Jingnan Liu

Ailing Yu, Lan Yao, Jingnan Liu et al.

Recent developments in Large Language Model (LLM)-based agents have shown impressive capabilities spanning multiple domains, exemplified by deep research systems that demonstrate superior performance on complex information-seeking and synthesis tasks. While general-purpose deep research agents have shown impressive capabilities, they struggle significantly with medical domain challenges, as evidenced by leading proprietary systems achieving limited accuracy on complex medical benchmarks. The key limitations are: (1) the model lacks sufficient dense medical knowledge for clinical reasoning, and (2) the framework is constrained by the absence of specialized retrieval tools tailored for medical contexts. We present a medical deep research agent that addresses these challenges through two core innovations. First, we develop a novel data synthesis framework using medical knowledge graphs, extracting the longest chains from subgraphs around rare medical entities to generate complex multi-hop question-answer pairs. Second, we integrate a custom-built private medical retrieval engine alongside general-purpose tools, enabling accurate medical information synthesis. Our approach generates 2100+ diverse trajectories across 12 medical specialties, each averaging 4.2 tool interactions. Through a two-stage training paradigm combining supervised fine-tuning and online reinforcement learning with composite rewards, our MedResearcher-R1-32B model demonstrates exceptional performance, establishing new state-of-the-art results on medical benchmarks while maintaining competitive performance on general deep research tasks. Our work demonstrates that strategic domain-specific innovations in architecture, tool design, and training data construction can enable smaller open-source models to outperform much larger proprietary systems in specialized domains.

Jingrong Wang, Bo Xu, Ronghe Jin et al.

Accurate, continuous, and reliable positioning is a critical component of achieving autonomous driving. However, in complex urban canyon environments, the vulnerability of a stand-alone sensor and non-line-of-sight (NLOS) caused by high buildings, trees, and elevated structures seriously affect positioning results. To address these challenges, a sky-view images segmentation algorithm based on Fully Convolutional Network (FCN) is proposed for GNSS NLOS detection. Building upon this, a novel NLOS detection and mitigation algorithm (named S-NDM) is extended to the tightly coupled Global Navigation Satellite Systems (GNSS), Inertial Measurement Units (IMU), and visual feature system which is called Sky-GVIO, with the aim of achieving continuous and accurate positioning in urban canyon environments. Furthermore, the system harmonizes Single Point Positioning (SPP) with Real-Time Kinematic (RTK) methodologies to bolster its operational versatility and resilience. In urban canyon environments, the positioning performance of S-NDM algorithm proposed in this paper is evaluated under different tightly coupled SPP-related and RTK-related models. The results exhibit that Sky-GVIO system achieves meter-level accuracy under SPP mode and sub-decimeter precision with RTK, surpassing the performance of GNSS/INS/Vision frameworks devoid of S-NDM. Additionally, the sky-view image dataset, inclusive of training and evaluation subsets, has been made publicly accessible for scholarly exploration at https://github.com/whuwangjr/sky-view-images .

Haoyang Hong, Jiajun Yin, Yuan Wang et al.

Multi-agent systems perform well on general reasoning tasks. However, the lack of training in specialized areas hinders their accuracy. Current training methods train a unified large language model (LLM) for all agents in the system. This may limit the performances due to different distributions underlying for different agents. Therefore, training multi-agent systems with distinct LLMs should be the next step to solve. However, this approach introduces optimization challenges. For example, agents operate at different frequencies, rollouts involve varying sub-agent invocations, and agents are often deployed across separate servers, disrupting end-to-end gradient flow. To address these issues, we propose M-GRPO, a hierarchical extension of Group Relative Policy Optimization designed for vertical Multi-agent systems with a main agent (planner) and multiple sub-agents (multi-turn tool executors). M-GRPO computes group-relative advantages for both main and sub-agents, maintaining hierarchical credit assignment. It also introduces a trajectory-alignment scheme that generates fixed-size batches despite variable sub-agent invocations. We deploy a decoupled training pipeline in which agents run on separate servers and exchange minimal statistics via a shared store. This enables scalable training without cross-server backpropagation. In experiments on real-world benchmarks (e.g., GAIA, XBench-DeepSearch, and WebWalkerQA), M-GRPO consistently outperforms both single-agent GRPO and multi-agent GRPO with frozen sub-agents, demonstrating improved stability and sample efficiency. These results show that aligning heterogeneous trajectories and decoupling optimization across specialized agents enhances tool-augmented reasoning tasks.

Xiuyuan Chen, Tao Sun, Dexin Su et al.

Current benchmarks for AI clinician systems, often based on multiple-choice exams or manual rubrics, fail to capture the depth, robustness, and safety required for real-world clinical practice. To address this, we introduce the GAPS framework, a multidimensional paradigm for evaluating \textbf{G}rounding (cognitive depth), \textbf{A}dequacy (answer completeness), \textbf{P}erturbation (robustness), and \textbf{S}afety. Critically, we developed a fully automated, guideline-anchored pipeline to construct a GAPS-aligned benchmark end-to-end, overcoming the scalability and subjectivity limitations of prior work. Our pipeline assembles an evidence neighborhood, creates dual graph and tree representations, and automatically generates questions across G-levels. Rubrics are synthesized by a DeepResearch agent that mimics GRADE-consistent, PICO-driven evidence review in a ReAct loop. Scoring is performed by an ensemble of large language model (LLM) judges. Validation confirmed our automated questions are high-quality and align with clinician judgment. Evaluating state-of-the-art models on the benchmark revealed key failure modes: performance degrades sharply with increased reasoning depth (G-axis), models struggle with answer completeness (A-axis), and they are highly vulnerable to adversarial perturbations (P-axis) as well as certain safety issues (S-axis). This automated, clinically-grounded approach provides a reproducible and scalable method for rigorously evaluating AI clinician systems and guiding their development toward safer, more reliable clinical practice.

Bohan Yang, Gang Liu, Yang Zhong et al.

Proton arc therapy (PAT) is an emerging and promising modality in radiotherapy, offering improved dose distribution and treatment robustness over intensity-modulated proton therapy. Yet, identifying the optimal energy layer (EL) sequence remains challenging due to the intensive computational demand and prolonged treatment delivery time. This study proposes an unsupervised deep learning model for fast EL pre-selection that minimizes EL switch (ELS) time while maintaining high plan quality. We introduce a novel data representation method, spot-count representation, which encodes the number of proton spots intersecting the target and organs at risk (OAR) in a matrix structured by sorted gantry angles and energy layers. This representation serves as the input of an U-Net style architecture, SPArc_dl, which is trained using a tri-objective function: maximizing spot-counts on target, minimizing spot-counts on OAR, and reducing ELS time. The model is evaluated on 35 nasopharyngeal cancer cases, and its performance is compared to SPArc_particle_swarm (SPArc_ps). SPArc_dl produces EL pre-selection that significantly improves both plan quality and delivery efficiency. Compared to SPArc_ps, it enhances the conformity index by 0.1 (p<0.01), reduces the homogeneity index by 0.71 (p<0.01), lowers the brainstem mean dose by 0.25 (p<0.01), and shortens the ELS time by 37.2% (p < 0.01). The results unintentionally reveal employing unchanged ELS is more time-wise efficient than descended ELS. SPArc_dl's inference time is within 1 second. However, SPArc_dl plan demonstrates limitation in robustness. The proposed spot-count representation lays a foundation for incorporating unsupervised deep learning approaches into EL pre-selection task. SPArc_dl is a fast tool for generating high-quality PAT plans by strategically pre-selecting EL to reduce delivery time while maintaining excellent dosimetric performance.

Yarong Luo, Yang Liu, Chi Guo et al.

This paper proposes a unified mathematical framework for inertial measurement unit (IMU) preintegration in inertial-aided navigation system in different frames under different motion condition. The navigation state is precisely discretized as three parts: local increment, global state, and global increment. The global increment can be calculated in different frames such as local geodetic navigation frame and earth-centered-earth-fixed frame. The local increment which is referred as the IMU preintegration can be calculated under different assumptions according to the motion of the agent and the grade of the IMU. Thus, it more accurate and more convenient for online state estimation of inertial-integrated navigation system under different environment. Furthermore, the covariance propagation based on left perturbation is proposed for the first time, which is independent of the inputs of the gyroscope and accelerometer. Finally, we show the monotonicity of the uncertainty for determinant optimality criteria and Rényi entropy optimality criteria.

Hailiang Tang, Xiaoji Niu, Tisheng Zhang et al.

Odometer has been proven to significantly improve the accuracy of the Global Navigation Satellite System / Inertial Navigation System (GNSS/INS) integrated vehicle navigation in GNSS-challenged environments. However, the odometer is inaccessible in many applications, especially for aftermarket devices. To apply forward speed aiding without hardware wheeled odometer, we propose OdoNet, an untethered one-dimensional Convolution Neural Network (CNN)-based pseudo-odometer model learning from a single Inertial Measurement Unit (IMU), which can act as an alternative to the wheeled odometer. Dedicated experiments have been conducted to verify the feasibility and robustness of the OdoNet. The results indicate that the IMU individuality, the vehicle loads, and the road conditions have little impact on the robustness and precision of the OdoNet, while the IMU biases and the mounting angles may notably ruin the OdoNet. Thus, a data-cleaning procedure is added to effectively mitigate the impacts of the IMU biases and the mounting angles. Compared to the process using only non-holonomic constraint (NHC), after employing the pseudo-odometer, the positioning error is reduced by around 68%, while the percentage is around 74% for the hardware wheeled odometer. In conclusion, the proposed OdoNet can be employed as an untethered pseudo-odometer for vehicle navigation, which can efficiently improve the accuracy and reliability of the positioning in GNSS-denied environments.

Hailiang Tang, Xiaoji Niu, Tisheng Zhang et al.

Inertial measurement unit (IMU) preintegration is widely used in factor graph optimization (FGO); e.g., in visual-inertial navigation system and global navigation satellite system/inertial navigation system (GNSS/INS) integration. However, most existing IMU preintegration models ignore the Earth's rotation and lack delicate integration processes, and these limitations severely degrade the INS accuracy. In this study, we construct a refined IMU preintegration model that incorporates the Earth's rotation, and analytically compute the covariance and Jacobian matrix. To mitigate the impact caused by sensors other than IMU in the evaluation system, FGO-based GNSS/INS integration is adopted to quantitatively evaluate the accuracy of the refined preintegration. Compared to a classic filtering-based GNSS/INS integration baseline, the employed FGO-based integration using the refined preintegration yields the same accuracy. In contrast, the existing rough preintegration yields significant accuracy degradation. The performance difference between the refined and rough preintegration models can exceed 200% for an industrial-grade MEMS module and 10% for a consumer-grade MEMS chip. Clearly, the Earth's rotation is the major factor to be considered in IMU preintegration in order to maintain the IMU precision, even for a consumer-grade IMU.

Yarong Luo, Chi Guo, Jingnan Liu

This paper proposes a equivariant filtering (EqF) framework for the inertial-integrated state estimation problem. As the kinematic system of the inertial-integrated navigation can be naturally modeling on the matrix Lie group $SE_2(3)$, the symmetry of the Lie group can be exploited to design a equivariant filter which extends the invariant extended Kalman filtering on the group affine system. Furthermore, details of the analytic state transition matrices for left invariant error and right invariant error are given.

Yarong Luo, Chi Guo, Shengyong You et al.

This paper proposes an $SE_2(3)$ based extended Kalman filtering (EKF) framework for the inertial-integrated state estimation problem. The error representation using the straight difference of two vectors in the inertial navigation system may not be reasonable as it does not take the direction difference into consideration. Therefore, we choose to use the $SE_2(3)$ matrix Lie group to represent the state of the inertial-integrated navigation system which consequently leads to the common frame error representation. With the new velocity and position error definition, we leverage the group affine dynamics with the autonomous error properties and derive the error state differential equation for the inertial-integrated navigation on the north-east-down (NED) navigation frame and the earth-centered earth-fixed (ECEF) frame, respectively, the corresponding EKF, terms as $SE_2(3)$ based EKF has also been derived. It provides a new perspective on the geometric EKF with a more sophisticated formula for the inertial-integrated navigation system. Furthermore, we design two new modified error dynamics on the NED frame and the ECEF frame respectively by introducing new auxiliary vectors. Finally the equivalence of the left-invariant EKF and left $SE_2(3)$ based EKF have been shown in navigation frame and ECEF frame.