arXiv subjects

Mo Yang

Publications and source records attributed to Mo Yang.

At least 19 recordsLinked to original sources

arXiv2026

SPOT-E: Test-Time Entropy Shaping with Visual Spotlights for Frozen VLMs

Vision-language models (VLMs) often underperform on evidence intensive tasks because decisive visual evidence are small, localized, and easy to overlook, leading to failures in evidence readout even when high-level reasoning is intact. Prior inference-time visual interventions can improve grounding without retraining, but they are largely open-loop and lack a mechanism to verify whether highlighted evidence is actually used. We study answer-span prediction entropy as a model-internal feedback signal and show that naive entropy minimization is ambiguous, since low entropy may arise from evidence-grounded confidence or shortcut collapse. To resolve this ambiguity, we introduce low-entropy anchors and an entropy-shaping objective that reduces answer uncertainty while preserving baseline high-confidence tokens. We instantiate this principle in SPOT-E, a plug-and-play test-time method that produces question-conditioned spotlights, optimized per instance via light-weight tuning based on Group Relative Policy Optimization (GRPO). Across all benchmarks and different VLM families, SPOT-E yields consistent gains and improved robustness under visual corruptions. Code is publicly available at: https://github.com/YinBo0927/SPOT-E

cs.CV↗

arXiv2026

From Trajectories to Phenotypes: Disease Progression as Structural Priors for Multi-organ Imaging Representation Learning

Imaging-derived phenotypes (IDPs) summarize multi-organ physiology but provide only static snapshots of diseases that evolve over time. In contrast, longitudinal electronic health records encode disease trajectories through temporal dependencies among past diagnosis events and comorbidity structure. We hypothesize that IDPs and disease trajectories contain partially shared disease-relevant structure. We propose a trajectory-aware distillation framework that transfers structural knowledge from a generative disease trajectory Transformer into an organ-wise IDP encoder. A population-scale trajectory model trained on longitudinal diagnosis sequences produces subject-level embeddings that supervise IDP representation learning via geometry-preserving alignment. During downstream prediction, trajectory and imaging representations can also be fused via cross-attention. Across 159 diseases in the UK Biobank cohort, trajectory-aware pretraining consistently improves both discrimination (AUC) and time-to-onset prediction (MAE), with the largest gains for low-prevalence diseases. Similarity relationships in IDP embedding space also align with those in trajectory space, providing supportive evidence for partially aligned representation geometry. These results suggest that population-scale generative disease models can serve as structural priors for data-limited imaging modalities, improving robustness under realistic cohort constraints.

cs.IR↗

arXiv2026

AD-CARE: A Guideline-grounded, Modality-agnostic LLM Agent for Real-world Alzheimer's Disease Diagnosis with Multi-cohort Assessment, Fairness Analysis, and Reader Study

Alzheimer's disease (AD) is a growing global health challenge as populations age, and timely, accurate diagnosis is essential to reduce individual and societal burden. However, real-world AD assessment is hampered by incomplete, heterogeneous multimodal data and variability across sites and patient demographics. Although large language models (LLMs) have shown promise in biomedicine, their use in AD has largely been confined to answering narrow, disease-specific questions rather than generating comprehensive diagnostic reports that support clinical decision-making. Here we expand LLM capabilities for clinical decision support by introducing AD-CARE, a modality-agnostic agent that performs guideline-grounded diagnostic assessment from incomplete, heterogeneous inputs without imputing missing modalities. By dynamically orchestrating specialized diagnostic tools and embedding clinical guidelines into LLM-driven reasoning, AD-CARE generates transparent, report-style outputs aligned with real-world clinical workflows. Across six cohorts comprising 10,303 cases, AD-CARE achieved 84.9% diagnostic accuracy, delivering 4.2%-13.7% relative improvements over baseline methods. Despite cohort-level differences, dataset-specific accuracies remain robust (80.4%-98.8%), and the agent consistently outperforms all baselines. AD-CARE reduced performance disparities across racial and age subgroups, decreasing the average dispersion of four metrics by 21%-68% and 28%-51%, respectively. In a controlled reader study, the agent improved neurologist and radiologist accuracy by 6%-11% and more than halved decision time. The framework yielded 2.29%-10.66% absolute gains over eight backbone LLMs and converges their performance. These results show that AD-CARE is a scalable, practically deployable framework that can be integrated into routine clinical workflows for multimodal decision support in AD.

cs.MA↗

arXiv2026

SWE-Pruner: Self-Adaptive Context Pruning for Coding Agents

LLM agents have demonstrated remarkable capabilities in software development, but their performance is hampered by long interaction contexts, which incur high API costs and latency. While various context compression approaches such as LongLLMLingua have emerged to tackle this challenge, they typically rely on fixed metrics such as PPL, ignoring the task-specific nature of code understanding. As a result, they frequently disrupt syntactic and logical structure and fail to retain critical implementation details. In this paper, we propose SWE-Pruner, a self-adaptive context pruning framework tailored for coding agents. Drawing inspiration from how human programmers "selectively skim" source code during development and debugging, SWE-Pruner performs task-aware adaptive pruning for long contexts. Given the current task, the agent formulates an explicit goal (e.g., "focus on error handling") as a hint to guide the pruning targets. A lightweight neural skimmer (0.6B parameters) is trained to dynamically select relevant lines from the surrounding context given the goal. Evaluations across four benchmarks and multiple models validate SWE-Pruner's effectiveness in various scenarios, achieving 23-54% token reduction on agent tasks like SWE-Bench Verified while even improving success rates, and up to 14.84x compression on single-turn tasks like LongCodeQA with minimal performance impact.

cs.SE↗

arXiv2026

Reasoning in Trees: Improving Retrieval-Augmented Generation for Multi-Hop Question Answering

Retrieval-Augmented Generation (RAG) has demonstrated significant effectiveness in enhancing large language models (LLMs) for complex multi-hop question answering (QA). For multi-hop QA tasks, current iterative approaches predominantly rely on LLMs to self-guide and plan multi-step exploration paths during retrieval, leading to substantial challenges in maintaining reasoning coherence across steps from inaccurate query decomposition and error propagation. To address these issues, we introduce Reasoning Tree Guided RAG (RT-RAG), a novel hierarchical framework for complex multi-hop QA. RT-RAG systematically decomposes multi-hop questions into explicit reasoning trees, minimizing inaccurate decomposition through structured entity analysis and consensus-based tree selection that clearly separates core queries, known entities, and unknown entities. Subsequently, a bottom-up traversal strategy employs iterative query rewriting and refinement to collect high-quality evidence, thereby mitigating error propagation. Comprehensive experiments show that RT-RAG substantially outperforms state-of-the-art methods by 7.0% F1 and 6.0% EM, demonstrating the effectiveness of RT-RAG in complex multi-hop QA.

cs.CL↗

arXiv2025

Safe Control of Multi-Agent Systems with Minimal Communication

In many multi-agent systems, communication is limited by bandwidth, latency, and energy constraints. Designing controllers that achieve coordination and safety with minimal communication is critical for scalable and reliable deployment. This paper presents a method for designing controllers that minimize inter-agent communication in multi-agent systems while satisfying safety and coordination requirements, while conforming to communication delay constraints. The control synthesis problem is cast as a rank minimization problem, where a convex relaxation is obtained via system level synthesis. Simulation results on various tasks, including trajectory tracking with relative and heterogeneous sensing, demonstrate that the proposed method significantly reduces inter-agent transmission compared to baseline approaches.

eess.SY↗

arXiv2024

Human Balancing on a Log: A Switched Multi-Layer Controller

We study the task of balancing a human on a log that is fixed in place. Balancing on a log is substantially more challenging than balancing on a flat surface due to increased instability -- nonetheless, we are able to balance by composing simple (e.g., PID, LQR) controllers in a bio-inspired switched multi-layer configuration. The controller consists of an upper-layer LQR planner (akin to the central nervous system) that coordinates ankle and hip torques, and lower-layer PID trackers (akin to local motor units) that follow this plan subject to nonlinear dynamics. The controller switches between three operational modes depending on the state of the human. The efficacy of the controller is verified in simulation, where our controller is able to stabilize the human for a variety of initial conditions and disturbances. We also introduce a controller that outputs muscle activations to perform the same balancing task.

eess.SY↗

arXiv2022

Asymmetric phenomenon of flow and heat transfer in charging process of thermal energy storage based on an entire domain model

Phase change energy storage is getting increasing attention as a representative technology to achieve carbon neutrality. The phase change process exists typical phenomenon of asymmetry that affects the energy storage performance. However, the mechanism of asymmetry is currently lack of elaboration because the half domain model is always used to simplify the numerical simulation and avoid the appearance of asymmetry. In this study, the entire domain model and boundary conditions were adopted and numerically simulated for the melting process of paraffin wax, i.e., the charging process of energy storage. The nonlinear dynamics method was applied to explain the asymmetric flow and heat transfer phenomenon. The three-stage characteristic based on charging speed was proposed for charging process and was explained by thermal conduction or natural convection. The asymmetric phenomenon was elaborated on the cause of formation, change mechanism, and effect evaluation. Results show that natural convection accounts for the multiple solutions (including the asymmetric ones) of charging process. It is also pointed out that asymmetric solutions can exist under symmetric geometry structure. To accurately solve the charging process, it is necessary to use the entire domain model. Both the charging speed and energy storage capacity have a positive correlation with asymmetry. Thus, a potential idea for energy storage application was proposed to increase the charging speed; that is, adding thermal disturbance during the melting process to destroy flow stability and to enhance convective heat transfer. The research methods and findings will support the development of energy storage and numerical investigation in other related areas.

physics.flu-dyn↗

arXiv2021

Dimming Down LED: An Open-source Threshold Implementation on Light Encryption Device (LED) Block Cipher

Lightweight block ciphers have been widely used in applications such as RFID tags, IoTs, and network sensors. Among them, with comparable parameters, the Light Encryption Device (LED) block cipher achieves the smallest area. However, implementation of encryption algorithms manifest side-channel leakage, therefore, it is crucial to protect their design against side-channel analyses. In this paper, we present a threshold implementation of the LED cipher which has 64-bit data input and 128-bit key. The presented design splits secret information among multiple shares to achieve a higher security level. We demonstrate that our implementation can protect against first-order power side-channel attacks. As a cost, the design area is almost doubled and the maximum operating frequency is degraded by 30%. To make our design verifiable, we have also open-sourced our design online.

cs.CR↗

arXiv2019

Tropical Cyclone Track Forecasting using Fused Deep Learning from Aligned Reanalysis Data

The forecast of tropical cyclone trajectories is crucial for the protection of people and property. Although forecast dynamical models can provide high-precision short-term forecasts, they are computationally demanding, and current statistical forecasting models have much room for improvement given that the database of past hurricanes is constantly growing. Machine learning methods, that can capture non-linearities and complex relations, have only been scarcely tested for this application. We propose a neural network model fusing past trajectory data and reanalysis atmospheric images (wind and pressure 3D fields). We use a moving frame of reference that follows the storm center for the 24h tracking forecast. The network is trained to estimate the longitude and latitude displacement of tropical cyclones and depressions from a large database from both hemispheres (more than 3000 storms since 1979, sampled at a 6 hour frequency). The advantage of the fused network is demonstrated and a comparison with current forecast models shows that deep learning methods could provide a valuable and complementary prediction. Moreover, our method can give a forecast for a new storm in a few seconds, which is an important asset for real-time forecasts compared to traditional forecasts.

physics.ao-ph↗

arXiv2019

Hybrid LBM-FVM and LBM-MCM Methods for Fluid Flow and Heat Transfer Simulation

The fluid flow and heat transfer problems encountered in industry applications span into different scales and there are different numerical methods for different scales problems. It is not possible to use single scale method to solve problems involving multiple scales. Multiscale methods are needed to solve problems involving multiple scales. In this chapter, meso-macro-multiscale methods are developed by combining various single scale numerical methods, including lattice Boltzmann method (LBM), finite volume method (FVM) and Monte Carlo method (MCM). Macroscale methods include FVM, while LBM and MCM belongs to mesoscale methods. Two strategies exist in combing these numerical methods. For the first one, the whole domain is divided into multiple subdomains and different domains use various numerical methods. Message passing among subdomains decides the accuracy of this type of multiscale numerical method. For the second one, various parameters are solved with different numerical methods. These two types of multiscale methods are both discussed in this chapter.

cs.CE↗

arXiv2018

Dynamical Model of Binary Asteroid Systems Using Binary Octahedrons

We used binary octahedrons to investigate the dynamical behaviors of binary asteroid systems. The mutual potential of the binary polyhedron method is derived from the fourth order to the sixth order. The irregular shapes, relative orbits, attitude angles, as well as the angular velocities of the binary asteroid system are included in the model. We investigated the relative trajectory of the secondary relative to the primary, the total angular momentum and total energy of the system, the three-axis attitude angular velocity of the binary system, as well as the angular momentum of the two components. The relative errors of total angular momentum and total energy indicate the calculation has a high precision. It is found that the influence of the orbital and attitude motion of the primary from the gravitational force of the secondary is obvious. This study is useful for understanding the complicated dynamical behaviors of the binary asteroid systems discovered in our Solar system.

astro-ph.EP↗

arXiv2016

Lattice Boltzmann Method Simulation of 3-D Melting Using Double MRT Model with Interfacial Tracking Method

Three-dimensional melting problems are investigated numerically with Lattice Boltzmann method (LBM). Regarding algorithm's accuracy and stability, Multiple-Relaxation-Time (MRT) models are employed to simplify the collision term in LBM. Temperature and velocity fields are solved with double distribution functions, respectively. 3-D melting problems are solved with double MRT models for the first time in this article. The key point for the numerical simulation of a melting problem is the methods to obtain the location of the melting front and this article uses interfacial tracking method. The interfacial tracking method combines advantages of both deforming and fixed grid approaches. The location of the melting front was obtained by calculating the energy balance at the solid-liquid interface. Various 3-D conduction controlled melting problems are solved firstly to verify the numerical method. Liquid fraction tendency and temperature distribution obtained from numerical methods agree with the analytical results well. The proposed double MRT model with interfacial tracking method is valid to solve 3-D melting problems. Different 3-D convection controlled melting problems are then solved with the proposed numerical method. Various locations of the heat surface have different melting front moving velocities, due to the natural convection effects. Rayleigh number's effects to the 3-D melting process is discussed.

physics.flu-dyn↗

arXiv2016

Effects of Slotted Structures on Nonlinear Characteristics of Natural Convection in a Cylinder with an Internal Concentric Slotted Annulus

Natural convection in a cylinder with an internally slotted annulus was solved by SIMPLE algorithm, and the effects of different slotted structures on nonlinear characteristics of natural convection were investigated. The results show that the equivalent thermal conductivity Keq increases with Rayleigh number, and reaches the maximum in the vertical orientation. Nonlinear results were obtained by simulating the fluid flow at different conditions. With increasing Rayleigh number, heat transfer is intensified and the state of heat transfer changes from the steady to unsteady. We investigated different slotted structures effects on natural convection, and analyze the corresponding nonlinear characteristics.

physics.flu-dyn↗

arXiv2016

Oscillatory Double-Diffusive Convection in a Horizontal Cavity with Soret and Dufour Effects

Oscillatory double-diffusive convection in horizontal cavity with Soret and Dufour effects is investigated numerically based on SIMPLE algorithm with QUICK scheme in non-uniform staggered grid system. The results show that double-diffusive convection develops from steady-state convection-dominated, periodic oscillatory, quasi-periodic oscillatory to chaotic flow, and finally return to periodic oscillation as buoyancy ratio increases. Moreover, fundamental frequency and fluctuation amplitude increase with buoyancy ratio. As Rayleigh number increases, transition trendy of oscillatory convection is similar to that of buoyancy ratio. But the return of periodic oscillation from chaos is not obtained as Rayleigh number increases. As aspect ratio decreases, the oscillatory convection evolves from periodic into steady-state. In addition, fundamental frequency increases at first and then decreases while fluctuation amplitude decreases with aspect ratio.

physics.comp-ph↗

arXiv2016

A multiscale nonequilibrium model for melting of metal powder bed subjected to constant heat flux

A multiscale numerical model based on nonequilibrium thermal effect for melting of metal powder bed subjected to constant heat flux is developed. The volume shrinkage due to density change is taken into account. The nonequilibrium model is discretized by an implicit finite difference method and solved numerically using an iterative tri-diagonal matrix algorithm. The evolutions of powder bed surface temperature and various interfacial locations as well as the melting temperature range during the melting process are investigated. The results show that liquid region, upper and lower parts of mushy zone are formed on the top of unsintered zone as the melting progresses. The duration of the preheating stage shortens and the melting rate accelerates as the initial porosity or initial temperature increases while particle size has much less effect on the melting process. The parametric study shows the melting temperature range of the powder bed widens with increasing initial porosity, decreasing initial temperature or increasing particle size.

cond-mat.mtrl-sci↗

arXiv2016

Structural, Dynamic, and Vibrational Properties during Heat Transfer in Si/Ge Superlattices: A Car-Parrinello Molecular Dynamics Study

The structural, dynamic, and vibrational properties during the heat transfer process in Si/Ge superlattices, are studied by analyzing the trajectories generated by the ab initio Car-Parrinello molecular dynamics simulation. The radial distribution functions and mean square displacements are calculated and further discussions are made to explain and probe the structural changes relating to the heat transfer phenomenon. Furthermore, the vibrational density of states of the two layers (Si/Ge) are computed and plotted to analyze the contributions of phonons with different frequencies to the heat conduction. Coherent heat conduction of the low frequency phonons is found and their contributions to facilitate heat transfer are confirmed. The Car-Parrinello molecular dynamics simulation outputs in the work show reasonable thermophysical results of the thermal energy transport process and shed light on the potential applications of treating the heat transfer in the superlattices of semiconductor materials from a quantum mechanical molecular dynamics simulation perspective.

physics.comp-ph↗

arXiv2016

Double MRT Thermal Lattice Boltzmann Method for Simulating Natural Convection of Low Prandtl Number Fluids

The purposes of this paper are testing an efficiency algorithm based on LBM and using it to analyze two-dimensional natural convection with low Prandtl number. Steady state or oscillatory results are obtained using double multiple-relaxation-time thermal lattice Boltzmann method. The velocity and temperature fields are solved using D2Q9 and D2Q5 models, respectively. With different Rayleigh number, the tested natural convection can either achieve to steady state or oscillatory. With fixed Rayleigh number, lower Prandtl number leads to a weaker convection effect, longer oscillation period and higher oscillation amplitude for the cases reaching oscillatory solutions. At fixed Prandtl number, higher Rayleigh number leads to a more notable convection effect and longer oscillation period. Double multiple-relaxation-time thermal lattice Boltzmann method is applied to simulate the low Prandtl number fluid natural convection. Rayleigh number and Prandtl number effects are also investigated when the natural convection results oscillate.

physics.comp-ph↗