privacy preservation. Despite considerable efforts have been made for FGL under either cross-device or cross-silo paradigm, how to effectively capture graph knowledge in a more complicated cross-silo cross-device environment remains an under-explored problem. However, this task is challenging because of the inherent hierarchy and heterogeneity of decentralized clients, diversified privacy constraints in different clients, and the cross-client graph integrity requirement. To this end, in this paper, we propose a Hierarchical Federated Graph Learning (HiFGL) framework for cross-silo cross-device FGL. Specifically, we devise a unified hierarchical architecture to safeguard federated GNN training on heterogeneous clients while ensuring graph integrity. Moreover, we propose a Secret Message Passing (SecMP) scheme to shield unauthorized access to subgraph-level and node-level sensitive information simultaneously. Theoretical analysis proves that HiFGL achieves multi-level privacy preservation with complexity guarantees. Extensive experiments on real-world datasets validate the superiority of the proposed framework against several baselines. Furthermore, HiFGL's versatile nature allows for its application in either solely cross-silo or cross-device settings, further broadening its utility in real-world FGL applications.

Rapid urbanization has significantly escalated traffic congestion, underscoring the need for advanced congestion prediction services to bolster intelligent transportation systems.As one of the world's largest ride-hailing platforms, DiDi places great emphasis on the accuracy of congestion prediction to enhance the effectiveness and reliability of their real-time services, such as travel time estimation and route planning.Despite numerous efforts have been made on congestion prediction, most of them fall short in handling heterogeneous and dynamic spatio-temporal dependencies (e.g., periodic and non-periodic congestions), particularly in the presence of noisy and incomplete traffic data.In this paper, we introduce a Congestion Prediction Mixture-of-Experts, CP-MoE, to address the above challenges.We first propose a sparsely-gated Mixture of Adaptive Graph Learners (MAGLs) with congestion-aware inductive biases to improve the model capacity for efficiently capturing complex spatio-temporal dependencies in varying traffic scenarios.Then, we devise two specialized experts to help identify stable trends and periodic patterns within the traffic data, respectively.By cascading these experts with MAGLs, CP-MoE delivers congestion predictions in a more robust and interpretable manner.Furthermore, an ordinal regression strategy is adopted to facilitate effective * Corresponding author.

Spatio-Temporal Graph Neural Network (STGNN) has been used as a common workhorse for traffic forecasting. However, most of them require prohibitive quadratic computational complexity to capture long-range spatio-temporal dependencies, thus hindering their applications to long historical sequences on large-scale road networks in the real-world. To this end, in this paper, we propose BigST, a linear complexity spatio-temporal graph neural network, to efficiently exploit long-range spatio-temporal dependencies for large-scale traffic forecasting. Specifically, we first propose a scalable long sequence feature extractor to encode node-wise long-range inputs ( e.g. , thousands of time-steps in the past week) into low-dimensional representations encompassing rich temporal dynamics. The resulting representations can be pre-computed and hence significantly reduce the computational overhead for prediction. Then, we build a linearized global spatial convolution network to adaptively distill time-varying graph structures, which enables fast runtime message passing along spatial dimensions in linear complexity. We empirically evaluate our model on two large-scale real-world traffic datasets. Extensive experiments demonstrate that BigST can scale to road networks with up to one hundred thousand nodes, while significantly improving prediction accuracy and efficiency compared to state-of-the-art traffic forecasting models.

Recent years have witnessed the fast development of deep learning techniques in quantitative trading. It still remains unclear how to exploit deep learning techniques to improve high-frequency trading (HFT). Indeed, there are two emerging challenges for the use of deep learning for HFT: (i) how to quantify fast-changing market conditions for tick-level signal prediction; (ii) how to establish a unified trading paradigm for different securities of diverse market conditions and severe signal sparsity. To this end, in this paper, we propose anOnlineAdaptive Sequence Learning(LEVER) framework, which consists of two distinct components to predict the HFT signals at the tick level for a variety of securities simultaneously. Specifically, we start with a single learner that adopts an encoder-decoder architecture for each security-based HFT signal prediction. In this single learner, an ordered encoder module first captures the variability patterns of the security's price curve by encoding the input indicator sequence from different time ranges. An unordered decoder module then outlines the pivot points of the price curve as support and resistance levels to quantify the market status. Based on the measured market condition, a prediction module further approximates the impacts of upcoming security data as the potential market momentum to detect the tick-level trading signals. To overcome the computational challenges and signal sparsity posed by online HFT for multiple securities, we develop a competitive active-meta learning paradigm to enhance the signal learners’ learning efficiency for online implementation. Finally, extensive experiments on real-world stock market data demonstrate the effectiveness of our deployed LEVER for improving the performances of the existing industry method by 0.27 in the Sharpe ratio and by 0.09% in a transaction-based return.

Real-time bus travel time is crucial for the smart public transportation system and is beneficial for improving user satisfaction for online map services. However, it faces great challenges due to fine-grained spatial dependencies and dynamic temporal dependencies. To address the above problem, we propose GBTTE, a novel end-to-end graph attention network framework to estimate bus travel time. Specifically, we construct a novel graph structure of bus routes and use a graph attention network to capture the fine-grained spatial features of bus routes. Then, we fully exploit the joint spatial-temporal relations of bus stops through a spatial-temporal graph attention network and also capture the dynamic correlation between the route and the bus transportation network with a cross graph attention network. Finally, we integrate the route representation, the spatial-temporal representation and contextual information to estimate bus travel time. Extensive experiments carried out on two large-scale real-world datasets demonstrate the effectiveness of GBTTE. In addition, GBTTE has been deployed in production at Baidu Maps, handling tens of millions of requests every day.

Medical tests are crucial for treatment decision making. However, over-testing can often occur in any medical speciality or level of expertise. Since over-testing usually results in a financial burden for patients and is also a waste of medical resources, this naturally leads to the question: which medical test items (MTIs) are necessary and should be prioritized for the target patients? It is a nontrivial task to identify the right MTIs due to the diversified health status of patients and the complicated prerequisites of therapies. To this end, in this paper, we propose a data-driven approach to evaluate the priority which should be given to MTIs by modeling the relationships between MTIs and therapies. Specifically, we first develop a dual hierarchical topic model (DHTM), which views the adopted hierarchical therapies as labeled topics and the MTI reports, i.e., the set of hierarchical attribute-value pairs (AVPs), as documents. Then, with the therapy-AVP distribution and the partial MTI reports of the target patient, we can scope the candidate therapies, which are further utilized to evaluate the accumulated gain of MTIs to be tested. Moreover, the next MTI recommendation is conducted based on the gains. Finally, extensive experiments on real-world medical data validate the effectiveness of our approach, and some interesting observations are also provided. The code is available at https://github.com/mtirec/MtiRec.

Time-of-arrival estimation or Estimated Time of Arrival (ETA) has become an indispensable building block of modern intelligent transportation systems. While many efforts have been made for time-of-arrival estimation, most of them have scalability and robustness issues when dealing with real-world large-scale ETA scenarios, where billions of vehicle trajectories and ETA requests have been continuously generating every day. To this end, in this paper, we propose a robust and scalable incremental ETA learning framework, iETA, to continuously exploit spatio-temporal traffic patterns from massive floating-car data and thus achieve better estimation performances. Specifically, we first build an incremental travel time predictor that can be incrementally updated based on newly generated traffic data. The incremental travel time predictor not only reduces the overall learning overhead but also improves the model's robustness toward urban traffic distribution shifts. Then, we propose a historical traffic knowledge consolidation module to preserve critical spatio-temporal knowledge from previous ETA predictors under the incremental learning setting. Moreover, to reduce interference induced by low-quality traffic data, we propose an adversarial training module to improve the learning robustness by proactively mitigating and resisting traffic noise perturbations. Finally, extensive experiments demonstrate the effectiveness and efficiency of the proposed system against state-of-the-art baselines in large-scale ETA scenarios. Most importantly, iETA has been deployed on the Didi Chuxing platform, handling real-time billions of ETA queries every day, and substantially improves the prediction accuracy.

Vehicle Energy Consumption (VEC) estimation aims to predict the total energy required for a given trip before it starts, which is of great importance to trip planning and transportation sustainability. Existing approaches mainly focus on extracting statistically significant factors from typical trips to improve the VEC estimation. However, the energy consumption of each vehicle may diverge widely due to the personalized driving behavior under varying travel contexts. To this end, this paper proposes a preference-aware meta-optimization framework Meta-Pec for personalized vehicle energy consumption estimation. Specifically, we first propose a spatiotemporal behavior learning module to capture the latent driver preference hidden in historical trips. Moreover, based on the memorization of driver preference, we devise a selection-based driving behavior prediction module to infer driver-specific driving patterns on a given route, which provides additional basis and supervision signals for VEC estimation. Besides, a driver-specific meta-optimization scheme is proposed to enable fast model adaption by learning and sharing transferable knowledge globally. Extensive experiments on two real-world datasets show the superiority of our proposed framework against ten numerical and data-driven machine learning baselines. The source code is available at https://github.com/usail-hkust/Meta-Pec.

Travel Time Estimation (TTE) aims to accurately forecast the expected trip duration from an origin to a destination. As one of the world's largest ride-hailing platforms, DiDi answers billions of TTE queries per day. The quality of TTE directly decides the customer's experience and the effectiveness of passenger-to-driver matching. However, existing studies mainly regard TTE as a deterministic regression problem and focus on improving the prediction accuracy of a single label, which overlooks the travel time uncertainty induced by various dynamic contextual factors. To this end, in this paper, we propose a probabilistic framework, ProbTTE, for uncertainty-aware travel time prediction. Specifically, the framework first transforms the single-label regression task to a multi-class classification problem to estimate the implicit travel time distribution. Moreover, we propose an adaptive local label-smoothing scheme to capture the ordinal inter-class relationship among soft travel time labels. Furthermore, we construct a route-wise log-normal distribution regularizer to absorb prior knowledge from large-scale historical trip data. By explicitly considering the travel uncertainty, the proposed approach not only improves the TTE accuracy but also provides additional travel time information to benefit downstream tasks in ride-hailing. Extensive experiments on real-world datasets demonstrate the superiority of the proposed framework compared with state-of-the-art travel time prediction algorithms. In addition, ProbTTE has been deployed in production at DiDi in late 2022 to empower various order dispatching services, and improves passenger and driver experiences significantly.

Machine learning-based forecasting models are commonly used in Intelligent Transportation Systems (ITS) to predict traffic patterns and provide city-wide services. However, most of the existing models are susceptible to adversarial attacks, which can lead to inaccurate predictions and negative consequences such as congestion and delays. Therefore, improving the adversarial robustness of these models is crucial for ITS. In this paper, we propose a novel framework for incorporating adversarial training into spatiotemporal traffic forecasting tasks. We demonstrate that traditional adversarial training methods designated for static domains cannot be directly applied to traffic forecasting tasks, as they fail to effectively defend against dynamic adversarial attacks. Then, we propose a reinforcement learning-based method to learn the optimal node selection strategy for adversarial examples, which simultaneously strengthens the dynamic attack defense capability and reduces the model overfitting. Additionally, we introduce a self-knowledge distillation regularization module to overcome the "forgetting issue" caused by continuously changing adversarial nodes during training. We evaluate our approach on two real-world traffic datasets and demonstrate its superiority over other baselines. Our method effectively enhances the adversarial robustness of spatiotemporal traffic forecasting models. The source code for our framework is available at https://github.com/usail-hkust/RDAT.

Urban vibrancy describes the prosperity, diversity, and accessibility of urban areas, which is vital to a city’s socio-economic development and sustainability. While many efforts have been made for statically measuring and evaluating urban vibrancy, there are few studies on the evolutionary process of urban vibrancy, yet we know little about the relationship between urban vibrancy evolution and sophisticated spatiotemporal dynamics. In this article, we make use of multi-sourced urban data to develop a data-driven framework, U-Evolve , to investigate urban vibrancy evolution. Specifically, we first exploit the spatiotemporal characteristics of urban areas to create multi-view time-dependent graphs. Then, we analyze the contextual features and graph patterns of multi-view time-dependent graphs in terms of informing future urban vibrancy variations. Our analysis validates the informativeness of multi-view time-dependent graphs for characterizing and informing future urban vibrancy evolution. After that, we construct a feature based model to forecast future urban vibrancy evolution and quantify each feature’s importance. Moreover, to further enhance the forecasting effectiveness, we propose a graph learning based model to capture spatiotemporal autocorrelation of urban areas based on multi-view time-dependent graphs in an end-to-end manner. Finally, extensive experiments on two metropolises, Beijing and Shanghai, demonstrate the effectiveness of our forecasting models. The U-Evolve framework has also been deployed in the production environment to deliver real-world urban development and planning insights for various cities in China.

We study the problem of balancing effectiveness and efficiency in automated feature selection. Feature selection is to find an optimal feature subset from large feature space. After exploring many feature selection methods, we observe a computational dilemma: 1) traditional feature selection (e.g., mRMR) is mostly efficient, but difficult to identify the best subset; 2) the emerging reinforced feature selection automatically navigates feature space to search the best subset, but is usually inefficient. Are automation and efficiency always apart from each other? Can we bridge the gap between effectiveness and efficiency under automation? Motivated by this dilemma, we aim to develop a novel feature space navigation method. In our preliminary work, we leveraged interactive reinforcement learning to accelerate feature selection by external trainer-agent interaction. Our preliminary work can be significantly improved by modeling the structured knowledge of its downstream task (e.g., decision tree) as learning feedback. In this journal version, we propose a novel interactive and closed-loop architecture to simultaneously model interactive reinforcement learning (IRL) and decision tree feedback (DTF). Specifically, IRL is to create an interactive feature selection loop and DTF is to feed structured feature knowledge back to the loop. The DTF improves IRL from two aspects. First, the tree-structured feature hierarchy generated by decision tree is leveraged to improve state representation. In particular, we represent the selected feature subset as an undirected graph of feature-feature correlations and a directed tree of decision features. We propose a new embedding method capable of empowering Graph Convolutional Network (GCN) to jointly learn state representation from both the graph and the tree. Second, the tree-structured feature hierarchy is exploited to develop a new reward scheme. In particular, we personalize reward assignment of agents based on decision tree feature importance. In addition, observing agents’ actions can also be a feedback, we devise another new reward scheme, to weigh and assign reward based on the selected frequency ratio of each agent in historical action records. Finally, we present extensive experiments with real-world datasets to demonstrate the improved performances of our method.

Predicting air quality in fine spatiotemporal granularity is of great importance for air pollution control and urban sustainability. However, existing studies are either focused on predicting station-wise future air quality, or inferring current air quality for unmonitored regions. How to accurately forecast future air quality for these unmonitored regions in a fine granularity remains an unexplored problem. In this paper, we propose the Self-Supervised Hierarchical Graph Neural Network (SSH-GNN), for fine-grained air quality forecasting in a semi-supervised way. Specifically, to augment spatially sparse air quality observations, SSH-GNN first approximates the city-wide air quality distribution based on historical readings and various urban contextual factors (e.g., weather conditions and traffic flows). Then, we propose a hierarchical recurrent graph neural network to make city-wide predictions, which encodes the spatial hierarchy of urban regions for long-range spatiotemporal correlation modeling. Moreover, by leveraging spatiotemporal self-supervision strategies, SSH-GNN exploits both universal topological and contextual patterns to further enhance the forecasting effectiveness. Extensive experiments on two real-world datasets show that SSH-GNN significantly outperforms the state-of-the-art algorithms.

Accurate and timely air quality and weather predictions are of great importance to urban governance and human livelihood. Though many efforts have been made for air quality or weather prediction, most of them simply employ one another as feature input, which ignores the inner-connection between two predictive tasks. On one hand, the accurate prediction of one task can help improve another task's performance. On the other hand, geospatially distributed air quality and weather monitoring stations provide additional hints for city-wide spatiotemporal dependency modeling. Inspired by the above two insights, in this paper, we propose a multi-view multi-adversarial approach, entitled MasterGNN$^{+}$, to jointly predict air quality and weather conditions. First, we devise a multi-view graph learning block to model spatial autocorrelation based on geographical distance and environmental context. Then, a dedicated evolved recurrent network is proposed to dynamically capture the long-range and independent temporal autocorrelation for each monitoring station and time slot. After that, we develop a multi-adversarial graph learning framework to against observation noise propagation introduced by spatiotemporal modeling. Moreover, we present an adaptive training strategy by formulating multi-adversarial learning as a multi-task learning problem. Finally, extensive experiments on two real-world datasets show that MasterGNN$^{+}$achieves the best performance compared with seven baselines on both air quality and weather prediction tasks.

Whilelearning to rank(LTR) has been widely used in web search to prioritize most relevant webpages among the retrieved contents subject to the input queries, the traditional LTR models fail to deliver decent performance due to two main reasons: 1) the lack of well-annotated query-webpage pairs with ranking scores to cover search queries of various popularity, and 2) ill-trained models based on a limited number of training samples with poor generalization performance. To improve the performance of LTR models, tremendous efforts have been done from above two aspects, such as enlarging training sets with pseudo-labels of ranking scores by self-training, or refining the features used for LTR through feature extraction and dimension reduction. Though LTR performance has been marginally increased, we still believe these methods could be further improved in the newly-fashioned “interpolating regime”. Specifically, instead of lowering the number of features used for LTR models, our work proposes to transform original data with random Fourier feature, so as to over-parameterize the downstream LTR models (e.g., GBRank or LightGBM) with features in ultra-high dimensionality and achieve superb generalization performance. Furthermore, rather than self-training with pseudo-labels produced by the same LTR model in a “self-tuned” fashion, the proposed method incorporates the diversity of prediction results between the listwise and pointwise LTR models while co-training both models with a cyclic labeling-prediction pipeline in a “ping-pong” manner. We deploy the proposedCo-trained andOver-parameterizedLTRsystemCOLTRat Baidu search and evaluateCOLTRwith a large number of baseline methods. The results show thatCOLTRcould achieve$\Delta NDCG_{4}$= 3.64%$\sim$4.92%, compared to baselines, under various ratios of labeled samples. We also conduct a 7-day A/B Test using the realistic web traffics of Baidu Search, where we can still observe significant performance improvement around$\Delta NDCG_{4}$= 0.17%$\sim$0.92% in real-world applications.COLTRperforms consistently both in online and offline experiments.

Public transportation plays a critical role in peoples daily life. It has been proven that public transportation is more environmentally sustainable, efficient, and economical than any other forms of travel. However, due to the increasing expansion of transportation networks and more complex travel situations, people are having difficulties in efficiently finding the most preferred route from one place to another through public transportation systems for both intra-city and inter-city trips. To this end, in this paper, we present Polestar++, a data-driven engine for intelligent and efficient public transportation routing. Specifically, we first propose a novel hierarchical public transportation graph(HPTG) to model both intra-city and inter-city public transportation in terms of various travel costs, such as time or distance. Then, we introduce a general route search algorithm coupled with an efficient station binding method for efficient route candidate generation. After that, we propose a two-pass route candidate ranking module to capture user preferences under dynamic travel situations. Finally, experiments on two real-world data sets demonstrate the advantages of Polestar++ in terms of both efficiency and effectiveness. To date, Polestar++ is servicing over 330 cities, answers over a hundred millions of queries each day, and achieves substantial improvement of user click ratio.

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.

Electric Vehicle (EV) has become preferable choices in modern transportation system due to its environmental and energy sustainability. However, in many large cities, EV drivers often fail to find proper spots for charging because of the limited charging infrastructures and spatiotemporally unbalanced charging demands. Indeed, the recent emergence of deep reinforcement learning provides great potential to improve charging experience over long-term horizons. In this paper, we propose RLCharge for intelligent EV charging station recommendation by jointly considering various long-term spatiotemporal factors. Specifically, by regarding each charging station as an agent, we formulate the problem as a multi-objective multi-agent reinforcement learning task. We first develop a multi-agent actor-critic framework with centralized training decentralized execution. Particularly, we propose a tailor designed centralized attentive critic with the delayed access strategy to coordinate the recommendation between geo-distributed agents during centralized training. Besides, we propose the spatio-temporal heterogeneous graph convolution module to handle the partial observability problem during decentralized execution. After that, to effectively optimize multiple divergent objectives, we develop a dynamic gradient re-weighting strategy to adaptively guide the optimization direction, and propose an adaptive imitation learning scheme to further accelerate and stabilize the policy convergence. Finally, extensive experiments on two real-world datasets demonstrate that RLCHARGE achieves the best comprehensive performance compared with ten baseline approaches.

Electric Vehicles (EVs) have been emerging as a promising low-carbon transport target. While a large number of public charging stations are available, the use of these stations is often imbalanced, causing many problems to Charging Station Operators (CSOs). To this end, in this paper, we propose a Multi-Agent Graph Convolutional Reinforcement Learning (MAGC) framework to enable CSOs to achieve more effective use of these stations by providing dynamic pricing for each of the continuously arising charging requests with optimizing multiple long-term commercial goals. Specifically, we first formulate this charging station request-specific dynamic pricing problem as a mixed competitive-cooperative multi-agent reinforcement learning task, where each charging station is regarded as an agent. Moreover, by modeling the whole charging market as a dynamic heterogeneous graph, we devise a multi-view heterogeneous graph attention networks to integrate complex interplay between agents induced by their diversified relationships. Then, we propose a shared meta generator to generate individual customized dynamic pricing policies for large-scale yet diverse agents based on the extracted meta characteristics. Finally, we design a contrastive heterogeneous graph pooling representation module to learn a condensed yet effective state action representation to facilitate policy learning of large-scale agents. Extensive experiments on two real-world datasets demonstrate the effectiveness of MAGC and empirically show that the overall use of stations can be improved if all the charging stations in a charging market embrace our dynamic pricing policy.

Deep Neural Network (DNN) based recommendation systems are widely used in the modern internet industry for a variety of services. However, the rapid expansion of application scenarios and the explosive global internet traffic growth have caused the industry to face increasing challenges to serve the complicated recommendation workflow regarding online recommendation efficiency and compute resource overhead. In this paper, we present a GPU-accelerated online serving system, namely Lion, which consists of the staged event-driven heterogeneous pipeline, unified memory manager, and automatic execution optimizer to handle web-scale traffic in a real-time and cost-effective way. Moreover, Lion provides a heterogeneous template library to enable fast development and migration for diverse in-house web-scale recommendation systems without requiring knowledge of heterogeneous programming. The system is currently deployed at Baidu, supporting over twenty recommendation services, including news feed, short video clips, and the search engine. Extensive experimental studies on five real-world deployed online recommendation services demonstrate the superiority of the proposed GPU-accelerated online serving system. Since launched in early 2020, Lion has answered billions of recommendation requests per day, and has helped Baidu successfully save millions of U.S. dollars in hardware and utility costs per year.

Talent demand and supply forecasting aims to model the variation of the labor market, which is crucial to companies for recruitment strategy adjustment and to job seekers for proactive career path planning. However, existing approaches either focus on talent demand or supply forecasting, but overlook the interconnection between demand-supply sequences among different companies and positions. To this end, in this paper, we propose a Dynamic Heterogeneous Graph Enhanced Meta-learning (DH-GEM) framework for fine-grained talent demand-supply joint prediction. Specifically, we first propose a Demand-Supply Joint Encoder-Decoder (DSJED) and a Dynamic Company-Position Heterogeneous Graph Convolutional Network (DyCP-HGCN) to respectively capture the intrinsic correlation between demand and supply sequences and company-position pairs. Moreover, a Loss-Driven Sampling based Meta-learner (LDSM) is proposed to optimize long-tail forecasting tasks with a few training data. Extensive experiments have been conducted on three real-world datasets to demonstrate the effectiveness of our approach compared with five baselines. DH-GEM has been deployed as a core component of the intelligent human resource system of a cooperative partner.

Data science competitions are becoming increasingly popular for enterprises collecting advanced innovative solutions and allowing contestants to sharpen their data science skills. Most existing studies about data science competitions have a focus on improving task-specific data science techniques, such as algorithm design and parameter tuning. However, little effort has been made to understand the data science competition itself. To this end, in this article, we shed light on the team’s competition performance, and investigate the team’s evolving performance in the crowd-sourcing competitive innovation context. Specifically, we first acquire and construct multi-sourced datasets of various data science competitions, including the KDD Cup 2019 machine learning competition and beyond. Then, we conduct an empirical analysis to identify and quantify a rich set of features that are significantly correlated with teams’ future performances. By leveraging team’s rank as a proxy, we observe “the stronger, the stronger” rule; that is, top-ranked teams tend to keep their advantages and dominate weaker teams for the rest of the competition. Our results also confirm that teams with diversified backgrounds tend to achieve better performances. After that, we formulate the team’s future rank prediction problem and propose the Multi-Task Representation Learning (MTRL) framework to model both static features and dynamic features. Extensive experimental results on four real-world data science competitions demonstrate the team’s future performance can be well predicted by using MTRL. Finally, we envision our study will not only help competition organizers to understand the competition in a better way, but also provide strategic implications to contestants, such as guiding the team formation and designing the submission strategy.

The understanding of people’s inter-regional mobility behaviors, such as predicting the next activity region (AR) or uncovering the intentions for regional mobility, is of great value to public administration or business interests. While there are numerous studies on human mobility, these studies are mainly from a statistical view or study movement behaviors within a region. The work on individual-level inter-regional mobility behavior is limited. To this end, in this article, we propose a dynamic region-relation-aware graph neural network (DRRGNN) for exploring individual mobility behaviors over ARs. Specifically, we aim at developing models that can answer three questions: (1) Which regions are the ARs? (2) Which region will be the next AR, and (3) Why do people make this regional mobility? To achieve these tasks, we first propose a method to find out people’s ARs. Then, the designed model integrates a dynamic graph convolution network (DGCN) and a recurrent neural network (RNN) to depict the evolution of relations between ARs and mine the regional mobility patterns. In the learning process, the model further considers peoples’ profiles and visited point-of-interest (POIs). Finally, extensive experiments on two real-world datasets show that the proposed model can significantly improve accuracy for both the next AR prediction and mobility intention prediction.

Transportation recommendation is one important map service in navigation applications. Previous transportation recommendation solutions fail to deliver satisfactory user experience because their recommendations only consider routes in one transportation mode (uni-modal, e.g., taxi, bus, cycle) and largely overlook situational context. In this work, we propose$\mathsf {Hydra}$, a multi-task deep learning based recommendation system that offers multi-modal transportation planning and is adaptive to various situational context (e.g., nearby point-of-interest (POI) distribution and weather). We leverage the availability of existing routing engines and big urban data, and design a novel two-level framework that integrates uni-modal and multi-modal (e.g., taxi-bus, bus-cycle) routes as well as heterogeneous urban data for intelligent multi-modal transportation recommendation. In addition to urban context features constructed from multi-source urban data, we learn the latent representations of users, origin-destination (OD) pairs and transportation modes based on user implicit feedbacks, which captures the collaborative transportation mode preferences of users and OD pairs. Moreover, we propose two models to recommend the proper route among various uni-modal and multi-modal transportation routes: (1) a light-weight gradient boosting decision tree (GBDT) based recommendation model; and (2) a multi-task wide and deep learning (MTWDL) based recommendation model. We also optimize the framework to support real-time, large-scale route query and recommendation. We deploy$\mathsf {Hydra}$on Baidu Maps,11.https://maps.baidu.com/.one of the world's largest map services. Real-world urban-scale experiments demonstrate the effectiveness and efficiency of our proposed system. Since its deployment in August 2018,$\mathsf {Hydra}$has answered over a hundred million route recommendation queries made by over ten million distinct users. The GBDT based model and MTWDL based model achieve 82.8 and 96.6 percent relative improvement of user click ratio, respectively.

The ability to predict city-wide parking availability is crucial for the successful development of Parking Guidance and Information (PGI) systems. The effective prediction of city-wide parking availability can boost parking efficiency, improve urban planning, and ultimately alleviate city congestion. However, it is a non-trivial task for city-wide parking availability prediction because of three major challenges: 1) the non-euclidean spatial autocorrelation among parking lots, 2) the dynamic temporal autocorrelation inside of and between parking lots, and 3) the scarcity of information about real-time parking availability obtained from real-time sensors (e.g., camera, ultrasonic sensor, and bluetooth sensor). To this end, we propose aSemi-supervisedHierarchicalRecurrent Graph Neural Network-X(SHARE-X) to predict parking availability of each parking lot within a city. Specifically, we first propose a hierarchical graph convolution module to model the non-euclidean spatial autocorrelation among parking lots. Along this line, a contextual graph convolution block and a multi-resolution soft clustering graph convolution block are respectively proposed to capture local and global spatial dependencies between parking lots. Moreover, we devise a hierarchical attentive recurrent network module to incorporate both short and long-term dynamic temporal dependencies of parking lots. Additionally, a parking availability approximation module is introduced to estimate missing real-time parking availabilities from both spatial and temporal domains. Finally, experiments on two real-world datasets demonstrate thatSHARE-Xoutperforms eight state-of-the-art baselines in parking availability prediction.

The fairness issue is very important in deploying machine learning models as algorithms widely used in human society can be easily in discrimination. Researchers have studied disparity on tabular data a lot and proposed many methods to relieve bias. However, studies towards unfairness in graph are still at early stage while graph data that often represent connections among people in real-world applications can easily give rise to fairness issues and thus should be attached to great importance. Fair representation learning is one of the most effective methods to relieve bias, which aims to generate hidden representations of input data while obfuscating sensitive information. In graph setting, learning fair representations of graph (also called fair graph embeddings) is effective to solve graph unfairness problems. However, most existing works of fair graph embeddings only study fairness in a coarse granularity (i.e., group fairness), but overlook individual fairness. In this paper, we study fair graph representations from different levels. Specifically, we consider both group fairness and individual fairness on graph. To debias graph embeddings, we propose FairGAE, a fair graph auto-encoder model, to derive unbiased graph embeddings based on the tailor-designed fair Graph Convolution Network (GCN) layers. Then, to achieve multi-level fairness, we design a Wasserstein distance based regularizer to learn the optimal transport for fairer embeddings. To overcome the efficiency concern, we further bring up Sinkhorn divergence as the approximations of Wasserstein cost for computation. Finally, we apply the learned unbiased embeddings into the node classification task and conduct extensive experiments on two real-world graph datasets to demonstrate the improved performances of our approach.

Real estate appraisal refers to the process of developing an unbiased opinion for real property's market value, which plays a vital role in decision-making for various players in the marketplace (e.g., real estate agents, appraisers, lenders, and buyers). However, it is a non-trivial task for accurate real estate appraisal because of three major challenges: (1) The complicated influencing factors for property value; (2) The asynchronously spatiotemporal dependencies among real estate transactions; (3) The diversified correlations between residential communities. To this end, we propose a Multi-Task Hierarchical Graph Representation Learning (MugRep) framework for accurate real estate appraisal. Specifically, by acquiring and integrating multi-source urban data, we first construct a rich feature set to profile the real estate from multiple perspectives~(e.g., geographical distribution, human mobility distribution, and resident demographics distribution). Then, an evolving real estate transaction graph and a corresponding event graph convolution module are proposed to incorporate asynchronously spatiotemporal dependencies among real estate transactions. Moreover, to further incorporate valuable knowledge from the view of residential communities, we devise a hierarchical heterogeneous community graph convolution module to capture diversified correlations between residential communities. Finally, an urban district partitioned multi-task learning module is introduced to generate differently distributed value opinions for real estate. Extensive experiments on two real-world datasets demonstrate the effectiveness of MugRep and its components and features.

In modern internet industries, deep learning based recommender systems have became an indispensable building block for a wide spectrum of applications, such as search engine, news feed, and short video clips. However, it remains challenging to carry the well-trained deep models for online real-time inference serving, with respect to the time-varying web-scale traffics from billions of users, in a cost-effective manner. In this work, we present JIZHI - a Model-as-a-Service system - that per second handles hundreds of millions of online inference requests to huge deep models with more than trillions of sparse parameters, for over twenty real-time recommendation services at Baidu, Inc. In JIZHI, the inference workflow of every recommendation request is transformed to a Staged Event-Driven Pipeline (SEDP), where each node in the pipeline refers to a staged computation or I/O intensive task processor. With traffics of real-time inference requests arrived, each modularized processor can be run in a fully asynchronized way and managed separately. Besides, JIZHI introduces the heterogeneous and hierarchical storage to further accelerate the online inference process by reducing unnecessary computations and potential data access latency induced by ultra-sparse model parameters. Moreover, an intelligent resource manager has been deployed to maximize the throughput of JIZHI over the shared infrastructure by searching the optimal resource allocation plan from historical logs and fine-tuning the load shedding policies over intermediate system feedback. Extensive experiments have been done to demonstrate the advantages of JIZHI from the perspectives of end-to-end service latency, system-wide throughput, and resource consumption. Since launched in July 2019, JIZHI has helped Baidu saved more than ten million US dollars in hardware and utility costs per year while handling 200% more traffics without sacrificing the inference efficiency.

Motivated by the success of pre-trained language models such as BERT in a broad range of natural language processing (NLP) tasks, recent research efforts have been made for adapting these models for different application domains. Along this line, existing domain-oriented models have primarily followed the vanilla BERT architecture and have a straightforward use of the domain corpus. However, domain-oriented tasks usually require accurate understanding of domain phrases, and such fine-grained phrase-level knowledge is hard to be captured by existing pre-training scheme. Also, the word co-occurrences guided semantic learning of pre-training models can be largely augmented by entity-level association knowledge. But meanwhile, there is a risk of introducing noise due to the lack of groundtruth word-level alignment. To address the issues, we provide a generalized domain-oriented approach, which leverages auxiliary domain knowledge to improve the existing pre-training framework from two aspects. First, to preserve phrase knowledge effectively, we build a domain phrase pool as auxiliary knowledge, meanwhile we introduce Adaptive Hybrid Masked Model to incorporate such knowledge. It integrates two learning modes, word learning and phrase learning, and allows them to switch between each other. Second, we introduce Cross Entity Alignment to leverage entity association as weak supervision to augment the semantic learning of pre-trained models. To alleviate the potential noise in this process, we introduce an interpretableOptimal Transport based approach to guide alignment learning. Experiments on four domain-oriented tasks demonstrate the superiority of our framework.

To cope with the fast-evolving business trend, it becomes critical for companies to continuously review their talent recruitment strategies by the timely forecast of talent demand in recruitment market. While many efforts have been made on recruitment market analysis, due to the sparsity of fine-grained talent demand time series and the complex temporal correlation of the recruitment market, there is still no effective approach for fine-grained talent demand forecast, which can quantitatively model the dynamics of the recruitment market. To this end, in this paper, we propose a data-driven neural sequential approach, namely Talent Demand Attention Network (TDAN), for forecasting fine-grained talent demand in the recruitment market. Specifically, we first propose to augment the univariate time series of talent demand at multiple grained levels and extract intrinsic attributes of both companies and job positions with matrix factorization techniques. Then, we design a Mixed Input Attention module to capture company trends and industry trends to alleviate the sparsity of fine-grained talent demand. Meanwhile, we design a Relation Temporal Attention module for modeling the complex temporal correlation that changes with the company and position. Finally, extensive experiments on a real-world recruitment dataset clearly validate the effectiveness of our approach for fine-grained talent demand forecast, as well as its interpretability for modeling recruitment trends. In particular, TDAN has been deployed as an important functional component of intelligent recruitment system of cooperative partner.

Feature selection is a fundamental component of data mining, aiming to select optimal feature subsets for downstream task.Recently, an emerging feature selection method called reinforced feature selection applies reinforcement learning into feature selection.Reinforced Feature Selection (RFS) automates feature selection process and can effectively find the optimal subset.Generally, RFS can be categorized into single-agent RFS and multi-agent RFS.Single-agent RFS uses one reinforcement learning agent to select features, but its action space is exponentially-increasing with feature number and can merely obtain local optima.Multi-agent RFS uses multiple agents to select features; this method can achieve global optima, but it needs to optimize as many policy networks as feature number which costs huge computational resources and thus becomes computationally inefficient.This dilemma naturally leads to a research question: How can we synthesize the advantages of single-agent RFS and multi-agent RFS while avoiding their disadvantages?To answer this question, we propose a Group-based Interactive Reinforced Feature Selection (GIRFS) framework.This framework balances single-agent RFS and multi-agent RFS for better feature selection.Specifically, we formulate the feature selection problem into a group-based RFS problem.In this formulation, we first assign the given features into several groups based on feature similarity measurement.Then, we create agents for each group, where each agent decides to select/deselect features in its corresponding group.This design balances the size of action space and number of policy networks and thus makes RFS more effective and efficient.Moreover, to further improve learning efficiency, we propose a hierarchical teacher-like trainer to provide external action advice for agents.This trainer provides advice by intra-group selection and inter-group selection and fuses knowledge from mRMR and decision tree to help agents explore and learn.Finally, we present extensive experiments on real-world datasets to demonstrate the improved performances of our method.

Query and Point-of-Interest (POI) matching, aiming at recommending the most relevant POIs from partial query keywords, has become one of the most essential functions in online navigation and ride-hailing applications. Existing methods for query-POI matching, such as Google Maps and Uber, have a natural focus on measuring the static semantic similarity between contextual information of queries and geographical information of POIs. However, it remains challenging for dynamic and personalized online query-POI matching because of the non-stationary and situational context-dependent query-POI relevance. Moreover, the large volume of online queries requires an adaptive and incremental model training strategy that is efficient and scalable in the online scenario. To this end, in this paper, we propose an Incremental Spatio-Temporal Graph Learning (IncreSTGL) framework for intelligent online query-POI matching. Specifically, we first model dynamic query-POI interactions as microscopic and macroscopic graphs. Then, we propose an incremental graph representation learning module to refine and update query-POI interaction graphs in an online incremental fashion, which includes: (i) a contextual graph attention operation quantifying query-POI correlation based on historical queries under dynamic situational context, (ii) a graph discrimination operation capturing the sequential query-POI relevance drift from a holistic view of personalized preference and social homophily, and (iii) a multi-level temporal attention operation summarizing the temporal variations of query-POI interaction graphs for subsequent query-POI matching. Finally, we introduce a lightweight semantic matching module for online query-POI similarity measurement. To demonstrate the effectiveness and efficiency of the proposed algorithm, we conduct extensive experiments on two real-world datasets collected from a leading online navigation and map service provider in China.

Electric Vehicle (EV) has become a preferable choice in the modern transportation system due to its environmental and energy sustainability. However, in many large cities, EV drivers often fail to find the proper spots for charging, because of the limited charging infrastructures and the spatiotemporally unbalanced charging demands. Indeed, the recent emergence of deep reinforcement learning provides great potential to improve the charging experience from various aspects over a long-term horizon. In this paper, we propose a framework, named Multi-Agent Spatio-Temporal Reinforcement Learning (Master), for intelligently recommending public accessible charging stations by jointly considering various long-term spatiotemporal factors. Specifically, by regarding each charging station as an individual agent, we formulate this problem as a multi-objective multi-agent reinforcement learning task. We first develop a multi-agent actor-critic framework with the centralized attentive critic to coordinate the recommendation between geo-distributed agents. Moreover, to quantify the influence of future potential charging competition, we introduce a delayed access strategy to exploit the knowledge of future charging competition during training. After that, to effectively optimize multiple learning objectives, we extend the centralized attentive critic to multi-critics and develop a dynamic gradient re-weighting strategy to adaptively guide the optimization direction. Finally, extensive experiments on two real-world datasets demonstrate that Master achieves the best comprehensive performance compared with nine baseline approaches.

In this paper, we study the problem of balancing effectiveness and efficiency in automated feature selection. Feature selection is to find the optimal feature subset from large-scale feature space, and is a fundamental intelligence for machine learning and predictive analysis. After exploring many feature selection methods, we observe a computational dilemma: 1) traditional feature selection methods (e.g., K-Best, decision tree based ranking, mRMR) are mostly efficient, but difficult to identify the best subset; 2) the emerging reinforced feature selection methods automatically navigate feature space to explore the best subset, but are usually inefficient. Are automation and efficiency always apart from each other? Can we bridge the gap between effectiveness and efficiency under automation? Motivated by such a computational dilemma, this study is to develop a novel feature space navigation method. To that end, we propose an Interactive Reinforced Feature Selection (IRFS) framework that guides agents by not just self-exploration experience, but also diverse external skilled trainers to accelerate learning for feature exploration. Specifically, we formulate the feature selection problem into an interactive reinforcement learning framework. In this framework, we first model two trainers skilled at different searching strategies: (1) KBest based trainer; (2) Decision Tree based trainer. We then develop two strategies: (1) to identify assertive and hesitant agents to diversify agent training, and (2) to enable the two trainers to take the teaching role in different stages to fuse the experience of the trainers and diversify teaching process. Such a hybrid teaching strategy can help agents to learn broader knowledge, and thereafter be more effective. Finally, we present extensive experiments on real-world datasets to demonstrate the improved performances of our method: more efficient than reinforced selection and more effective than classic feature selection.

The competitive relationship of Points of Interest (POIs) refers to the degree of competition between two POIs for business opportunities from third parties in an urban area. Existing studies for competitive analysis usually focus on mining competitive relationships of entities, such as companies or products, from textual data. However, there are few studies which have a focus on competitive analysis for POIs. Indeed, the growing availability of user behavior data about POIs, such as POI reviews and human mobility data, enables a new paradigm for understanding the competitive relationships among POIs. To this end, in this paper, we study how to predict the POI competitive relationship. Along this line, a very first challenge is how to integrate heterogeneous user behavior data with the spatial features of POIs. As a solution, we first build a heterogeneous POI information network (HPIN) from POI reviews and map search data. Then, we develop a graph neural network-based deep learning framework, named DeepR, for POI competitive relationship prediction based on HPIN. Specifically, DeepR contains two components: a spatial adaptive graph neural network (SA-GNN) and a POI pairwise knowledge extraction learning (PKE) model. The SA-GNN is a novel GNN architecture with incorporating POI's spatial information and location distribution by a specially designed spatial oriented aggregation layer and spatial-dependency attentive propagation mechanism. In addition, PKE is devised to distill the POI pairwise knowledge in HPIN being useful for relationship prediction into condensate vectors with relational graph convolution and cross attention. Finally, extensive experiments on two real-world datasets demonstrate the effectiveness of our method.

Public transportation plays a critical role in people's daily life. It has been proven that public transportation is more environmentally sustainable, efficient, and economical than any other forms of travel. However, due to the increasing expansion of transportation networks and more complex travel situations, people are having difficulties in efficiently finding the most preferred route from one place to another through public transportation systems. To this end, in this paper, we present Polestar, a data-driven engine for intelligent and efficient public transportation routing.Specifically, we first propose a novel Public Transportation Graph (PTG) to model public transportation system in terms of various travel costs, such as time or distance. Then, we introduce a general route search algorithm coupled with an efficient station binding method for efficient route candidate generation. After that, we propose a two-pass route candidate ranking module to capture user preferences under dynamic travel situations. Finally, experiments on two real-world data sets demonstrate the advantages of Polestar in terms of both efficiency and effectivenes Indeed, in early 2019, Polestar has been deployed on Baidu Maps, one of the world's largest map services. To date, Polestar is servicing over 330 cities, answers over a hundred millions of queries each day, and achieves substantial improvement of user click ratio.

Existing spatial object recommendation algorithms generally treat objects identically when ranking them. However, spatial objects often cover different levels of spatial granularity and thereby are heterogeneous. For example, one user may prefer to be recommended a region (say Manhattan), while another user might prefer a venue (say a restaurant). Even for the same user, preferences can change at different stages of data exploration. In this paper, we study how to support top-k spatial object recommendations at varying levels of spatial granularity, enabling spatial objects at varying granularity, such as a city, suburb, or building, as a Point of Interest (POI). To solve this problem, we propose the use of a POI tree, which captures spatial containment relationships between POIs. We design a novel multi-task learning model called MPR (short for Multi-level POI Recommendation), where each task aims to return the top-k POIs at a certain spatial granularity level. Each task consists of two subtasks: (i) attribute-based representation learning; (ii) interaction-based representation learning. The first subtask learns the feature representations for both users and POIs, capturing attributes directly from their profiles. The second subtask incorporates user-POI interactions into the model. Additionally, MPR can provide insights into why certain recommendations are being made to a user based on three types of hints: user-aspect, POI-aspect, and interaction-aspect. We empirically validate our approach using two real-life datasets, and show promising performance improvements over several state-of-the-art methods.

In location-based services, such as navigation and ride-hailing, it is an essential function to match a query with Point-of-Interests (POIs) for efficient destination retrieval. Indeed, due to the space limit and real-time requirement, such services usually require intermediate POI matching results when only partial search keywords are typed. While there are numerous retrieval models for general textual semantic matching, few attempts have been made for query-POI matching by considering the integration of rich spatio-temporal factors and dynamic user preferences. To this end, in this paper, we develop a spatio-temporal dual graph attention network ~(STDGAT), which can jointly model dynamic situational context and users' sequential behaviors for intelligent query-POI matching. Specifically, we first utilize a semantic representation block to model semantic correlations among incomplete texts as well as various spatio-temporal factors captured by location and time. Next, we propose a novel dual graph attention network to capture two types of query-POI relevance, where one models global query-POI interaction and another one models time-evolving user preferences on destination POIs. Moreover, we also incorporate spatio-temporal factors into the dual graph attention network so that the query-POI relevance can be generalized to the sophisticated situational context. After that, a pairwise fusion strategy is introduced to extract the salient global feature representatives for both queries and POIs. Finally, several cold-start strategies and training methods are proposed to improve the matching effectiveness and training efficiency. Extensive experiments on two real-world datasets demonstrate the performances of our approach compared with state-of-the-art baselines. The results show that our model achieves significant improvement in terms of matching accuracy even with only partial query keywords are given.

Multi-modal transportation recommendation aims to provide the most appropriate travel route with various transportation modes according to certain criteria. After analyzing large-scale navigation data, we find that route representations exhibit two patterns: spatio-temporal autocorrelations within transportation networks and the semantic coherence of route sequences. However, there are few studies that consider both patterns when developing multi-modal transportation systems. To this end, in this paper, we study multi-modal transportation recommendation with unified route representation learning by exploiting both spatio-temporal dependencies in transportation networks and the semantic coherence of historical routes. Specifically, we propose to unify both dynamic graph representation learning and hierarchical multi-task learning for multi-modal transportation recommendations. Along this line, we first transform the multi-modal transportation network into time-dependent multi-view transportation graphs and propose a spatiotemporal graph neural network module to capture the spatial and temporal autocorrelation. Then, we introduce a coherent-aware attentive route representation learning module to project arbitrary-length routes into fixed-length representation vectors, with explicit modeling of route coherence from historical routes. Moreover, we develop a hierarchical multi-task learning module to differentiate route representations for different transport modes, and this is guided by the final recommendation feedback as well as multiple auxiliary tasks equipped in different network layers. Extensive experimental results on two large-scale real-world datasets demonstrate the performance of the proposed system outperforms eight baselines.

Transportation recommendation is one important map service in navigation applications. Previous transportation recommendation solutions fail to deliver satisfactory user experience because their recommendations only consider routes in one transportation mode (uni-modal, e.g., taxi, bus, cycle) and largely overlook situational context. In this work, we propose Hydra, a recommendation system that offers multi-modal transportation planning and is adaptive to various situational context (e.g., nearby point-of-interest (POI) distribution and weather). We leverage the availability of existing routing engines and big urban data, and design a novel two-level framework that integrates uni-modal and multi-modal (e.g., taxi-bus, bus-cycle) routes as well as heterogeneous urban data for intelligent multi-modal transportation recommendation. In addition to urban context features constructed from multi-source urban data, we learn the latent representations of users, origin-destination (OD) pairs and transportation modes based on user implicit feedbacks, which captures the collaborative transportation mode preferences of users and OD pairs. A gradient boosting tree based model is then introduced to recommend the proper route among various uni-modal and multi-modal transportation routes. We also optimize the framework to support real-time, large-scale route query and recommendation. We deploy Hydra on Baidu Maps, one of the world's largest map services. Real-world urban-scale experiments demonstrate the effectiveness and efficiency of our proposed system. Since its deployment in August 2018, Hydra has answered over a hundred million route recommendation queries made by over ten million distinct users with 82.8% relative improvement of user click ratio.

Profiling driver behaviors and designing appropriate pricing models are essential for auto insurance companies to gain profits and attract customers (drivers). The existing approaches either rely on static demographic information like age, or model only coarse-grained driving behaviors. They are therefore ineffective to yield accurate risk predictions over time for appropriate pricing, resulting in profit decline or even financial loss. Moreover, existing pricing strategies seldom take profit maximization into consideration, especially under the enterprise constraints. The recent growth of vehicle telematics data (vehicle sensing data) brings new opportunities to auto insurance industry, because of its sheer size and fine-grained mobility for profiling drivers. But, how to fuse these sparse, inconsistent and heterogeneous data is still not well addressed. To tackle these problems, we propose a unified PPP (Profile-Price-Profit) framework, working on the real-world large-scale vehicle telematics data and insurance data. PPP profiles drivers' fine-grained behaviors by considering various driving features from the trajectory perspective. Then, to predict drivers' risk probabilities, PPP leverages the group-level insight and categorizes drivers' different temporal risk change patterns into groups by ensemble learning. Next, the pricing model in PPP incorporates both the demographic analysis and the mobility factors of driving risk and mileage, to generate personalized insurance price for supporting flexible premium periods. Finally, the maximal profit problem proves to be NP-Complete. Then, an efficient heuristic-based dynamic programming is proposed. Extensive experimental results demonstrated that, PPP effectively predicts the driver's risk and outperforms the current company's pricing strategy (in industry) and the state-of-the-art approach. PPP also achieves near the maximal profit (difference by only 3%) for the company, and lowers the total price for the drivers.

In modern big data warehouse systems, we observe a common phenomenon that a column of data values can be derived from one or several other columns by transforming and concatenating these columns. We call this relationship between columns a Soft Concatenation Mapping (SCM). SCMs imply significant redundancy in the schema or data, and therefore can be exploited for data integration or data compression. In this paper, we formalize the problem of SCM detection and prove it is NP-hard. We then propose efficient approximate algorithms to detect all SCMs or an optimal set of SCMs in a table. Our experiments on both real-world and synthetic datasets show promising results.

In this paper, we present TICC, an automatic data compression component that can transparently eliminate data redundancies across columns in column-oriented database systems. We further propose two approaches to integrate inter-column compression into existing database systems. One approach is to use User Defined Functions (UDFs), and the other is native. We implement these two approaches on top of Hive based on the ORC file, a common data format in column stores, and evaluate the performance of TICC using real-world datasets. The experimental results demonstrate that TICC can significantly reduce the storage overhead and process a variety of queries over large-scale data with up to 20% performance improvement over the original Hive.

UnionPay's inter-bank transaction settlement platform (ITSP) generates a huge amount of bankcard transaction data everyday, recording different bankcard activities. In order to unleash the business value of these data, UnionPay has built a customized data warehouse based on Hadoop to manage and query the massive data imported from ITSP. However, the original system suffers from low storage utilization due to various types of data redundancy. Such data redundancy is caused by the long-term evolution of the system architecture. It dramatically wastes storage space, degrades query performance and leads to data inconsistency problem. In order to address these issues, we have developed UStore, an optimized storage system to reduce most data redundancies and improve query performance. In this paper, we present the design and implementation of UStore in detail. We test the performance of UStore on UnionPay's real data and the results show significant improvements in both storage utilization and query performance. To date, UStore has been deployed to process over 15 years' bankcard transaction data (over 3PB in plain text format) in UnionPay.