arXiv subjects

Jiawei Tang

Publications and source records attributed to Jiawei Tang.

At least 19 recordsLinked to original sources

arXiv2026

TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents

Document parsing aims to transform unstructured documents into structured and machine-readable representations. Recent advances in Vision-Language Models (VLMs) have significantly advanced document parsing. However, existing approaches still face two major challenges. First, decoupled VLM-based methods heavily rely on accurate layout analysis, where geometric distortions in camera-captured documents can introduce cascading errors. Second, although end-to-end VLM-based methods alleviate the dependence on explicit layout detection, they often suffer from redundant generation, hallucinations, and insufficient structural reasoning in high-resolution scenarios. To address these challenges, we propose TeleOCR, a unified framework for document parsing. TeleOCR introduces deformation-aware learning to incorporate geometric perception into VLMs and proposes an adaptive sampling mechanism for complex layout representation. Furthermore, a content-structure decoupled learning strategy is developed to explicitly model formula grammars and table structures, enabling more effective structured representation learning. Extensive experiments demonstrate that TeleOCR achieves state-of-the-art performance across diverse document parsing benchmarks. It obtains overall scores of 96.87, 88.53 and 78.41 on OmniDocBench v1.6, Wild-OmniDocBench, and PureDocBench, respectively, and ranks first in the ICDAR 2026 Sci-ImageMiner Challenge. These results validate the effectiveness and generalization capability of TeleOCR in complex document parsing scenarios.

cs.CV↗

arXiv2026

Variational Inference for Evidential Deep Learning

While Deep Neural Networks (DNNs) achieve remarkable performance, their tendency to produce overconfident predictions. Evidential Deep Learning (EDL) mitigates this by formulating predictions as a Dirichlet distribution over class probabilities to explicitly quantify epistemic uncertainty. However, we found that the conventional EDL suffers from two fundamental limitations: a Kullback-Leibler (KL) penalty that only suppresses the evidence of negative classes, producing excessively high evidence therefore decreasing the model's ability to quantify uncertainty, and an absence in theoretical guarantee of setting Dirichlet parameter $\alpha=e+1$. In this paper, we propose a mathematically principled framework, Variational Inference Evidential Deep Learning (VI-EDL). By reformulating evidential learning through the lens of variational inference, we derive an Evidence Lower Bound (ELBO), which prevents the evidence from growing excessively. Theoretically, we rigorously establish a generalization bound and reveal how the predicted uncertainty, feature and network complexity affect this bound, and why setting $\boldsymbol{\alpha} = \mathbf{e} + \mathbf{1}$ can minimize it. Extensive experiments on standard visual and medical datasets demonstrate that VI-EDL achieves state-of-the-art performance, showing excellent performance in out-of-distribution detection, noise detection and autonomous driving scenario. The code is available in https://github.com/seutjw/VI-EDL.

cs.LG↗

arXiv2025

Towards Better IncomLDL: We Are Unaware of Hidden Labels in Advance

Label distribution learning (LDL) is a novel paradigm that describe the samples by label distribution of a sample. However, acquiring LDL dataset is costly and time-consuming, which leads to the birth of incomplete label distribution learning (IncomLDL). All the previous IncomLDL methods set the description degrees of "missing" labels in an instance to 0, but remains those of other labels unchanged. This setting is unrealistic because when certain labels are missing, the degrees of the remaining labels will increase accordingly. We fix this unrealistic setting in IncomLDL and raise a new problem: LDL with hidden labels (HidLDL), which aims to recover a complete label distribution from a real-world incomplete label distribution where certain labels in an instance are omitted during annotation. To solve this challenging problem, we discover the significance of proportional information of the observed labels and capture it by an innovative constraint to utilize it during the optimization process. We simultaneously use local feature similarity and the global low-rank structure to reveal the mysterious veil of hidden labels. Moreover, we theoretically give the recovery bound of our method, proving the feasibility of our method in learning from hidden labels. Extensive recovery and predictive experiments on various datasets prove the superiority of our method to state-of-the-art LDL and IncomLDL methods.

cs.LG↗

arXiv2025

AliBoost: Ecological Boosting Framework in Alibaba Platform

Maintaining a healthy ecosystem in billion-scale online platforms is challenging, as users naturally gravitate toward popular items, leaving cold and less-explored items behind. This ''rich-get-richer'' phenomenon hinders the growth of potentially valuable cold items and harms the platform's ecosystem. Existing cold-start models primarily focus on improving initial recommendation performance for cold items but fail to address users' natural preference for popular content. In this paper, we introduce AliBoost, Alibaba's ecological boosting framework, designed to complement user-oriented natural recommendations and foster a healthier ecosystem. AliBoost incorporates a tiered boosting structure and boosting principles to ensure high-potential items quickly gain exposure while minimizing disruption to low-potential items. To achieve this, we propose the Stacking Fine-Tuning Cold Predictor to enhance the foundation CTR model's performance on cold items for accurate CTR and potential prediction. AliBoost then employs an Item-oriented Bidding Boosting mechanism to deliver cold items to the most suitable users while balancing boosting speed with user-personalized preferences. Over the past six months, AliBoost has been deployed across Alibaba's mainstream platforms, successfully cold-starting over a billion new items and increasing both clicks and GMV of cold items by over 60% within 180 days. Extensive online analysis and A/B testing demonstrate the effectiveness of AliBoost in addressing ecological challenges, offering new insights into the design of billion-scale recommender systems.

cs.IR↗

arXiv2025

Concentration Distribution Learning from Label Distributions

Label distribution learning (LDL) is an effective method to predict the relative label description degree (a.k.a. label distribution) of a sample. However, the label distribution is not a complete representation of an instance because it overlooks the absolute intensity of each label. Specifically, it's impossible to obtain the total description degree of hidden labels that not in the label space, which leads to the loss of information and confusion in instances. To solve the above problem, we come up with a new concept named background concentration to serve as the absolute description degree term of the label distribution and introduce it into the LDL process, forming the improved paradigm of concentration distribution learning. Moreover, we propose a novel model by probabilistic methods and neural networks to learn label distributions and background concentrations from existing LDL datasets. Extensive experiments prove that the proposed approach is able to extract background concentrations from label distributions while producing more accurate prediction results than the state-of-the-art LDL methods. The code is available in https://github.com/seutjw/CDL-LD.

cs.LG↗

arXiv2025

Lie-algebra Adaptive Tracking Control for Rigid Body Dynamics

Adaptive tracking control for rigid body dynamics is of critical importance in control and robotics, particularly for addressing uncertainties or variations in system model parameters. However, most existing adaptive control methods are designed for systems with states in vector spaces, often neglecting the manifold constraints inherent to robotic systems. In this work, we propose a novel Lie-algebra-based adaptive control method that leverages the intrinsic relationship between the special Euclidean group and its associated Lie algebra. By transforming the state space from the group manifold to a vector space, we derive a linear error dynamics model that decouples model parameters from the system state. This formulation enables the development of an adaptive optimal control method that is both geometrically consistent and computationally efficient. Extensive simulations demonstrate the effectiveness and efficiency of the proposed method. We have made our source code publicly available to the community to support further research and collaboration.

cs.RO↗

arXiv2024

Breaking the Midas Spell:Understanding Progressive Novice-AI Collaboration in Spatial Design

In spatial design, Artificial Intelligence (AI) tools often generate the entire spatial design outcome in a single automated step, rather than engaging users in a deepening and iterative process. This significantly reduces users' involvement, learning, and creative capabilities, leading to a superficial understanding of spatial design. We conducted a Wizard-of-Oz study, where Novices and AI (acted by experimenters) worked together to finish spatial design tasks using various AI models. We identified typical function and workflow patterns adopted by the participants, leading to the understanding of the opportunities and challenges in the human-AI co-creation process. Based on insights gathered from this research, we proposed some design implications of the novice-AI collaboration system that aims to democratize spatial design through a progressive, iterative co-creation process.

cs.HC↗

arXiv2024

Bias-VarianceTrade-off in Kalman Filter-Based Disturbance Observers

The performance of disturbance observers is strongly influenced by the level of prior knowledge about the disturbance model. The simultaneous input and state estimation (SISE) algorithm is widely recognized for providing unbiased minimum-variance estimates under arbitrary disturbance models. In contrast, the Kalman filter-based disturbance observer (KF-DOB) achieves minimum mean-square error estimation when the disturbance model is fully specified. However, practical scenarios often fall between these extremes, where only partial knowledge of the disturbance model is available. This paper investigates the inherent bias-variance trade-off in KF-DOB when the disturbance model is incomplete. We further show that SISE can be interpreted as a special case of KF-DOB, where the disturbance noise covariance tends to infinity. To address this trade-off, we propose two novel estimators: the multi-kernel correntropy Kalman filter-based disturbance observer (MKCKF-DOB) and the interacting multiple models Kalman filter-based disturbance observer (IMMKF-DOB). Simulations verify the effectiveness of the proposed methods.

eess.SP↗

arXiv2024

Leveraging Inter-Chunk Interactions for Enhanced Retrieval in Large Language Model-Based Question Answering

Retrieving external knowledge and prompting large language models with relevant information is an effective paradigm to enhance the performance of question-answering tasks. Previous research typically handles paragraphs from external documents in isolation, resulting in a lack of context and ambiguous references, particularly in multi-document and complex tasks. To overcome these challenges, we propose a new retrieval framework IIER, that leverages Inter-chunk Interactions to Enhance Retrieval. This framework captures the internal connections between document chunks by considering three types of interactions: structural, keyword, and semantic. We then construct a unified Chunk-Interaction Graph to represent all external documents comprehensively. Additionally, we design a graph-based evidence chain retriever that utilizes previous paths and chunk interactions to guide the retrieval process. It identifies multiple seed nodes based on the target question and iteratively searches for relevant chunks to gather supporting evidence. This retrieval process refines the context and reasoning chain, aiding the large language model in reasoning and answer generation. Extensive experiments demonstrate that IIER outperforms strong baselines across four datasets, highlighting its effectiveness in improving retrieval and reasoning capabilities.

cs.CL↗

arXiv2024

MINER-RRT*: A Hierarchical and Fast Trajectory Planning Framework in 3D Cluttered Environments

Trajectory planning for quadrotors in cluttered environments has been challenging in recent years. While many trajectory planning frameworks have been successful, there still exists potential for improvements, particularly in enhancing the speed of generating efficient trajectories. In this paper, we present a novel hierarchical trajectory planning framework to reduce computational time and memory usage called MINER-RRT*, which consists of two main components. First, we propose a sampling-based path planning method boosted by neural networks, where the predicted heuristic region accelerates the convergence of rapidly-exploring random trees. Second, we utilize the optimal conditions derived from the quadrotor's differential flatness properties to construct polynomial trajectories that minimize control effort in multiple stages. Extensive simulation and real-world experimental results demonstrate that, compared to several state-of-the-art (SOTA) approaches, our method can generate high-quality trajectories with better performance in 3D cluttered environments.

cs.RO↗

arXiv2024

Merged ChemProt-DrugProt for Relation Extraction from Biomedical Literature

The extraction of chemical-gene relations plays a pivotal role in understanding the intricate interactions between chemical compounds and genes, with significant implications for drug discovery, disease understanding, and biomedical research. This paper presents a data set created by merging the ChemProt and DrugProt datasets to augment sample counts and improve model accuracy. We evaluate the merged dataset using two state of the art relationship extraction algorithms: Bidirectional Encoder Representations from Transformers (BERT) specifically BioBERT, and Graph Convolutional Networks (GCNs) combined with BioBERT. While BioBERT excels at capturing local contexts, it may benefit from incorporating global information essential for understanding chemical-gene interactions. This can be achieved by integrating GCNs with BioBERT to harness both global and local context. Our results show that by integrating the ChemProt and DrugProt datasets, we demonstrated significant improvements in model performance, particularly in CPR groups shared between the datasets. Incorporating the global context using GCN can help increase the overall precision and recall in some of the CPR groups over using just BioBERT.

cs.CL↗

arXiv2024

GMPC: Geometric Model Predictive Control for Wheeled Mobile Robot Trajectory Tracking

The configuration of most robotic systems lies in continuous transformation groups. However, in mobile robot trajectory tracking, many recent works still naively utilize optimization methods for elements in vector space without considering the manifold constraint of the robot configuration. In this letter, we propose a geometric model predictive control (MPC) framework for wheeled mobile robot trajectory tracking. We first derive the error dynamics of the wheeled mobile robot trajectory tracking by considering its manifold constraint and kinematic constraint simultaneously. After that, we utilize the relationship between the Lie group and Lie algebra to convexify the tracking control problem, which enables us to solve the problem efficiently. Thanks to the Lie group formulation, our method tracks the trajectory more smoothly than existing nonlinear MPC. Simulations and physical experiments verify the effectiveness of our proposed methods. Our pure Python-based simulation platform is publicly available to benefit further research in the community.

eess.SY↗

arXiv2024

Direct Shooting Method for Numerical Optimal Control: A Modified Transcription Approach

Direct shooting is an efficient method to solve numerical optimal control. It utilizes the Runge-Kutta scheme to discretize a continuous-time optimal control problem making the problem solvable by nonlinear programming solvers. However, conventional direct shooting raises a contradictory dynamics issue when using an augmented state to handle {high-order} systems. This paper fills the research gap by considering the direct shooting method for {high-order} systems. We derive the modified Euler and Runge-Kutta-4 methods to transcribe the system dynamics constraint directly. Additionally, we provide the global error upper bounds of our proposed methods. A set of benchmark optimal control problems shows that our methods provide more accurate solutions than existing approaches.

eess.SY↗

arXiv2023

KnowledgeNavigator: Leveraging Large Language Models for Enhanced Reasoning over Knowledge Graph

Large language model (LLM) has achieved outstanding performance on various downstream tasks with its powerful natural language understanding and zero-shot capability, but LLM still suffers from knowledge limitation. Especially in scenarios that require long logical chains or complex reasoning, the hallucination and knowledge limitation of LLM limit its performance in question answering (QA). In this paper, we propose a novel framework KnowledgeNavigator to address these challenges by efficiently and accurately retrieving external knowledge from knowledge graph and using it as a key factor to enhance LLM reasoning. Specifically, KnowledgeNavigator first mines and enhances the potential constraints of the given question to guide the reasoning. Then it retrieves and filters external knowledge that supports answering through iterative reasoning on knowledge graph with the guidance of LLM and the question. Finally, KnowledgeNavigator constructs the structured knowledge into effective prompts that are friendly to LLM to help its reasoning. We evaluate KnowledgeNavigator on multiple public KGQA benchmarks, the experiments show the framework has great effectiveness and generalization, outperforming previous knowledge graph enhanced LLM methods and is comparable to the fully supervised models.

cs.CL↗

arXiv2023

Label Distribution Learning from Logical Label

Label distribution learning (LDL) is an effective method to predict the label description degree (a.k.a. label distribution) of a sample. However, annotating label distribution (LD) for training samples is extremely costly. So recent studies often first use label enhancement (LE) to generate the estimated label distribution from the logical label and then apply external LDL algorithms on the recovered label distribution to predict the label distribution for unseen samples. But this step-wise manner overlooks the possible connections between LE and LDL. Moreover, the existing LE approaches may assign some description degrees to invalid labels. To solve the above problems, we propose a novel method to learn an LDL model directly from the logical label, which unifies LE and LDL into a joint model, and avoids the drawbacks of the previous LE methods. Extensive experiments on various datasets prove that the proposed approach can construct a reliable LDL model directly from the logical label, and produce more accurate label distribution than the state-of-the-art LE methods.

cs.LG↗

arXiv2021

nvBench: A Large-Scale Synthesized Dataset for Cross-Domain Natural Language to Visualization Task

NL2VIS - which translates natural language (NL) queries to corresponding visualizations (VIS) - has attracted more and more attention both in commercial visualization vendors and academic researchers. In the last few years, the advanced deep learning-based models have achieved human-like abilities in many natural language processing (NLP) tasks, which clearly tells us that the deep learning-based technique is a good choice to push the field of NL2VIS. However, a big balk is the lack of benchmarks with lots of (NL, VIS) pairs. We present nvBench, the first large-scale NL2VIS benchmark, containing 25,750 (NL, VIS) pairs from 750 tables over 105 domains, synthesized from (NL, SQL) benchmarks to support cross-domain NL2VIS task. The quality of nvBench has been extensively validated by 23 experts and 300+ crowd workers. Deep learning-based models training using nvBench demonstrate that nvBench can push the field of NL2VIS.

cs.HC↗

arXiv2020

Analyzing the Design Space of Re-opening Policies and COVID-19 Outcomes in the US

Recent re-opening policies in the US, following a period of social distancing measures, introduced a significant increase in daily COVID-19 infections, calling for a roll-back or substantial revisiting of these policies in many states. The situation is suggestive of difficulties modeling the impact of partial distancing/re-opening policies on future epidemic spread for purposes of choosing safe alternatives. More specifically, one needs to understand the impact of manipulating the availability of social interaction venues (e.g., schools, workplaces, and retail establishments) on virus spread. We introduce a model, inspired by social networks research, that answers the above question. Our model compartmentalizes interaction venues into categories we call mixing domains, enabling one to predict COVID-19 contagion trends in different geographic regions under different what if assumptions on partial re-opening of individual domains. We apply our model to several highly impacted states showing (i) how accurately it predicts the extent of current resurgence (from available policy descriptions), and (ii) what alternatives might be more effective at mitigating the second wave. We further compare policies that rely on partial venue closure to policies that espouse wide-spread periodic testing instead (i.e., in lieu of social distancing). Our models predict that the benefits of (mandatory) testing out-shadow the benefits of partial venue closure, suggesting that perhaps more efforts should be directed to such a mitigation strategy.

physics.soc-ph↗

arXiv2019

Custom Edge-Element FEM Solver and its Application to Eddy-Current Simulation of Realistic 2M-Element Human Brain Phantom

Extensive research papers of three-dimensional computational techniques are widely used for the investigation of human brain pathophysiology. Eddy current analyzing could provide an indication of conductivity change within a biological body. A significant obstacle to current trend analyses is the development of a numerically stable and efficiency-finite element scheme that performs well at low frequency and does not require a large number of degrees of freedom. Here, a custom finite element method (FEM) solver based on edge elements is proposed using the weakly coupled theory, which separates the solution into two steps. First, the background field (the magnetic vector potential on each edge) is calculated and stored. Then, the electric scalar potential on each node is obtained by FEM based on Galerkin formulations. Consequently, the electric field and eddy current distribution in the object can be obtained. This solver is more efficient than typical commercial solvers since it reduces the vector eddy current equation to a scalar one, and reduces the meshing domain to just the eddy current region. It can therefore tackle complex eddy current calculations for models with much larger numbers of elements, such as those encountered in eddy current computation in biological tissues. An example is presented with a realistic human brain mesh of 2 million elements. In addition, with this solver, the equivalent magnetic field induced from the excitation coil is applied, and therefore there is no need to mesh the excitation coil. In combination, these significantly increase the efficiency of the solver.

physics.med-ph↗