arXiv subjects

Jiawei Lu

Publications and source records attributed to Jiawei Lu.

16 recordsLinked to original sources

arXiv2026

MathGen: Revealing the Illusion of Mathematical Competence through Text-to-Image Generation

Modern generative models have demonstrated the ability to solve challenging mathematical problems. In many real-world settings, however, mathematical solutions must be expressed visually through diagrams, plots, geometric constructions, and structured symbolic layouts, where correctness depends on precise visual composition. This naturally raises the question of whether generative models can still do so when the answer must be rendered visually rather than written in text? To study this problem, we introduce MathGen, a rigorous benchmark of 420 problems spanning seven core domains, including 350 Clean-Scene problems and 70 paired Open-Scene problems. Each problem is evaluated under a Script-as-a-Judge protocol with problem-specific verification criteria implemented through reusable executable scripts for deterministic and reproducible evaluation. Experiments on representative open-source and proprietary text-to-image models show that mathematical fidelity remains a major bottleneck: even the best closed-source model reaches only 53.7% overall accuracy, while open-source models achieve just 1--11%, often near 0% on structured tasks, particularly those requiring precise geometric and functional rendering. Overall, current T2I models remain far from reliable at even elementary mathematical visual generation.

cs.CV↗

arXiv2026

Rollout-Training Co-Design for Efficient LLM-Based Multi-Agent Reinforcement Learning

Despite algorithm-level innovations for multi-agent reinforcement learning (MARL), the underlying networked infrastructure for large-scale MARL training remains underexplored. Existing training frameworks primarily optimize for single-agent scenarios and fail to address the unique system-level challenges of MARL, including rollout-training synchronization barriers, rollout load imbalance, and training resource underutilization. To bridge this gap, we propose FlexMARL, the first end-to-end training framework that holistically optimizes rollout, training, and their orchestration for large-scale LLM-based MARL. Specifically, FlexMARL introduces the joint orchestrator to manage data flow under the rollout-training disaggregated architecture. Building upon the experience store, a novel micro-batch driven asynchronous pipeline eliminates the synchronization barriers while providing strong consistency guarantees. Rollout engine adopts a parallel sampling scheme combined with hierarchical load balancing, which adapts to skewed inter/intra-agent request patterns. Training engine achieves on-demand hardware binding through agent-centric resource allocation. The training states of different agents are swapped via unified and location-agnostic communication. Empirical results on a large-scale production cluster demonstrate that FlexMARL achieves up to 7.3x speedup and improves hardware utilization by up to 5.6x compared to existing frameworks.

cs.LG↗

arXiv2025

MatPedia: A Universal Generative Foundation for High-Fidelity Material Synthesis

Physically-based rendering (PBR) materials are fundamental to photorealistic graphics, yet their creation remains labor-intensive and requires specialized expertise. While generative models have advanced material synthesis, existing methods lack a unified representation bridging natural image appearance and PBR properties, leading to fragmented task-specific pipelines and inability to leverage large-scale RGB image data. We present MatPedia, a foundation model built upon a novel joint RGB-PBR representation that compactly encodes materials into two interdependent latents: one for RGB appearance and one for the four PBR maps encoding complementary physical properties. By formulating them as a 5-frame sequence and employing video diffusion architectures, MatPedia naturally captures their correlations while transferring visual priors from RGB generation models. This joint representation enables a unified framework handling multiple material tasks--text-to-material generation, image-to-material generation, and intrinsic decomposition--within a single architecture. Trained on MatHybrid-410K, a mixed corpus combining PBR datasets with large-scale RGB images, MatPedia achieves native $1024\times1024$ synthesis that substantially surpasses existing approaches in both quality and diversity.

cs.CV↗

arXiv2025

HiMA-Ecom: Enabling Joint Training of Hierarchical Multi-Agent E-commerce Assistants

Hierarchical multi-agent systems based on large language models (LLMs) have become a common paradigm for building AI assistants in vertical domains such as e-commerce, where a master agent coordinates multiple specialized sub-agents. Despite their practical importance, realistic benchmarks for training and evaluating such systems remain scarce, and joint optimization across functionally distinct agents is still challenging. To address this gap, we introduce HiMA-Ecom, the first hierarchical multi-agent benchmark tailored for e-commerce scenarios. HiMA-Ecom contains 22.8K instances, including agent-specific supervised fine-tuning samples with memory and system-level input-output pairs for joint multi-agent reinforcement learning. Building upon it, a joint training method named HiMA-R1 is proposed. It presents Variance-Reduction Group Relative Policy Optimization (VR-GRPO), which employs initial trajectory-based Monte Carlo sampling to mitigate the exponential joint action space and selects informative agent groups for efficient updates based on reward variance. Furthermore, an adaptive memory evolution mechanism that repurposes GRPO rewards as cost-free supervisory signals is designed to eliminate repetitive reasoning and accelerate convergence. Experiments on HiMA-Ecom demonstrate that our method, built upon smaller 3B/7B open-source models, achieves performance comparable to that of larger LLMs, such as DeepSeek-R1, and surpasses DeepSeek-V3 by an average of 6\%.

cs.AI↗

arXiv2025

Bridging the Gap: Enabling Natural Language Queries for NoSQL Databases through Text-to-NoSQL Translation

NoSQL databases are core data infrastructure, yet natural-language access to them remains underdeveloped: correct query generation must recover how a non-relational data model represents entities, nested paths, arrays, missing fields, and dynamic keys. This paper studies Text-to-NoSQL, translating natural-language requests into executable NoSQL queries, instantiated with MongoDB aggregation pipelines over schema-less document stores. We present TEND, short for Text-to-NoSQL Dataset, an execution-verified benchmark with 1,210 MongoDB-native tasks across 11 databases. To our knowledge, TEND is the first Text-to-NoSQL benchmark whose database worlds are MongoDB-native by design: experts manually define collection boundaries, nested arrays, optional and sparse paths, polymorphic shapes, and dynamic-key conventions; these worlds are populated with real data and verified through frozen MongoDB execution, so TEND evaluates schema-less document reasoning rather than SQL-to-MQL transfer. We further introduce SAG, a Schema-as-Data Grounding solver that induces path and value grounding from stored-document evidence before bounded MQL generation, execution-grounded repair, and result-consistency selection. Evaluation uses bounded column-tolerant execution accuracy (EXC) as the headline metric, complemented by a graded result-set F1 and a mutually exclusive execution-outcome decomposition. Experiments show that LLMs with strong NL2SQL performance degrade substantially on TEND, validating Text-to-NoSQL as a distinct schema-less document reasoning problem.

cs.DB↗

arXiv2024

TPLogAD: Unsupervised Log Anomaly Detection Based on Event Templates and Key Parameters

Log-system is an important mechanism for recording the runtime status and events of Web service systems, and anomaly detection in logs is an effective method of detecting problems. However, manual anomaly detection in logs is inefficient, error-prone, and unrealistic. Existing log anomaly detection methods either use the indexes of event templates, or form vectors by embedding the fixed string part of the template as a sentence, or use time parameters for sequence analysis. However, log entries often contain features and semantic information that cannot be fully represented by these methods, resulting in missed and false alarms. In this paper, we propose TPLogAD, a universal unsupervised method for analyzing unstructured logs, which performs anomaly detection based on event templates and key parameters. The itemplate2vec and para2vec included in TPLogAD are two efficient and easy-to-implement semantic representation methods for logs, detecting anomalies in event templates and parameters respectively, which has not been achieved in previous work. Additionally, TPLogAD can avoid the interference of log diversity and dynamics on anomaly detection. Our experiments on four public log datasets show that TPLogAD outperforms existing log anomaly detection methods.

cs.LG↗

arXiv2024

Bayesian Deep Learning Approach for Real-time Lane-based Arrival Curve Reconstruction at Intersection using License Plate Recognition Data

The acquisition of real-time and accurate traffic arrival information is of vital importance for proactive traffic control systems, especially in partially connected vehicle environments. License plate recognition (LPR) data that record both vehicle departures and identities are proven to be desirable in reconstructing lane-based arrival curves in previous works. Existing LPR databased methods are predominantly designed for reconstructing historical arrival curves. For real-time reconstruction of multi-lane urban roads, it is pivotal to determine the lane choice of real-time link-based arrivals, which has not been exploited in previous studies. In this study, we propose a Bayesian deep learning approach for real-time lane-based arrival curve reconstruction, in which the lane choice patterns and uncertainties of link-based arrivals are both characterized. Specifically, the learning process is designed to effectively capture the relationship between partially observed link-based arrivals and lane-based arrivals, which can be physically interpreted as lane choice proportion. Moreover, the lane choice uncertainties are characterized using Bayesian parameter inference techniques, minimizing arrival curve reconstruction uncertainties, especially in low LPR data matching rate conditions. Real-world experiment results conducted in multiple matching rate scenarios demonstrate the superiority and necessity of lane choice modeling in reconstructing arrival curves.

cs.LG↗

arXiv2024

Instructive Code Retriever: Learn from Large Language Model's Feedback for Code Intelligence Tasks

Recent studies proposed to leverage large language models (LLMs) with In-Context Learning (ICL) to handle code intelligence tasks without fine-tuning. ICL employs task instructions and a set of examples as demonstrations to guide the model in generating accurate answers without updating its parameters. While ICL has proven effective for code intelligence tasks, its performance heavily relies on the selected examples. Previous work has achieved some success in using BM25 to retrieve examples for code intelligence tasks. However, existing approaches lack the ability to understand the semantic and structural information of queries, resulting in less helpful demonstrations. Moreover, they do not adapt well to the complex and dynamic nature of user queries in diverse domains. In this paper, we introduce a novel approach named Instructive Code Retriever (ICR), which is designed to retrieve examples that enhance model inference across various code intelligence tasks and datasets. We enable ICR to learn the semantic and structural information of the corpus by a tree-based loss function. To better understand the correlation between queries and examples, we incorporate the feedback from LLMs to guide the training of the retriever. Experimental results demonstrate that our retriever significantly outperforms state-of-the-art approaches. We evaluate our model's effectiveness on various tasks, i.e., code summarization, program synthesis, and bug fixing. Compared to previous state-of-the-art algorithms, our method achieved improvements of 50.0% and 90.0% in terms of BLEU-4 for two code summarization datasets, 74.6% CodeBLEU on program synthesis dataset, and increases of 3.6 and 3.2 BLEU-4 on two bug fixing datasets.

cs.SE↗

arXiv2024

GenesisTex2: Stable, Consistent and High-Quality Text-to-Texture Generation

Large-scale text-guided image diffusion models have shown astonishing results in text-to-image (T2I) generation. However, applying these models to synthesize textures for 3D geometries remains challenging due to the domain gap between 2D images and textures on a 3D surface. Early works that used a projecting-and-inpainting approach managed to preserve generation diversity but often resulted in noticeable artifacts and style inconsistencies. While recent methods have attempted to address these inconsistencies, they often introduce other issues, such as blurring, over-saturation, or over-smoothing. To overcome these challenges, we propose a novel text-to-texture synthesis framework that leverages pretrained diffusion models. We first introduce a local attention reweighing mechanism in the self-attention layers to guide the model in concentrating on spatial-correlated patches across different views, thereby enhancing local details while preserving cross-view consistency. Additionally, we propose a novel latent space merge pipeline, which further ensures consistency across different viewpoints without sacrificing too much diversity. Our method significantly outperforms existing state-of-the-art techniques regarding texture consistency and visual quality, while delivering results much faster than distillation-based methods. Importantly, our framework does not require additional training or fine-tuning, making it highly adaptable to a wide range of models available on public platforms.

cs.CV↗

arXiv2024

Boosting Column Generation with Graph Neural Networks for Joint Rider Trip Planning and Crew Shift Scheduling

Optimizing service schedules is pivotal to the reliable, efficient, and inclusive on-demand mobility. This pressing challenge is further exacerbated by the increasing needs of an aging population, the oversubscription of existing services, and the lack of effective solution methods. This study addresses the intricacies of service scheduling, by jointly optimizing rider trip planning and crew scheduling for a complex dynamic mobility service. The resulting optimization problems are extremely challenging computationally for state-of-the-art methods. To address this fundamental gap, this paper introduces the Joint Rider Trip Planning and Crew Shift Scheduling Problem (JRTPCSSP) and a novel solution method, called Attention and Gated GNN-Informed Column Generation (AGGNNI-CG), that hybridizes column generation and machine learning to obtain near-optimal solutions to the JRTPCSSP with real-life constraints of the application. The key idea of the machine-learning component is to dramatically reduce the number of paths to explore in the pricing problem, accelerating the most time-consuming component of the column generation. The machine learning component is a graph neural network with an attention mechanism and a gated architecture, which is particularly suited to cater for the different input sizes coming from daily operations. AGGNNI-CG has been applied to a challenging, real-world dataset from the Paratransit system of Chatham County in Georgia. It produces substantial improvements compared to the baseline column generation approach, which typically cannot produce high-quality feasible solutions in reasonable time on large-scale complex instances. AGGNNI-CG also produces significant improvements in service quality compared to the existing system.

math.OC↗

arXiv2023

A Robust Integrated Multi-Strategy Bus Control System via Deep Reinforcement Learning

An efficient urban bus control system has the potential to significantly reduce travel delays and streamline the allocation of transportation resources, thereby offering enhanced and user-friendly transit services to passengers. However, bus operation efficiency can be impacted by bus bunching. This problem is notably exacerbated when the bus system operates along a signalized corridor with unpredictable travel demand. To mitigate this challenge, we introduce a multi-strategy fusion approach for the longitudinal control of connected and automated buses. The approach is driven by a physics-informed deep reinforcement learning (DRL) algorithm and takes into account a variety of traffic conditions along urban signalized corridors. Taking advantage of connected and autonomous vehicle (CAV) technology, the proposed approach can leverage real-time information regarding bus operating conditions and road traffic environment. By integrating the aforementioned information into the DRL-based bus control framework, our designed physics-informed DRL state fusion approach and reward function efficiently embed prior physics and leverage the merits of equilibrium and consensus concepts from control theory. This integration enables the framework to learn and adapt multiple control strategies to effectively manage complex traffic conditions and fluctuating passenger demands. Three control variables, i.e., dwell time at stops, speed between stations, and signal priority, are formulated to minimize travel duration and ensure bus stability with the aim of avoiding bus bunching. We present simulation results to validate the effectiveness of the proposed approach, underlining its superior performance when subjected to sensitivity analysis, specifically considering factors such as traffic volume, desired speed, and traffic signal conditions.

eess.SY↗

arXiv2023

Revitalizing Public Transit in Low Ridership Areas: An Exploration of On-Demand Multimodal Transit Systems

Public transit plays an essential role in mitigating traffic congestion, reducing emissions, and enhancing travel accessibility and equity. One of the critical challenges in designing public transit systems is distributing finite service supplies temporally and spatially to accommodate time-varying and space-heterogeneous travel demands. Particularly, for regions with low or scattered ridership, there is a dilemma in designing traditional transit lines and corresponding service frequencies. Dense transit lines and high service frequency increase operation costs, while sparse transit lines and low service frequency result in poor accessibility and long passenger waiting time. In the coming era of Mobility-as-a-Service, the aforementioned challenge is expected to be addressed by on-demand services. In this study, we design an On-Demand Multimodel Transit System (ODMTS) for regions with low or scattered travel demands, in which some low-ridership bus lines are replaced with flexible on-demand ride-sharing shuttles. In the proposed ODMTS, riders within service regions can request shuttles to finish their trips or to connect to fixed-route services such as bus, metro, and light rail. Leveraging the integrated transportation system modeling platform, POLARIS, a simulation-based case study is conducted to assess the effectiveness of this system in Austin, Texas.

math.OC↗

arXiv2023

A Big-Data Driven Framework to Estimating Vehicle Volume based on Mobile Device Location Data

Vehicle volume serves as a critical metric and the fundamental basis for traffic signal control, transportation project prioritization, road maintenance plans and more. Traditional methods of quantifying vehicle volume rely on manual counting, video cameras, and loop detectors at a limited number of locations. These efforts require significant labor and cost for expansions. Researchers and private sector companies have also explored alternative solutions such as probe vehicle data, while still suffering from a low penetration rate. In recent years, along with the technological advancement in mobile sensors and mobile networks, Mobile Device Location Data (MDLD) have been growing dramatically in terms of the spatiotemporal coverage of the population and its mobility. This paper presents a big-data driven framework that can ingest terabytes of MDLD and estimate vehicle volume at a larger geographical area with a larger sample size. The proposed framework first employs a series of cloud-based computational algorithms to extract multimodal trajectories and trip rosters. A scalable map matching and routing algorithm is then applied to snap and route vehicle trajectories to the roadway network. The observed vehicle counts on each roadway segment are weighted and calibrated against ground truth control totals, i.e., Annual Vehicle-Miles of Travel (AVMT), and Annual Average Daily Traffic (AADT). The proposed framework is implemented on the all-street network in the state of Maryland using MDLD for the entire year of 2019. Results indicate that our proposed framework produces reliable vehicle volume estimates and also demonstrate its transferability and the generalization ability.

cs.CY↗

arXiv2022

Decentralized Vehicle Coordination: The Berkeley DeepDrive Drone Dataset and Consensus-Based Models

A significant portion of roads, particularly in densely populated developing countries, lacks explicitly defined right-of-way rules. These understructured roads pose substantial challenges for autonomous vehicle motion planning, where efficient and safe navigation relies on understanding decentralized human coordination for collision avoidance. This coordination, often termed "social driving etiquette," remains underexplored due to limited open-source empirical data and suitable modeling frameworks. In this paper, we present a novel dataset and modeling framework designed to study motion planning in these understructured environments. The dataset includes 20 aerial videos of representative scenarios, an image dataset for training vehicle detection models, and a development kit for vehicle trajectory estimation. We demonstrate that a consensus-based modeling approach can effectively explain the emergence of priority orders observed in our dataset, and is therefore a viable framework for decentralized collision avoidance planning.

cs.RO↗

arXiv2022

Pose Guided Image Generation from Misaligned Sources via Residual Flow Based Correction

Generating new images with desired properties (e.g. new view/poses) from source images has been enthusiastically pursued recently, due to its wide range of potential applications. One way to ensure high-quality generation is to use multiple sources with complementary information such as different views of the same object. However, as source images are often misaligned due to the large disparities among the camera settings, strong assumptions have been made in the past with respect to the camera(s) or/and the object in interest, limiting the application of such techniques. Therefore, we propose a new general approach which models multiple types of variations among sources, such as view angles, poses, facial expressions, in a unified framework, so that it can be employed on datasets of vastly different nature. We verify our approach on a variety of data including humans bodies, faces, city scenes and 3D objects. Both the qualitative and quantitative results demonstrate the better performance of our method than the state of the art.

cs.CV↗

arXiv2018

Intertrochanteric Fracture Visualization and Analysis Using a Map Projection Technique

Understanding intertrochanteric fracture distribution is an important topic in orthopaedics due to its high morbidity and mortality. The intertrochanteric fracture can contain high-dimensional information including complicated 3D fracture lines, which often make it difficult to visualize or to obtain valuable statistics for clinical diagnosis and prognosis applications. This paper proposed a map projection technique to map the high-dimensional information into a 2D parametric space. This method can preserve the 3D proximal femur surface and structure while visualizing the entire fracture line with a single plot/view. Using this method and a standardization technique, a total of 100 patients with different ages and genders are studied based on the original radiographs acquired by CT scan. The comparison shows that the proposed map projection representation is more efficient and rich in information visualization than the conventional heat map technique. Using the proposed method, a fracture probability can be obtained at any location in the 2D parametric space, from which the most probable fracture region can be accurately identified. The study shows that age and gender have significant influences on intertrochanteric fracture frequency and fracture line distribution.

q-bio.QM↗