Qingyi Wang

Team Seedance, De Chen, Liyang Chen et al. · gatech

Seedance 2.0 is a new native multi-modal audio-video generation model, officially released in China in early February 2026. Compared with its predecessors, Seedance 1.0 and 1.5 Pro, Seedance 2.0 adopts a unified, highly efficient, and large-scale architecture for multi-modal audio-video joint generation. This allows it to support four input modalities: text, image, audio, and video, by integrating one of the most comprehensive suites of multi-modal content reference and editing capabilities available in the industry to date. It delivers substantial, well-rounded improvements across all key sub-dimensions of video and audio generation. In both expert evaluations and public user tests, the model has demonstrated performance on par with the leading levels in the field. Seedance 2.0 supports direct generation of audio-video content with durations ranging from 4 to 15 seconds, with native output resolutions of 480p and 720p. For multi-modal inputs as reference, its current open platform supports up to 3 video clips, 9 images, and 3 audio clips. In addition, we provide Seedance 2.0 Fast version, an accelerated variant of Seedance 2.0 designed to boost generation speed for low-latency scenarios. Seedance 2.0 has delivered significant improvements to its foundational generation capabilities and multi-modal generation performance, bringing an enhanced creative experience for end users.

Qingyi Wang, Shenhao Wang, Dingyi Zhuang et al.

Recent studies have significantly improved the prediction accuracy of travel demand using graph neural networks. However, these studies largely ignored uncertainty that inevitably exists in travel demand prediction. To fill this gap, this study proposes a framework of probabilistic graph neural networks (Prob-GNN) to quantify the spatiotemporal uncertainty of travel demand. This Prob-GNN framework is substantiated by deterministic and probabilistic assumptions, and empirically applied to the task of predicting the transit and ridesharing demand in Chicago. We found that the probabilistic assumptions (e.g. distribution tail, support) have a greater impact on uncertainty prediction than the deterministic ones (e.g. deep modules, depth). Among the family of Prob-GNNs, the GNNs with truncated Gaussian and Laplace distributions achieve the highest performance in transit and ridesharing data. Even under significant domain shifts, Prob-GNNs can predict the ridership uncertainty in a stable manner, when the models are trained on pre-COVID data and tested across multiple periods during and after the COVID-19 pandemic. Prob-GNNs also reveal the spatiotemporal pattern of uncertainty, which is concentrated on the afternoon peak hours and the areas with large travel volumes. Overall, our findings highlight the importance of incorporating randomness into deep learning for spatiotemporal ridership prediction. Future research should continue to investigate versatile probabilistic assumptions to capture behavioral randomness, and further develop methods to quantify uncertainty to build resilient cities.

Yunhan Zheng, Qingyi Wang, Dingyi Zhuang et al.

Short-term demand forecasting for on-demand ride-hailing services is one of the fundamental issues in intelligent transportation systems. However, previous travel demand forecasting research predominantly focused on improving prediction accuracy, ignoring fairness issues such as systematic underestimations of travel demand in disadvantaged neighborhoods. This study investigates how to measure, evaluate, and enhance prediction fairness between disadvantaged and privileged communities in spatial-temporal demand forecasting of ride-hailing services. A two-pronged approach is taken to reduce the demand prediction bias. First, we develop a novel deep learning model architecture, named socially aware neural network (SA-Net), to integrate the socio-demographics and ridership information for fair demand prediction through an innovative socially-aware convolution operation. Second, we propose a bias-mitigation regularization method to mitigate the mean percentage prediction error gap between different groups. The experimental results, validated on the real-world Chicago Transportation Network Company (TNC) data, show that the de-biasing SA-Net can achieve better predictive performance in both prediction accuracy and fairness. Specifically, the SA-Net improves prediction accuracy for both the disadvantaged and privileged groups compared with the state-of-the-art models. When coupled with the bias mitigation regularization method, the de-biasing SA-Net effectively bridges the mean percentage prediction error gap between the disadvantaged and privileged groups, and also protects the disadvantaged regions against systematic underestimation of TNC demand. Our proposed de-biasing method can be adopted in many existing short-term travel demand estimation models, and can be utilized for various other spatial-temporal prediction tasks such as crime incidents predictions.

Qingyi Wang, Shenhao Wang, Yunhan Zheng et al.

Classical demand modeling analyzes travel behavior using only low-dimensional numeric data (i.e. sociodemographics and travel attributes) but not high-dimensional urban imagery. However, travel behavior depends on the factors represented by both numeric data and urban imagery, thus necessitating a synergetic framework to combine them. This study creates a theoretical framework of deep hybrid models with a crossing structure consisting of a mixing operator and a behavioral predictor, thus integrating the numeric and imagery data into a latent space. Empirically, this framework is applied to analyze travel mode choice using the MyDailyTravel Survey from Chicago as the numeric inputs and the satellite images as the imagery inputs. We found that deep hybrid models outperform both the traditional demand models and the recent deep learning in predicting the aggregate and disaggregate travel behavior with our supervision-as-mixing design. The latent space in deep hybrid models can be interpreted, because it reveals meaningful spatial and social patterns. The deep hybrid models can also generate new urban images that do not exist in reality and interpret them with economic theory, such as computing substitution patterns and social welfare changes. Overall, the deep hybrid models demonstrate the complementarity between the low-dimensional numeric and high-dimensional imagery data and between the traditional demand modeling and recent deep learning. It generalizes the latent classes and variables in classical hybrid demand models to a latent space, and leverages the computational power of deep learning for imagery while retaining the economic interpretability on the microeconomics foundation.

Shuzheng Si, Qingyi Wang, Haozhe Zhao et al.

Recognizing whether outputs from large language models (LLMs) contain faithfulness hallucination is crucial for real-world applications, e.g., retrieval-augmented generation and summarization. In this paper, we introduce FaithLens, a cost-efficient and effective faithfulness hallucination detection model that can jointly provide binary predictions and corresponding explanations to improve trustworthiness. To achieve this, we first synthesize training data with explanations via advanced LLMs and apply a well-defined data filtering strategy to ensure label correctness, explanation quality, and data diversity. Subsequently, we fine-tune the model on these well-curated training data as a cold start and further optimize it with rule-based reinforcement learning, using rewards for both prediction correctness and explanation quality. Results on 12 diverse tasks show that the 8B-parameter FaithLens outperforms advanced models such as GPT-4.1 and o3. Also, FaithLens can produce high-quality explanations, delivering a distinctive balance of trustworthiness, efficiency, and effectiveness.

Zhouhong Gu, Xiaoxuan Zhu, Yin Cai et al.

Large language model based multi-agent systems have demonstrated significant potential in social simulation and complex task resolution domains. However, current frameworks face critical challenges in system architecture design, cross-domain generalizability, and performance guarantees, particularly as task complexity and number of agents increases. We introduces AgentGroupChat-V2, a novel framework addressing these challenges through three core innovations: (1) a divide-and-conquer fully parallel architecture that decomposes user queries into hierarchical task forest structures enabling dependency management and distributed concurrent processing. (2) an adaptive collaboration engine that dynamically selects heterogeneous LLM combinations and interaction modes based on task characteristics. (3) agent organization optimization strategies combining divide-and-conquer approaches for efficient problem decomposition. Extensive experiments demonstrate AgentGroupChat-V2's superior performance across diverse domains, achieving 91.50% accuracy on GSM8K (exceeding the best baseline by 5.6 percentage points), 30.4% accuracy on competition-level AIME (nearly doubling other methods), and 79.20% pass@1 on HumanEval. Performance advantages become increasingly pronounced with higher task difficulty, particularly on Level 5 MATH problems where improvements exceed 11 percentage points compared to state-of-the-art baselines. These results confirm that AgentGroupChat-V2 provides a comprehensive solution for building efficient, general-purpose LLM multi-agent systems with significant advantages in complex reasoning scenarios. Code is available at https://github.com/MikeGu721/AgentGroupChat-V2.

Shuzheng Si, Haozhe Zhao, Yu Lei et al.

Many real-world tasks require language models (LMs) to reason over complex contexts that exceed their parametric knowledge. This calls for context learning, where LMs directly learn relevant knowledge from the given context. An intuitive solution is inference-time skill augmentation: extracting the rules and procedures from context into natural-language skills. However, constructing such skills for context learning scenarios faces two challenges: the prohibitive cost of manual skill annotation for long, technically dense contexts, and the lack of external feedback for automated skill construction, since there is no automatic signal to tell whether a proposed skill is helpful. In this paper, we propose Ctx2Skill, a self-evolving framework that autonomously discovers, refines, and selects context-specific skills without human supervision or external feedback. At its core, a multi-agent self-play loop has a Challenger that generates probing tasks and rubrics, a Reasoner that attempts to solve them guided by an evolving skill set, and a neutral Judge that provides binary feedback. Crucially, both the Challenger and the Reasoner evolve through accumulated skills: dedicated Proposer and Generator agents analyze failure cases and synthesize them into targeted skill updates for both sides, enabling automated skill discovery and refinement. To prevent adversarial collapse caused by increasingly extreme task generation and over-specialized skill accumulation, we further introduce a Cross-time Replay mechanism that identifies the skill set achieving the best balance across representative cases for the Reasoner side, ensuring robust and generalizable skill evolution. The resulting skills can be plugged into any language model to obtain better context learning capability. Evaluated on four context learning tasks from CL-bench, Ctx2Skill consistently improves solving rates across backbone models.

Qingyi Wang, Yuebing Liang, Yunhan Zheng et al.

Generative AI offers new opportunities for automating urban planning by creating site-specific urban layouts and enabling flexible design exploration. However, existing approaches often struggle to produce realistic and practical designs at scale. Therefore, we adapt a state-of-the-art Stable Diffusion model, extended with ControlNet, to generate high-fidelity satellite imagery conditioned on land use descriptions, infrastructure, and natural environments. To overcome data availability limitations, we spatially link satellite imagery with structured land use and constraint information from OpenStreetMap. Using data from three major U.S. cities, we demonstrate that the proposed diffusion model generates realistic and diverse urban landscapes by varying land-use configurations, road networks, and water bodies, facilitating cross-city learning and design diversity. We also systematically evaluate the impacts of varying language prompts and control imagery on the quality of satellite imagery generation. Our model achieves high FID and KID scores and demonstrates robustness across diverse urban contexts. Qualitative assessments from urban planners and the general public show that generated images align closely with design descriptions and constraints, and are often preferred over real images. This work establishes a benchmark for controlled urban imagery generation and highlights the potential of generative AI as a tool for enhancing planning workflows and public engagement.

Mingyi He, Yuebing Liang, Shenhao Wang et al.

Urban design is a multifaceted process that demands careful consideration of site-specific constraints and collaboration among diverse professionals and stakeholders. The advent of generative artificial intelligence (GenAI) offers transformative potential by improving the efficiency of design generation and facilitating the communication of design ideas. However, most existing approaches are not well integrated with human design workflows. They often follow end-to-end pipelines with limited control, overlooking the iterative nature of real-world design. This study proposes a stepwise generative urban design framework that integrates multimodal diffusion models with human expertise to enable more adaptive and controllable design processes. Instead of generating design outcomes in a single end-to-end process, the framework divides the process into three key stages aligned with established urban design workflows: (1) road network and land use planning, (2) building layout planning, and (3) detailed planning and rendering. At each stage, multimodal diffusion models generate preliminary designs based on textual prompts and image-based constraints, which can then be reviewed and refined by human designers. We design an evaluation framework to assess the fidelity, compliance, and diversity of the generated designs. Experiments using data from Chicago and New York City demonstrate that our framework outperforms baseline models and end-to-end approaches across all three dimensions. This study underscores the benefits of multimodal diffusion models and stepwise generation in preserving human control and facilitating iterative refinements, laying the groundwork for human-AI interaction in urban design solutions.

Team Seedance, Heyi Chen, Siyan Chen et al.

Recent strides in video generation have paved the way for unified audio-visual generation. In this work, we present Seedance 1.5 pro, a foundational model engineered specifically for native, joint audio-video generation. Leveraging a dual-branch Diffusion Transformer architecture, the model integrates a cross-modal joint module with a specialized multi-stage data pipeline, achieving exceptional audio-visual synchronization and superior generation quality. To ensure practical utility, we implement meticulous post-training optimizations, including Supervised Fine-Tuning (SFT) on high-quality datasets and Reinforcement Learning from Human Feedback (RLHF) with multi-dimensional reward models. Furthermore, we introduce an acceleration framework that boosts inference speed by over 10X. Seedance 1.5 pro distinguishes itself through precise multilingual and dialect lip-syncing, dynamic cinematic camera control, and enhanced narrative coherence, positioning it as a robust engine for professional-grade content creation. Seedance 1.5 pro is now accessible on Volcano Engine at https://console.volcengine.com/ark/region:ark+cn-beijing/experience/vision?type=GenVideo.

Dingyi Zhuang, Qingyi Wang, Yunhan Zheng et al.

Transportation mode share analysis is important to various real-world transportation tasks as it helps researchers understand the travel behaviors and choices of passengers. A typical example is the prediction of communities' travel mode share by accounting for their sociodemographics like age, income, etc., and travel modes' attributes (e.g. travel cost and time). However, there exist only limited efforts in integrating the structure of the urban built environment, e.g., road networks, into the mode share models to capture the impacts of the built environment. This task usually requires manual feature engineering or prior knowledge of the urban design features. In this study, we propose deep hybrid models (DHM), which directly combine road networks and sociodemographic features as inputs for travel mode share analysis. Using graph embedding (GE) techniques, we enhance travel demand models with a more powerful representation of urban structures. In experiments of mode share prediction in Chicago, results demonstrate that DHM can provide valuable spatial insights into the sociodemographic structure, improving the performance of travel demand models in estimating different mode shares at the city level. Specifically, DHM improves the results by more than 20\% while retaining the interpretation power of the choice models, demonstrating its superiority in interpretability, prediction accuracy, and geographical insights.

Da Zhang, Qingyi Wang, Shaojie Song et al.

Estimating health benefits of reducing fossil fuel use from improved air quality provides important rationales for carbon emissions abatement. Simulating pollution concentration is a crucial step of the estimation, but traditional approaches often rely on complicated chemical transport models that require extensive expertise and computational resources. In this study, we develop a novel and succinct machine learning framework that is able to provide precise and robust annual average fine particle (PM2.5) concentration estimations directly from a high-resolution fossil energy use data set. The accessibility and applicability of this framework show great potentials of machine learning approaches for integrated assessment studies. Applications of the framework with Chinese data reveal highly heterogeneous health benefits of reducing fossil fuel use in different sectors and regions in China with a mean of \$34/tCO2 and a standard deviation of \$84/tCO2. Reducing rural and residential coal use offers the highest co-benefits with a mean of \$360/tCO2. Our findings prompt careful policy designs to maximize cost-effectiveness in the transition towards a carbon-neutral energy system.

Guang Hua, Qingyi Wang, Dengpan Ye et al.

Power system frequency could be captured by digital recordings and extracted to compare with a reference database for forensic time-stamp verification. It is known as the electric network frequency (ENF) criterion, enabled by the properties of random fluctuation and intra-grid consistency. In essence, this is a task of matching a short random sequence within a long reference, and the reliability of this criterion is mainly concerned with whether this match could be unique and correct. In this paper, we comprehensively analyze the factors affecting the reliability of ENF matching, including length of test recording, length of reference, temporal resolution, and signal-to-noise ratio (SNR). For synthetic analysis, we incorporate the first-order autoregressive (AR) ENF model and propose an efficient time-frequency domain (TFD) noisy ENF synthesis method. Then, the reliability analysis schemes for both synthetic and real-world data are respectively proposed. Through a comprehensive study we reveal that while the SNR is an important external factor to determine whether time-stamp verification is viable, the length of test recording is the most important inherent factor, followed by the length of reference. However, the temporal resolution has little impact on the matching process.

Shenhao Wang, Qingyi Wang, Jinhua Zhao

It is an enduring question how to combine revealed preference (RP) and stated preference (SP) data to analyze travel behavior. This study presents a framework of multitask learning deep neural networks (MTLDNNs) for this question, and demonstrates that MTLDNNs are more generic than the traditional nested logit (NL) method, due to its capacity of automatic feature learning and soft constraints. About 1,500 MTLDNN models are designed and applied to the survey data that was collected in Singapore and focused on the RP of four current travel modes and the SP with autonomous vehicles (AV) as the one new travel mode in addition to those in RP. We found that MTLDNNs consistently outperform six benchmark models and particularly the classical NL models by about 5% prediction accuracy in both RP and SP datasets. This performance improvement can be mainly attributed to the soft constraints specific to MTLDNNs, including its innovative architectural design and regularization methods, but not much to the generic capacity of automatic feature learning endowed by a standard feedforward DNN architecture. Besides prediction, MTLDNNs are also interpretable. The empirical results show that AV is mainly the substitute of driving and AV alternative-specific variables are more important than the socio-economic variables in determining AV adoption. Overall, this study introduces a new MTLDNN framework to combine RP and SP, and demonstrates its theoretical flexibility and empirical power for prediction and interpretation. Future studies can design new MTLDNN architectures to reflect the speciality of RP and SP and extend this work to other behavioral analysis.

Shenhao Wang, Qingyi Wang, Jinhua Zhao

While deep neural networks (DNNs) have been increasingly applied to choice analysis showing high predictive power, it is unclear to what extent researchers can interpret economic information from DNNs. This paper demonstrates that DNNs can provide economic information as complete as classical discrete choice models (DCMs). The economic information includes choice predictions, choice probabilities, market shares, substitution patterns of alternatives, social welfare, probability derivatives, elasticities, marginal rates of substitution (MRS), and heterogeneous values of time (VOT). Unlike DCMs, DNNs can automatically learn the utility function and reveal behavioral patterns that are not prespecified by domain experts. However, the economic information obtained from DNNs can be unreliable because of the three challenges associated with the automatic learning capacity: high sensitivity to hyperparameters, model non-identification, and local irregularity. To demonstrate the strength and challenges of DNNs, we estimated the DNNs using a stated preference survey, extracted the full list of economic information from the DNNs, and compared them with those from the DCMs. We found that the economic information either aggregated over trainings or population is more reliable than the disaggregate information of the individual observations or trainings, and that even simple hyperparameter searching can significantly improve the reliability of the economic information extracted from the DNNs. Future studies should investigate other regularizations and DNN architectures, better optimization algorithms, and robust DNN training methods to address DNNs' three challenges, to provide more reliable economic information from DNN-based choice models.