arXiv subjects

Qiang He

Publications and source records attributed to Qiang He.

At least 19 recordsLinked to original sources

arXiv2026

FANS: Federated Adaptive Network Search Learning for Heterogeneous Devices

Heterogeneous Federated Learning (HFL) aims to train models across devices with diverse resource budgets while preserving data privacy. Existing HFL methods typically bind training to a small predefined menu of model configurations, which limits architectural coverage. To address this bottleneck, we introduce Federated Adaptive Network Search (FANS), a hypernetwork-based framework that learns a shared architecture space rather than a fixed set of client models. To optimize this shared space efficiently, we propose the Federated Parallel Scaling (FPS) algorithm, which jointly trains multiple sampled subnetworks in parallel with self-distillation so that larger sampled subnetworks can supervise smaller ones during local updates. We evaluate FANS on CIFAR-10, CIFAR-100, and MNLI using ResNet-18, DenseNet-121, and BERT-base, respectively. Across all benchmarks, FANS expands the feasible subnetwork pool by orders of magnitude (e.g., 4,680 candidates for ResNet-18 vs. 4 in existing methods) and improves the average accuracy-efficiency trade-off relative to representative HFL baselines. Device heterogeneity is emulated through resource tiers, and evaluation covers accuracy, parameter count, and MACs.

cs.LG↗

arXiv2026

Multi-dimensional training-priority weighting based on physical information propagation paths: a unified residual-weighting framework for physics-informed neural networks

Physics-informed neural networks (PINNs) have shown promise for solving partial differential equations (PDEs); however, their synchronous optimization treats residuals of different regions and constraints equally, which is inconsistent with the progressive "from source to response" physical information propagation path, degrading training stability and accuracy. Existing causal training methods focus mainly on the temporal dimension, lacking a unified characterization of spatial and boundary dimensions. To address this, we define a unified class of training priorities according to the physical information propagation path: premise regions should be learned before dependent regions; temporal, spatial, and boundary priorities are instances of this principle. Using neural tangent kernel (NTK) dynamics, we theoretically analyze why standard PINNs do not obey this priority: their residual convergence order is governed by the NTK spectrum and is independent of the propagation path. Accordingly, we propose a unified multi-dimensional priority-constraint framework that partitions the domain along the propagation path and constructs negative-exponential residual weights, converting the physical propagation order into a training priority. For cases with coexisting priorities, we introduce a directional compatibility coefficient to clarify that "orthogonal directions can be coupled multiplicatively in synergy, whereas coaxial opposite directions cannot." Benchmark cases show that this method consistently improves the convergence behavior and prediction accuracy of PINNs on problems with clear propagation paths or constraint-dominated structures, without modifying the network architecture and with controllable additional computational cost.

cs.LG↗

arXiv2026

One Model for All: Multi-Objective Controllable Language Models

Aligning large language models (LLMs) with human preferences is critical for enhancing LLMs' safety, helpfulness, humor, faithfulness, etc. Current reinforcement learning from human feedback (RLHF) mainly focuses on a fixed reward learned from average human ratings, which may weaken the adaptability and controllability of varying preferences. However, creating personalized LLMs requires aligning LLMs with individual human preferences, which is non-trivial due to the scarce data per user and the diversity of user preferences in multi-objective trade-offs, varying from emphasizing empathy in certain contexts to demanding efficiency and precision in others. Can we train one LLM to produce personalized outputs across different user preferences on the Pareto front? In this paper, we introduce Multi-Objective Control (MOC), which trains a single LLM to directly generate responses in the preference-defined regions of the Pareto front. Our approach introduces multi-objective optimization (MOO) principles into RLHF to train an LLM as a preference-conditioned policy network. We improve the computational efficiency of MOC by applying MOO at the policy level, enabling us to fine-tune a 7B-parameter model on a single A6000 GPU. Extensive experiments demonstrate the advantages of MOC over baselines in three aspects: (i) controllability of LLM outputs w.r.t. user preferences on the trade-off among multiple rewards; (ii) quality and diversity of LLM outputs, measured by the hyper-volume of multiple solutions achieved; and (iii) generalization to unseen preferences. These results highlight MOC's potential for real-world applications requiring scalable and customizable LLMs.

cs.LG↗

arXiv2026

TCATSeg: A Tooth Center-Wise Attention Network for 3D Dental Model Semantic Segmentation

Accurate semantic segmentation of 3D dental models is essential for digital dentistry applications such as orthodontics and dental implants. However, due to complex tooth arrangements and similarities in shape among adjacent teeth, existing methods struggle with accurate segmentation, because they often focus on local geometry while neglecting global contextual information. To address this, we propose TCATSeg, a novel framework that combines local geometric features with global semantic context. We introduce a set of sparse yet physically meaningful superpoints to capture global semantic relationships and enhance segmentation accuracy. Additionally, we present a new dataset of 400 dental models, including pre-orthodontic samples, to evaluate the generalization of our method. Extensive experiments demonstrate that TCATSeg outperforms state-of-the-art approaches.

cs.CV↗

arXiv2025

Scaling Behaviors of LLM Reinforcement Learning Post-Training: An Empirical Study in Mathematical Reasoning

While scaling laws for large language models (LLMs) during pre-training have been extensively studied, their behavior under reinforcement learning (RL) post-training remains largely unexplored. This paper presents a systematic empirical investigation of scaling behaviors in RL-based post-training, with a particular focus on mathematical reasoning. Based on a set of experiments across the full Qwen2.5 dense model series (0.5B to 72B), we characterize how model scale, data volume, and computational budget interact to shape performance. Our analysis leads to four key findings: 1. Larger models consistently exhibit superior learning efficiency on both compute and data metrics. 2. The relationship between test loss, compute, and data can be modeled by a predictive power-law which is robust across both base and instruction-tuned models. 3. Although larger models exhibit higher learning efficiency, the analytical learning efficiency term k(N) in the power-law reveals a latent saturation trend in learning efficiency as model size continues to increase. 4. In data-constrained regimes, repeated reuse of high-quality data proves highly effective, as final performance is primarily governed by the total number of optimization steps rather than the uniqueness of samples. Collectively, these results provide a principled foundation and practical guidelines for efficiently scaling the reasoning capabilities of LLMs through RL post-training.

cs.LG↗

arXiv2025

Pareto Multi-Objective Alignment for Language Models

Large language models (LLMs) are increasingly deployed in real-world applications that require careful balancing of multiple, often conflicting, objectives, such as informativeness versus conciseness, or helpfulness versus creativity. However, current alignment methods, primarily based on RLHF, optimize LLMs toward a single reward function, resulting in rigid behavior that fails to capture the complexity and diversity of human preferences. This limitation hinders the adaptability of LLMs to practical scenarios, making multi-objective alignment (MOA) a critical yet underexplored area. To bridge this gap, we propose Pareto Multi-Objective Alignment (PAMA), a principled and computationally efficient algorithm designed explicitly for MOA in LLMs. In contrast to computationally prohibitive multi-objective optimization (MOO) methods, PAMA transforms multi-objective RLHF into a convex optimization with a closed-form solution, significantly enhancing scalability. Traditional MOO approaches suffer from prohibitive O(n^2*d) complexity, where d represents the number of model parameters, typically in the billions for LLMs, rendering direct optimization infeasible. PAMA reduces this complexity to O(n) where n is the number of objectives, enabling optimization to be completed within milliseconds. We provide theoretical guarantees that PAMA converges to a Pareto stationary point, where no objective can be improved without degrading at least one other. Extensive experiments across language models ranging from 125M to 7B parameters demonstrate PAMA's robust and effective MOA capabilities, aligning with its theoretical advantages. PAMA provides a highly efficient solution to the MOA problem that was previously considered intractable, offering a practical and theoretically grounded approach to aligning LLMs with diverse human values, paving the way for versatile and adaptable real-world AI deployments.

cs.LG↗

arXiv2025

Task Adaptation from Skills: Information Geometry, Disentanglement, and New Objectives for Unsupervised Reinforcement Learning

Unsupervised reinforcement learning (URL) aims to learn general skills for unseen downstream tasks. Mutual Information Skill Learning (MISL) addresses URL by maximizing the mutual information between states and skills but lacks sufficient theoretical analysis, e.g., how well its learned skills can initialize a downstream task's policy. Our new theoretical analysis in this paper shows that the diversity and separability of learned skills are fundamentally critical to downstream task adaptation but MISL does not necessarily guarantee these properties. To complement MISL, we propose a novel disentanglement metric LSEPIN. Moreover, we build an information-geometric connection between LSEPIN and downstream task adaptation cost. For better geometric properties, we investigate a new strategy that replaces the KL divergence in information geometry with Wasserstein distance. We extend the geometric analysis to it, which leads to a novel skill-learning objective WSEP. It is theoretically justified to be helpful to downstream task adaptation and it is capable of discovering more initial policies for downstream tasks than MISL. We finally propose another Wasserstein distance-based algorithm PWSEP that can theoretically discover all optimal initial policies.

cs.LG↗

arXiv2025

Over-PINNs: Enhancing Physics-Informed Neural Networks via Higher-Order Partial Derivative Overdetermination of PDEs

Partial differential equations (PDEs) serve as the cornerstone of mathematical physics. In recent years, Physics-Informed Neural Networks (PINNs) have significantly reduced the dependence on large datasets by embedding physical laws directly into the training of neural networks. However, when dealing with complex problems, the accuracy of PINNs still has room for improvement. To address this issue, we introduce the Over-PINNs framework, which leverages automatic differentiation (AD) to generate higher-order auxiliary equations that impose additional physical constraints. These equations are incorporated as extra loss terms in the training process, effectively enhancing the model's ability to capture physical information through an "overdetermined" approach. Numerical results illustrate that this method exhibits strong versatility in solving various types of PDEs. It achieves a significant improvement in solution accuracy without incurring substantial additional computational costs.

cs.LG↗

arXiv2025

Learning from Loss Landscape: Generalizable Mixed-Precision Quantization via Adaptive Sharpness-Aware Gradient Aligning

Mixed Precision Quantization (MPQ) has become an essential technique for optimizing neural network by determining the optimal bitwidth per layer. Existing MPQ methods, however, face a major hurdle: they require a computationally expensive search for quantization policies on large-scale datasets. To resolve this issue, we introduce a novel approach that first searches for quantization policies on small datasets and then generalizes them to large-scale datasets. This approach simplifies the process, eliminating the need for large-scale quantization fine-tuning and only necessitating model weight adjustment. Our method is characterized by three key techniques: sharpness-aware minimization for enhanced quantization generalization, implicit gradient direction alignment to handle gradient conflicts among different optimization objectives, and an adaptive perturbation radius to accelerate optimization. Both theoretical analysis and experimental results validate our approach. Using the CIFAR10 dataset (just 0.5\% the size of ImageNet training data) for MPQ policy search, we achieved equivalent accuracy on ImageNet with a significantly lower computational cost, while improving efficiency by up to 150% over the baselines.

cs.CV↗

arXiv2025

NuWa: Deriving Lightweight Class-Specific Vision Transformers for Edge Devices

Vision Transformers (ViTs) often need to be compressed for deployment on resource-constrained edge devices like drones and smart vehicles. However, existing model compression methods ignore that many edge devices only require the knowledge of specific classes for their applications. As a result, the derived all-class ViTs retain redundant knowledge and perform suboptimally on these classes. We discovered that simply replacing the calibration dataset with class-specific data does not suffice to address this issue, as these methods face two fundamental limitations. First, they overlook the existence of class-detrimental weights, which interfere with specialization, while removing them can improve class-specific performance. Second, the diversity of target classes and resource constraints on edge devices demand numerous customized models. Existing methods are time-consuming and computationally expensive, thus unscalable. In this work, we present NuWa, a cost-efficient method that addresses these challenges by deriving small ViTs from base ViTs for edge devices with specific class requirements. NuWa performs self-knowledge purification to prune class-detrimental weights and efficiently derives compact ViTs through closed-form optimization. Without post-pruning retraining, the derived edge ViTs surpass the base ViT in class-specific accuracy and accelerate inference. Comprehensive experiments demonstrate that NuWa outperforms state-of-the-art training-free pruning methods on class-specific tasks by up to 29.00\% in accuracy. Compared with the best-performing training-dependent pruning method, NuWa achieves a 33.69x pruning speedup and reduces pruning cost by up to 99.83\%, with only a 0.61\% average accuracy loss. Project Page: https://github.com/CGCL-codes/NuWa.

cs.CV↗

arXiv2024

Diffgrasp: Whole-Body Grasping Synthesis Guided by Object Motion Using a Diffusion Model

Generating high-quality whole-body human object interaction motion sequences is becoming increasingly important in various fields such as animation, VR/AR, and robotics. The main challenge of this task lies in determining the level of involvement of each hand given the complex shapes of objects in different sizes and their different motion trajectories, while ensuring strong grasping realism and guaranteeing the coordination of movement in all body parts. Contrasting with existing work, which either generates human interaction motion sequences without detailed hand grasping poses or only models a static grasping pose, we propose a simple yet effective framework that jointly models the relationship between the body, hands, and the given object motion sequences within a single diffusion model. To guide our network in perceiving the object's spatial position and learning more natural grasping poses, we introduce novel contact-aware losses and incorporate a data-driven, carefully designed guidance. Experimental results demonstrate that our approach outperforms the state-of-the-art method and generates plausible whole-body motion sequences.

cs.CV↗

arXiv2024

Advancing DRL Agents in Commercial Fighting Games: Training, Integration, and Agent-Human Alignment

Deep Reinforcement Learning (DRL) agents have demonstrated impressive success in a wide range of game genres. However, existing research primarily focuses on optimizing DRL competence rather than addressing the challenge of prolonged player interaction. In this paper, we propose a practical DRL agent system for fighting games named Sh\=ukai, which has been successfully deployed to Naruto Mobile, a popular fighting game with over 100 million registered users. Sh\=ukai quantifies the state to enhance generalizability, introducing Heterogeneous League Training (HELT) to achieve balanced competence, generalizability, and training efficiency. Furthermore, Sh\=ukai implements specific rewards to align the agent's behavior with human expectations. Sh\=ukai's ability to generalize is demonstrated by its consistent competence across all characters, even though it was trained on only 13% of them. Additionally, HELT exhibits a remarkable 22% improvement in sample efficiency. Sh\=ukai serves as a valuable training partner for players in Naruto Mobile, enabling them to enhance their abilities and skills.

cs.AI↗

arXiv2024

An Interdisciplinary Perspective of the Built-Environment Microbiome

The built environment provides an excellent setting for interdisciplinary research on the dynamics of microbial communities. The system is simplified compared to many natural settings, and to some extent the entire environment can be manipulated, from architectural design, to materials use, air flow, human traffic, and capacity to disrupt microbial communities through cleaning. Here we provide an overview of the ecology of the microbiome in the built environment. We address niche space and refugia, population and community (metagenomic) dynamics, spatial ecology within a building, including the major microbial transmission mechanisms, as well as evolution. We also address the landscape ecology connecting microbiomes between physically separated buildings. At each stage we pay particular attention to the actual and potential interface between disciplines, such as ecology, epidemiology, materials science, and human social behavior. We end by identifying some opportunities for future interdisciplinary research on the microbiome of the built environment.

q-bio.PE↗

arXiv2024

Adaptive Regularization of Representation Rank as an Implicit Constraint of Bellman Equation

Representation rank is an important concept for understanding the role of Neural Networks (NNs) in Deep Reinforcement learning (DRL), which measures the expressive capacity of value networks. Existing studies focus on unboundedly maximizing this rank; nevertheless, that approach would introduce overly complex models in the learning, thus undermining performance. Hence, fine-tuning representation rank presents a challenging and crucial optimization problem. To address this issue, we find a guiding principle for adaptive control of the representation rank. We employ the Bellman equation as a theoretical foundation and derive an upper bound on the cosine similarity of consecutive state-action pairs representations of value networks. We then leverage this upper bound to propose a novel regularizer, namely BEllman Equation-based automatic rank Regularizer (BEER). This regularizer adaptively regularizes the representation rank, thus improving the DRL agent's performance. We first validate the effectiveness of automatic control of rank on illustrative experiments. Then, we scale up BEER to complex continuous control tasks by combining it with the deterministic policy gradient method. Among 12 challenging DeepMind control tasks, BEER outperforms the baselines by a large margin. Besides, BEER demonstrates significant advantages in Q-value approximation. Our code is available at https://github.com/sweetice/BEER-ICLR2024.

cs.LG↗

arXiv2023

The Case for Controls: Identifying outbreak risk factors through case-control comparisons

Investigations of infectious disease outbreaks often focus on identifying place- and context-dependent factors responsible for emergence and spread, resulting in phenomenological narratives ill-suited to developing generalizable predictive and preventive measures. We contend that case-control hypothesis testing is a more powerful framework for epidemiological investigation. The approach, widely used in medical research, involves identifying counterfactuals, with case-control comparisons drawn to test hypotheses about the conditions that manifest outbreaks. Here we outline the merits of applying a case-control framework as epidemiological study design. We first describe a framework for iterative multidisciplinary interrogation to discover minimally sufficient sets of factors that can lead to disease outbreaks. We then lay out how case-control comparisons can respectively center on pathogen(s), factor(s), or landscape(s) with vignettes focusing on pathogen transmission. Finally, we consider how adopting case-control approaches can promote evidence-based decision making for responding to and preventing outbreaks.

q-bio.PE↗

arXiv2023

EdgeDis: Enabling Fast, Economical, and Reliable Data Dissemination for Mobile Edge Computing

Mobile edge computing (MEC) enables web data caching in close geographic proximity to end users. Popular data can be cached on edge servers located less than hundreds of meters away from end users. This ensures bounded latency guarantees for various latency-sensitive web applications. However, transmitting a large volume of data out of the cloud onto many geographically-distributed web servers individually can be expensive. In addition, web content dissemination may be interrupted by various intentional and accidental events in the volatile MEC environment, which undermines dissemination efficiency and subsequently incurs extra transmission costs. To tackle the above challenges, we present a novel scheme named EdgeDis that coordinates data dissemination by distributed consensus among those servers. We analyze EdgeDis's validity theoretically and evaluate its performance experimentally. Results demonstrate that compared with baseline and state-of-the-art schemes, EdgeDis: 1) is 5.97x - 7.52x faster; 2) reduces dissemination costs by 48.21% to 91.87%; and 3) reduces performance loss caused by dissemination failures by up to 97.30% in time and 96.35% in costs.

cs.DC↗

arXiv2023

Keep Various Trajectories: Promoting Exploration of Ensemble Policies in Continuous Control

The combination of deep reinforcement learning (DRL) with ensemble methods has been proved to be highly effective in addressing complex sequential decision-making problems. This success can be primarily attributed to the utilization of multiple models, which enhances both the robustness of the policy and the accuracy of value function estimation. However, there has been limited analysis of the empirical success of current ensemble RL methods thus far. Our new analysis reveals that the sample efficiency of previous ensemble DRL algorithms may be limited by sub-policies that are not as diverse as they could be. Motivated by these findings, our study introduces a new ensemble RL algorithm, termed \textbf{T}rajectories-awar\textbf{E} \textbf{E}nsemble exploratio\textbf{N} (TEEN). The primary goal of TEEN is to maximize the expected return while promoting more diverse trajectories. Through extensive experiments, we demonstrate that TEEN not only enhances the sample diversity of the ensemble policy compared to using sub-policies alone but also improves the performance over ensemble RL algorithms. On average, TEEN outperforms the baseline ensemble DRL algorithms by 41\% in performance on the tested representative environments.

cs.LG↗

arXiv2023

SpeechMirror: A Multimodal Visual Analytics System for Personalized Reflection of Online Public Speaking Effectiveness

As communications are increasingly taking place virtually, the ability to present well online is becoming an indispensable skill. Online speakers are facing unique challenges in engaging with remote audiences. However, there has been a lack of evidence-based analytical systems for people to comprehensively evaluate online speeches and further discover possibilities for improvement. This paper introduces SpeechMirror, a visual analytics system facilitating reflection on a speech based on insights from a collection of online speeches. The system estimates the impact of different speech techniques on effectiveness and applies them to a speech to give users awareness of the performance of speech techniques. A similarity recommendation approach based on speech factors or script content supports guided exploration to expand knowledge of presentation evidence and accelerate the discovery of speech delivery possibilities. SpeechMirror provides intuitive visualizations and interactions for users to understand speech factors. Among them, SpeechTwin, a novel multimodal visual summary of speech, supports rapid understanding of critical speech factors and comparison of different speech samples, and SpeechPlayer augments the speech video by integrating visualization of the speaker's body language with interaction, for focused analysis. The system utilizes visualizations suited to the distinct nature of different speech factors for user comprehension. The proposed system and visualization techniques were evaluated with domain experts and amateurs, demonstrating usability for users with low visualization literacy and its efficacy in assisting users to develop insights for potential improvement.

cs.HC↗