arXiv subjects

Yuhao Fu

Publications and source records attributed to Yuhao Fu.

At least 19 recordsLinked to original sources

arXiv2026

Do Humans Bargain Differently with AI? Evidence from Alternating-Offer Games

Artificial intelligence increasingly participates in economic interactions not only as a tool, but also as an autonomous bargaining counterpart negotiating on behalf of firms, platforms, and consumers. Yet little is known about how humans respond psychologically and strategically when bargaining with such agents in dynamic settings. We study this question in a laboratory experiment using a three-stage alternating-offer bargaining game in which participants negotiate in real time with either another human or a GPT-based AI agent. We also introduce a human-beneficiary condition in which the AI agent's earnings may affect another participant's payment. Agreements are not reached earlier in human-human bargaining than in human-AI bargaining, but they are reached significantly earlier when the AI's payoff affects another participant's payoff. Human proposers offer more to human opponents than to AI agents, whereas responders become significantly more willing to accept unfair AI offers when AI earnings may benefit another human. These findings suggest that fairness and reciprocity toward AI are weaker and more conditional than toward humans, but partially remerge when AI outcomes affect real people. The results have implications for the design of AI negotiation systems and broader human-AI economic interactions.

econ.GN↗

arXiv2026

Do people rely on ChatGPT more than their peers to detect deepfake news?

This experimental study investigates how people rely on different sources of advice when detecting AI-generated fake news (deepfake news). In a laboratory deepfake detection task, student participants identified the proportion of human-written (non-AI-generated) content in synthetic deepfake news articles and received advice from ChatGPT (GPT-4), human peers, or linguistic experts. The results show that participants rely more on ChatGPT than on human peers when detecting GPT-2-generated deepfake news. Participants also rely more on linguistic experts than on peers, while the relative reliance on experts versus ChatGPT is mixed across experimental waves, potentially reflecting time trends in beliefs about AI-based detection. Importantly, in the additional experiment conducted in 2025 under the same experimental procedure, participants relied more on linguistic experts than on ChatGPT. Moreover, performance improvements reflect the joint role of reliance and advice quality, arising primarily when participants rely on high-quality advice. Overall, relying on AI to detect AI-generated deepfakes can improve detection outcomes, but only when AI-based detection tools are of sufficiently high quality. These findings highlight the dual role of GAI as both a source of deepfakes and a tool for mitigating related risks.

econ.GN↗

arXiv2026

Organizing Principles for Moir\'e Quantum Matter

Moir\'e flat bands in van der Waals bilayers are usually discussed through a small set of mechanisms associated with the $\Gamma$ and $K$ valleys of hexagonal crystals, and more recently with $M$-valleys systems. Here we show that this view is incomplete. The momentum-space location and effective local orbital character of the monolayer's band edge, in conjunction with the moir\'e symmetry and the symmetry representations of the resulting bands, provide a general set of organizing variables for the emergent low-energy moir\'e Hamiltonian. Applying fully relaxed first-principles calculations, band unfolding and symmetry-representation analysis to more than 600 commensurate twisted bilayers spanning all 2D lattice classes, we identify several routes to moir\'e quantum matter beyond the conventional single-orbital paradigm. The resulting flat bands realize trigonal, honeycomb, square, checkerboard and kagome-like Hubbard models with single-orbital, multi-orbital and multi-site Hilbert spaces; spin-orbit-coupled multi-orbital flat bands exhibit symmetry-indicated topology beyond the conventional $K$-valley setting; and nonsymmorphic moir\'e symmetries enforce semimetallic flat-band connectivity. Analogous quasi-one-dimensional flat-band structures are found in $M$-valley hexagonal systems and $X$-valley square or rectangular systems resulting from emergent momentum-space nonsymmorphic symmetries. Separately, coupled multi-valley manifolds with kagome-like connectivity are identified in several systems whose parent band edges lie at non-high-symmetry points. These results establish a valley-orbital-symmetry framework for connecting parent-material electronic structure to emergent moir\'e Hamiltonians relevant to correlated, topological and symmetry-enforced moir\'e phases.

cond-mat.mtrl-sci↗

arXiv2026

TENT: A Declarative Slice Spraying Engine for Performant and Resilient Data Movement in Disaggregated LLM Serving

Modern GPU clusters rely on complex, heterogeneous interconnects. As large language model (LLM) serving shifts toward agentic reasoning, KVCache becomes a first-class mobile asset, driving frequent migrations and massive elephant flows that dominate the execution critical path. Operating Mooncake Transfer Engine (TE) on thousands of GPUs exposed a fundamental flaw in existing frameworks: imperative, early-binding path selection. This rigidity results in state-blind striping that ignores congestion and grey failures, leading to bandwidth stranding. This also produces operational fragility where routine faults require manual intervention. We present TENT, a declarative orchestration engine that decouples transfer intent from physical execution. By abstracting interconnects into a unified resource pool, TENT shifts path resolution from initialization to slice-time late binding. Applications simply declare transfer intents, while TENT dynamically "sprays" fine-grained slices across rails based on real-time telemetry and predictive cost modeling. This orchestration eliminates head-of-line (HoL) blocking and enables transparent, sub-50 ms self-healing by rerouting slices around failures or degradations without application-level intervention. TENT serves as the production data plane for LLM inference and reinforcement learning (RL) pipelines at multiple industrial clusters. Our evaluation shows that TE+ outperforms state-of-the-art baselines, including Mooncake TE, NIXL, and UCCL. In LLM inference with SGLang HiCache, TE+ achieves up to 1.36x higher throughput and 26% lower P90 time-to-first-token (TTFT) than Mooncake TE. In RL pipelines, TENT accelerates parameter updates in Moonshot Checkpoint Engine by 20-26%.

cs.DC↗

arXiv2025

Spoofing-aware Prompt Learning for Unified Physical-Digital Facial Attack Detection

Real-world face recognition systems are vulnerable to both physical presentation attacks (PAs) and digital forgery attacks (DFs). We aim to achieve comprehensive protection of biometric data by implementing a unified physical-digital defense framework with advanced detection. Existing approaches primarily employ CLIP with regularization constraints to enhance model generalization across both tasks. However, these methods suffer from conflicting optimization directions between physical and digital attack detection under same category prompt spaces. To overcome this limitation, we propose a Spoofing-aware Prompt Learning for Unified Attack Detection (SPL-UAD) framework, which decouples optimization branches for physical and digital attacks in the prompt space. Specifically, we construct a learnable parallel prompt branch enhanced with adaptive Spoofing Context Prompt Generation, enabling independent control of optimization for each attack type. Furthermore, we design a Cues-awareness Augmentation that leverages the dual-prompt mechanism to generate challenging sample mining tasks on data, significantly enhancing the model's robustness against unseen attack types. Extensive experiments on the large-scale UniAttackDataPlus dataset demonstrate that the proposed method achieves significant performance improvements in unified attack detection tasks.

cs.CV↗

arXiv2025

Cloud-Fog-Edge Collaborative Computing for Sequential MIoT Workflow: A Two-Tier DDPG-Based Scheduling Framework

The Medical Internet of Things (MIoT) demands stringent end-to-end latency guarantees for sequential healthcare workflows deployed over heterogeneous cloud-fog-edge infrastructures. Scheduling these sequential workflows to minimize makespan is an NP-hard problem. To tackle this challenge, we propose a Two-tier DDPG-based scheduling framework that decomposes the scheduling decision into a hierarchical process: a global controller performs layer selection (edge, fog, or cloud), while specialized local controllers handle node assignment within the chosen layer. The primary optimization objective is the minimization of the workflow makespan. Experiments results validate our approach, demonstrating increasingly superior performance over baselines as workflow complexity rises. This trend highlights the frameworks ability to learn effective long-term strategies, which is critical for complex, large-scale MIoT scheduling scenarios.

cs.LG↗

arXiv2025

Every FLOP Counts: Scaling a 300B Mixture-of-Experts LING LLM without Premium GPUs

In this technical report, we tackle the challenges of training large-scale Mixture of Experts (MoE) models, focusing on overcoming cost inefficiency and resource limitations prevalent in such systems. To address these issues, we present two differently sized MoE large language models (LLMs), namely Ling-Lite and Ling-Plus (referred to as "Bailing" in Chinese, spelled B\v{a}il\'ing in Pinyin). Ling-Lite contains 16.8 billion parameters with 2.75 billion activated parameters, while Ling-Plus boasts 290 billion parameters with 28.8 billion activated parameters. Both models exhibit comparable performance to leading industry benchmarks. This report offers actionable insights to improve the efficiency and accessibility of AI development in resource-constrained settings, promoting more scalable and sustainable technologies. Specifically, to reduce training costs for large-scale MoE models, we propose innovative methods for (1) optimization of model architecture and training processes, (2) refinement of training anomaly handling, and (3) enhancement of model evaluation efficiency. Additionally, leveraging high-quality data generated from knowledge graphs, our models demonstrate superior capabilities in tool use compared to other models. Ultimately, our experimental findings demonstrate that a 300B MoE LLM can be effectively trained on lower-performance devices while achieving comparable performance to models of a similar scale, including dense and MoE models. Compared to high-performance devices, utilizing a lower-specification hardware system during the pre-training phase demonstrates significant cost savings, reducing computing costs by approximately 20%. The models can be accessed at https://huggingface.co/inclusionAI.

cs.LG↗

arXiv2025

InfiFusion: A Unified Framework for Enhanced Cross-Model Reasoning via LLM Fusion

We introduce InfiFusion, an efficient training pipeline designed to integrate multiple domain-specialized Large Language Models (LLMs) into a single pivot model, effectively harnessing the strengths of each source model. Traditional fusion methods either merge model parameters directly or rely on knowledge distillation with rigid assumptions, limiting their flexibility and efficiency. InfiFusion overcomes these limitations by enhancing Universal Logit Distillation (ULD) with Top-K selection and Logits Standardization. We propose two fusion strategies: Pairwise Fusion (InfiFusion$_p$), where each source model knowledge is distilled individually into the pivot model followed by merging and Unified Fusion (InfiFusion$_u$), where knowledge from all source models is distilled simultaneously into the pivot model. InfiFusion outperforms the state-of-the-art models, such as Qwen-2.5-14B-Instruct and Phi-4, across 11 widely applied benchmarks covering reasoning, coding, mathematics, and instruction-following tasks. Notably, InfiFusion achieves this superior performance while significantly reduces computational costs, completing full training with only 160 H800 GPU hours compared to the millions typically required for traditional LLM training.

cs.CL↗

arXiv2024

SolarDesign: An Online Photovoltaic Device Simulation and Design Platform

SolarDesign (https://solardesign.cn/) is an online photovoltaic device simulation and design platform that provides engineering modeling analysis for crystalline silicon solar cells, as well as emerging high-efficiency solar cells such as organic, perovskite, and tandem cells. The platform offers user-updatable libraries of basic photovoltaic materials and devices, device-level multi-physics simulations involving optical-electrical-thermal interactions, and circuit-level compact model simulations based on detailed balance theory. Employing internationally advanced numerical methods, the platform accurately, rapidly, and efficiently solves optical absorption, electrical transport, and compact circuit models. It achieves multi-level photovoltaic simulation technology from ``materials to devices to circuits'' with fully independent intellectual property rights. Compared to commercial software, the platform achieves high accuracy and improves speed by more than an order of magnitude. Additionally, it can simulate unique electrical transport processes in emerging solar cells, such as quantum tunneling, exciton dissociation, and ion migration.

physics.optics↗

arXiv2024

Unconstrained Model Merging for Enhanced LLM Reasoning

Recent advancements in building domain-specific large language models (LLMs) have shown remarkable success, especially in tasks requiring reasoning abilities like logical inference over complex relationships and multi-step problem solving. However, creating a powerful all-in-one LLM remains challenging due to the need for proprietary data and vast computational resources. As a resource-friendly alternative, we explore the potential of merging multiple expert models into a single LLM. Existing studies on model merging mainly focus on generalist LLMs instead of domain experts, or the LLMs under the same architecture and size. In this work, we propose an unconstrained model merging framework that accommodates both homogeneous and heterogeneous model architectures with a focus on reasoning tasks. A fine-grained layer-wise weight merging strategy is designed for homogeneous models merging, while heterogeneous model merging is built upon the probabilistic distribution knowledge derived from instruction-response fine-tuning data. Across 7 benchmarks and 9 reasoning-optimized LLMs, we reveal key findings that combinatorial reasoning emerges from merging which surpasses simple additive effects. We propose that unconstrained model merging could serve as a foundation for decentralized LLMs, marking a notable progression from the existing centralized LLM framework. This evolution could enhance wider participation and stimulate additional advancement in the field of artificial intelligence, effectively addressing the constraints posed by centralized models.

cs.CL↗

arXiv2024

General theory for longitudinal nonreciprocal charge transport

The longitudinal nonreciprocal charge transport (NCT) in crystalline materials is a highly non-trivial phenomenon, motivating the design of next generation two-terminal rectification devices (e.g., semiconductor diodes beyond PN junctions). The practical application of such devices is built upon crystalline materials whose longitudinal NCT occurs at room temperature and under low magnetic field. However, materials of this type are rather rare and elusive, and theory guiding the discovery of these materials is lacking. Here, we develop such a theory within the framework of semiclassical Boltzmann transport theory. By symmetry analysis, we classify the complete 122 magnetic point groups with respect to the longitudinal NCT phenomenon. The symmetry-adapted Hamiltonian analysis further uncovers a previously overlooked mechanism for this phenomenon. Our theory guides the first-principles prediction of longitudinal NCT in multiferroic \epsilon-Fe2O3 semiconductor that possibly occurs at room temperature, without the application of external magnetic field. These findings advance our fundamental understandings of longitudinal NCT in crystalline materials, and aid the corresponding materials discoveries.

cond-mat.mtrl-sci↗

arXiv2023

Engineering ferroelectricity in monoclinic hafnia

Ferroelectricity in the complementary metal-oxide semiconductor (CMOS)-compatible hafnia (HfO$_2$) is crucial for the fabrication of high-integration nonvolatile memory devices. However, the capture of ferroelectricity in HfO$_2$ requires the stabilization of thermodynamically-metastable orthorhombic or rhombohedral phases, which entails the introduction of defects (e.g., dopants and vacancies) and pays the price of crystal imperfections, causing unpleasant wake-up and fatigue effects. Here, we report a theoretical strategy on the realization of robust ferroelectricity in HfO$_2$-based ferroelectrics by designing a series of epitaxial (HfO$_2$)$_1$/(CeO$_2$)$_1$ superlattices. The advantages of the designated ferroelectric superlattices are defects free, and most importantly, on the base of the thermodynamically stable monoclinic phase of HfO$_2$. Consequently, this allows the creation of superior ferroelectric properties with an electric polarization $>$25 $\mu$C/cm$^2$ and an ultralow polarization-switching energy barrier at $\sim$2.5 meV/atom. Our work may open an entirely new route towards the fabrication of high-performance HfO$_2$ based ferroelectric devices.

cond-mat.mtrl-sci↗

arXiv2022

Inorganic Crystal Structure Prototype Database based on Unsupervised Learning of Local Atomic Environments

Recognition of structure prototypes from tremendous known inorganic crystal structures has been an important subject beneficial for material science research and new materials design. The existing databases of inorganic crystal structure prototypes were mostly constructed by classifying materials in terms of the crystallographic space group information. Herein, we employed a distinct strategy to construct the inorganic crystal structure prototype database, relying on the classification of materials in terms of local atomic environments (LAE) accompanied by unsupervised machine learning method. Specifically, we adopted a hierarchical clustering approach onto all experimentally known inorganic crystal structures data to identify structure prototypes. The criterion for hierarchical clustering is the LAE represented by the state-of-the-art structure fingerprints of the improved bond-orientational order parameters and the smooth overlap of atomic positions. This allows us to build up a LAE-based Inorganic Crystal Structure Prototype Database (LAE-ICSPD) containing 15,613 structure prototypes with defined stoichiometries. In addition, we have developed a Structure Prototype Generator Infrastructure (SPGI) package, which is a useful toolkit for structure prototype generation. Our developed SPGI toolkit and LAE-ICSPD are beneficial for investigating inorganic materials in a global way as well as accelerating materials discovery process in the data-driven mode.

cond-mat.mtrl-sci↗

arXiv2021

JAMIP: an artificial-intelligence aided data-driven infrastructure for computational materials informatics

Materials informatics has emerged as a promisingly new paradigm for accelerating materials discovery and design. It exploits the intelligent power of machine learning methods in massive materials data from experiments or simulations to seek for new materials, functionality, principles, etc. Developing specialized facility to generate, collect, manage, learn and mine large-scale materials data is crucial to materials informatics. We herein developed an artificial-intelligence-aided data-driven infrastructure named Jilin Artificial-intelligence aided Materials-design Integrated Package (JAMIP), which is an open-source Python framework to meet the research requirements of computational materials informatics. It is integrated by materials production factory, high-throughput first-principles calculations engine, automatic tasks submission and monitoring progress, data extraction, management and storage system, and artificial intelligence machine learning based data mining functions. We have integrated specific features such as inorganic crystal structure prototype database to facilitate high-throughput calculations and essential modules associated with machine learning studies of functional materials. We demonstrated how our developed code is useful in exploring materials informatics of optoelectronic semiconductors by taking halide perovskites as typical case. By obeying the principles of automation, extensibility, reliability and intelligence, the JAMIP code is a promisingly powerful tool contributing to the fast-growing field of computational materials informatics.

cond-mat.mtrl-sci↗

arXiv2020

Characterization of rattling in relation to thermal conductivity: ordered half-Heusler semiconductors

The factors that affect the thermal conductivity of semiconductors is a topic of great scientific interest, especially in relation to thermoelectrics. Key developments have been the concept of the phonon-glass-electron-crystal (PGEC) and the related idea of rattling to achieve this. We use first principles phonon and thermal conductivity calculations in order to explore the concept of rattling for stoichiometric ordered half-Heusler compounds. These compounds can be regarded as filled zinc blende materials, and the filling atom could be viewed as a rattler if it is weakly bound. We use two simple metrics, one related to the frequency and the other to bond frustration and anharmonicity. We find that both measures correlate with thermal conductivity. This suggests that both may be useful in screening materials for low thermal conductivity.

cond-mat.mtrl-sci↗

arXiv2019

Understanding the lattice thermal conductivity of SrTiO3 from an ab initio perspective

We present a detailed analysis of the structure dependence of the lattice thermal conductivity of SrTiO3. We have used both ab initio Molecular Dynamic simulations and Density Functional Theory calculations to decouple the effect of different structural distortions on the thermal conductivity. We have identified two main mechanisms for tuning the thermal conductivity when a distortion is applied. First, the modification of the acoustic-modes energy dispersion when a change in the lattice parameters is imposed and second, the low energy polar modes. In particular and counterintuitively, we have found that an increase in the angle of the oxygen octahedral rotations increases the thermal conductivity due to its coupling to these polar modes.

cond-mat.mtrl-sci↗

arXiv2019

Tuning from Frustrated Magnetism to Superconductivity in Quasi-One-Dimensional KCr$_3$As$_3$ Through Hydrogen Doping

We report the charge doping of KCr$_3$As$_3$ via H intercalation. We show that the previously reported ethanol bath deintercalation of K$_2$Cr$_3$As$_3$ to KCr$_3$As$_3$ has a secondary effect whereby H from the bath enters the quasi-one-dimensional Cr$_6$As$_6$ chains. Furthermore, we find that - contrary to previous interpretations - the difference between non-superconducting as-grown KCr$_3$As$_3$ samples and superconducting hydrothermally annealed samples is not a change in crystallinity but due to charge doping, with the latter treatment increasing the H concentration in the CrAs tubes effectively electron-doping the 133 compound. These results suggest a new stoichiometry KH$_x$Cr$_3$As$_3$, that superconductivity arises from a suppressed magnetic order via a tunable parameter and pave the way for the first charge-doped phase diagram in these materials.

cond-mat.supr-con↗

arXiv2019

Structural Instability and Magnetism of Superconducting KCr$_3$As$_3$

We find a lattice instability in the superconductor KCr$_3$As$_3$, corresponding to a distortion of the Cr metallic wires in the crystal structure. This distortion couples strongly to both the electronic and magnetic properties, in particular by making the electronic structure much more nearly one-dimensional, and by shifting the compound away from magnetism. We discuss the implications of these results in the context of the possibly unconventional superconductivity of this phase.

cond-mat.supr-con↗