arXiv subjects

Yao Wei

Publications and source records attributed to Yao Wei.

At least 19 recordsLinked to original sources

arXiv2026

Correlation-driven phonon renormalisation and the equation of state of $\gamma$-cerium

We investigate the thermodynamic properties of elemental cerium by assessing the crucial role of phonon free energy within the framework of dynamical mean-field theory (DMFT). While conventional density functional theory (DFT) often fails to capture the intricate energy landscape of $f$-electron materials, our approach integrates many-body electronic correlations with lattice dynamics to achieve a more rigorous description of the equation of state. We calculate the total energy as a function of the lattice constant at both the DFT and DFT+DMFT levels, subsequently incorporating the vibrational free energy derived from the phonon density of states. Our findings reveal that electronic renormalisation of the force constants significantly alters the phonon spectra, particularly in the strongly correlated $\gamma$-phase. By applying these phonon corrections to the energy profiles, we observe a substantial refinement in the predicted equilibrium volumes. Using principal-component-based machine learning, we interpolate phonon dispersions continuously from a finite set of first-principles calculations and compare them to experiment, finding significantly closer agreement compared to conventional DFT and DFT+U calculations that neglect dynamical many-body correlations. This study underlines the necessity of accounting for both electronic and vibrational entropy when evaluating the phase stability and structural transitions of lanthanide systems under varying pressures and temperatures.

cond-mat.str-el↗

arXiv2026

FlowOVD: Learning Generative Latent Flows for Zero-shot Open-vocabulary Detection

Open-vocabulary object detection (OVD) has achieved remarkable progress through large-scale vision-language pre-training. Existing methods, however, typically formulate OVD as a discriminative prediction problem, where decoder queries are either static or initialized from encoder features, thus limiting their diversity and flexibility. In this paper, we introduce a generative perspective by modeling decoder query generation as a continuous transport process in latent space. We propose FlowOVD, a text-conditioned query generation framework based on rectified flow that progressively transforms text-agnostic queries into text-guided queries. By introducing continuous latent query dynamics into a vision-language model (VLM) based detector, our method avoids heuristic discrete query construction and enables more expressive semantic alignment for open-vocabulary detection. Without requiring additional training data, FlowOVD achieves 49.5 AP on COCO and 31.5 AP on LVIS, outperforming GroundingDINO by +1.2 AP (+2.5 %) and +4.1 AP (+15.0 %), respectively. The larger gain on the challenging long-tailed LVIS benchmark further highlights the effectiveness of continuous query generation for open-vocabulary generalization.

cs.CV↗

arXiv2026

GLM-5: from Vibe Coding to Agentic Engineering

We present GLM-5, a next-generation foundation model designed to transition the paradigm of vibe coding to agentic engineering. Building upon the agentic, reasoning, and coding (ARC) capabilities of its predecessor, GLM-5 adopts DSA to significantly reduce training and inference costs while maintaining long-context fidelity. To advance model alignment and autonomy, we implement a new asynchronous reinforcement learning infrastructure that drastically improves post-training efficiency by decoupling generation from training. Furthermore, we propose novel asynchronous agent RL algorithms that further improve RL quality, enabling the model to learn from complex, long-horizon interactions more effectively. Through these innovations, GLM-5 achieves state-of-the-art performance on major open benchmarks. Most critically, GLM-5 demonstrates unprecedented capability in real-world coding tasks, surpassing previous baselines in handling end-to-end software engineering challenges. Code, models, and more information are available at https://github.com/zai-org/GLM-5.

cs.LG↗

arXiv2025

GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models

We present GLM-4.5, an open-source Mixture-of-Experts (MoE) large language model with 355B total parameters and 32B activated parameters, featuring a hybrid reasoning method that supports both thinking and direct response modes. Through multi-stage training on 23T tokens and comprehensive post-training with expert model iteration and reinforcement learning, GLM-4.5 achieves strong performance across agentic, reasoning, and coding (ARC) tasks, scoring 70.1% on TAU-Bench, 91.0% on AIME 24, and 64.2% on SWE-bench Verified. With much fewer parameters than several competitors, GLM-4.5 ranks 3rd overall among all evaluated models and 2nd on agentic benchmarks. We release both GLM-4.5 (355B parameters) and a compact version, GLM-4.5-Air (106B parameters), to advance research in reasoning and agentic AI systems. Code, models, and more information are available at https://github.com/zai-org/GLM-4.5.

cs.CL↗

arXiv2025

Impact of electronic correlations on the superconductivity of high-pressure CeH$_9$

Rare-earth superhydrides have attracted considerable attention because of their high critical superconducting temperature under extreme pressures. They are known to have localized valence electrons, implying strong electronic correlations. However, such many-body effects are rarely included in first-principles studies of rare-earth superhydrides because of the complexity of their high-pressure phases. In this work, we use a combined density functional theory and dynamical mean-field theory approach to study both electrons and phonons in the prototypical rare-earth superhydride CeH$_9$, shedding light on the impact of electronic correlations on its critical temperature for phonon-mediated superconductivity. Our findings indicate that electronic correlations result in a larger electronic density at the Fermi level, a bigger superconducting gap, and softer vibrational modes associated with hydrogen atoms. Together, the inclusion of these correlation signatures increases the Migdal-Eliashberg superconducting critical temperature from 47 K to 96 K, close to the measured 95 K. Our results reconcile experimental observations and theoretical predictions for CeH$_9$ and herald a path towards the quantitative modeling of phonon-mediated superconductivity for interacting electron systems.

cond-mat.supr-con↗

arXiv2025

SWE-Dev: Building Software Engineering Agents with Training and Inference Scaling

Large language models (LLMs) have advanced rapidly from conversational problem solving to addressing real-world tasks involving tool use, such as software engineering (SWE). Recent LLM-powered toolkits, such as OpenAI Codex and Cursor, have offered end-to-end automation of the software development process. However, building effective SWE agents remains challenging due to the lack of high-quality training data and effective test cases. To address this issue, we present SWE-Dev, an SWE agent built upon open-source LLMs. First, we develop a robust pipeline to synthesize test cases for patch evaluation. Second, we scale up agent trajectories to construct the training data for building SWE-Dev. Experiments on the SWE-bench-Verified benchmark show that the SWE-Dev models can achieve top performance among all open SWE agents. Specifically, the success rates of the SWE-Dev 7B and 32B parameter models reach 23.4% and 36.6%, respectively, outperforming state-of-the-art open-source models. All code, models, and datasets are publicly available at https://github.com/THUDM/SWE-Dev.

cs.AI↗

arXiv2025

InspectionV3: Enhancing Tobacco Quality Assessment with Deep Convolutional Neural Networks for Automated Workshop Management

The problems that tobacco workshops encounter include poor curing, inconsistencies in supplies, irregular scheduling, and a lack of oversight, all of which drive up expenses and worse quality. Large quantities make manual examination costly, sluggish, and unreliable. Deep convolutional neural networks have recently made strides in capabilities that transcend those of conventional methods. To effectively enhance them, nevertheless, extensive customization is needed to account for subtle variations in tobacco grade. This study introduces InspectionV3, an integrated solution for automated flue-cured tobacco grading that makes use of a customized deep convolutional neural network architecture. A scope that covers color, maturity, and curing subtleties is established via a labelled dataset consisting of 21,113 images spanning 20 quality classes. Expert annotators performed preprocessing on the tobacco leaf images, including cleaning, labelling, and augmentation. Multi-layer CNN factors use batch normalization to describe domain properties like as permeability and moisture spots, and so account for the subtleties of the workshop. Its expertise lies in converting visual patterns into useful information for enhancing workflow. Fast notifications are made possible by real-time, on-the-spot grading that matches human expertise. Images-powered analytics dashboards facilitate the tracking of yield projections, inventories, bottlenecks, and the optimization of data-driven choices. More labelled images are assimilated after further retraining, improving representational capacities and enabling adaptations for seasonal variability. Metrics demonstrate 97% accuracy, 95% precision and recall, 96% F1-score and AUC, 95% specificity; validating real-world viability.

cs.CV↗

arXiv2025

AccioScene: Compositional 3D Scene Generation via Graph Diffusion and Interaction-driven Critics

This paper presents a framework for generating 3D indoor scenes from text prompts. Existing methods often formulate scene synthesis as an object layout prediction problem conditioned on a single input modality, such as a text description, room shape, or scene graph. This design can lead to object collisions and limited functional plausibility, reducing its practical applicability. To address these limitations, we introduce a multi-stage pipeline that better reflects practical scene creation scenarios. Given a text prompt describing partial scene content, our method first uses graph diffusion to produce a contextually coherent scene graph and then predicts a realistic object layout. In addition, we incorporate lightweight human-object interaction priors to encourage human-centric and functional arrangements, with explicit spatial constraints to reduce interpenetration. Our approach generates coherent 3D scenes with viable layouts that better support human interaction. Experiments on the 3D-FRONT dataset demonstrate that our method achieves competitive or state-of-the-art performance compared with existing approaches, while improving the physical plausibility of generated scenes.

cs.LG↗

arXiv2024

Quantum Embedding of Non-local Quantum Many-body Interactions in Prototypal Anti-tumor Vaccine Metalloprotein on Near Term Quantum Computing Hardware

The real world obeys quantum physics and quantum computing presents an alternative way to map physical problems to systems that follow the same laws. Such computation fundamentally constitutes a better way to understand the most challenging quantum problems. One such problem is the accurate simulation of highly correlated quantum systems. Due to the high dimensionality of the problem classical computers require considerable computer power to accurately predict material properties, especially when strong electron interactions are present. Still, modern day quantum hardware has many limitations and only allows for modeling of very simple systems. Here we present for the first time a quantum computer model simulation of a complex hemocyanin molecule, which is an important respiratory protein involved in various physiological processes such as oxygen transport and immune defence, and is also used as a key component in therapeutic vaccines for cancer. To better characterise the mechanism by which hemocyanin transports oxygen, variational quantum eigensolver (VQE) based on fermionic excitations and quantum embedding methods is used in the context of dynamic mean field theory to solve Anderson impurity model (AIM). Finally, it is concluded that the magnetic structure of hemocyanin is largely influenced by the many-body correction and that the computational effort for solving correlated electron systems could be substantially reduced with the introduction of quantum computing algorithms. We encourage the use of the Hamiltonian systems presented in this paper as a benchmark for testing quantum computing algorithms efficiency for chemistry applications.

quant-ph↗

arXiv2024

Global Optimization of Molybdenum Subnanoclusters on Graphene: a Consistent Approach Towards Catalytic Applications

The development of novel sub-nanometer clusters (SNCs) catalysts with superior catalytic performance depends on the precise control of clusters' atomistic sizes, shapes, and accurate deposition onto surfaces. The intrinsic complexity of the adsorption process complicates the ability to achieve an atomistic understanding of the most relevant structure-reactivity relationships hampering the rational design of novel catalytic materials. In most cases, existing computational approaches rely on just a few structures to draw conclusions on clusters' reactivity thereby neglecting the complexity of the existing energy landscapes thus leading to insufficient sampling and, most likely, unreliable predictions. Moreover, modelling of the actual experimental procedure that is responsible for the deposition of SNCs on surfaces is often not done even though in some cases this procedure may enhance the significance of certain (e.g., metastable) adsorption geometries. This study proposes a novel systematic approach that utilizes global search techniques, specifically, the particle swarm optimization (PSO) method, in conjunction with \textit{ab-initio} calculations, to simulate {\it all stages} in the beam experiments, from predicting the most relevant SNCs structures in the beam and on a surface, to their reactivity. To illustrate the main steps of our approach, we consider the deposition of Molybdenum SNC of 6 Mo atoms on a free-standing graphene surface, as well as their catalytic properties with respect to the CO molecule dissociation reaction. Even though our calculations are not exhaustive and serve only to produce an illustration of the method, they are still able to provide insight into the complicated energy landscape of Mo SNCs on graphene demonstrating the catalytic activity of Mo SNCs and the importance of performing statistical sampling of available configurations...

cond-mat.mtrl-sci↗

arXiv2024

Characterization of Spin-Orbit Effects in Superconductors In$_5$Bi$_3$ and In$_5$Sb$_3$

We report a first principles computational analysis of two phonon-mediated superconductors, In$_{5}$Bi$_{3}$ and In$_{5}$Sb$_{3}$. We show that spin-orbit coupling leads to splitting of electron bands around the Fermi energy, resulting in a suppression of the electronic density of states in both compounds. In In$_{5}$Bi$_{3}$, the spin-orbit coupling is essential for the dynamical stability of the experimentally observed phase, and the calculated superconducting critical temperature is in close agreement with measurements. In In$_{5}$Sb$_{3}$, the spin-orbit coupling significantly reduces the calculated superconducting critical temperature compared to calculations neglecting relativistic effects. Our work emphasises the subtle interplay between spin-orbit interactions and phonon-mediated superconductivity.

cond-mat.supr-con↗

arXiv2024

Planner3D: LLM-enhanced graph prior meets 3D indoor scene explicit regularization

Compositional 3D scene synthesis has diverse applications across a spectrum of industries such as robotics, films, and video games, as it closely mirrors the complexity of real-world multi-object environments. Conventional works typically employ shape retrieval based frameworks which naturally suffer from limited shape diversity. Recent progresses have been made in object shape generation with generative models such as diffusion models, which increases the shape fidelity. However, these approaches separately treat 3D shape generation and layout generation. The synthesized scenes are usually hampered by layout collision, which suggests that the scene-level fidelity is still under-explored. In this paper, we aim at generating realistic and reasonable 3D indoor scenes from scene graph. To enrich the priors of the given scene graph inputs, large language model is utilized to aggregate the global-wise features with local node-wise and edge-wise features. With a unified graph encoder, graph features are extracted to guide joint layout-shape generation. Additional regularization is introduced to explicitly constrain the produced 3D layouts. Benchmarked on the SG-FRONT dataset, our method achieves better 3D scene synthesis, especially in terms of scene-level fidelity. The source code will be released after publication.

cs.CV↗

arXiv2024

Phoneme-Based Proactive Anti-Eavesdropping with Controlled Recording Privilege

The widespread smart devices raise people's concerns of being eavesdropped on. To enhance voice privacy, recent studies exploit the nonlinearity in microphone to jam audio recorders with inaudible ultrasound. However, existing solutions solely rely on energetic masking. Their simple-form noise leads to several problems, such as high energy requirements and being easily removed by speech enhancement techniques. Besides, most of these solutions do not support authorized recording, which restricts their usage scenarios. In this paper, we design an efficient yet robust system that can jam microphones while preserving authorized recording. Specifically, we propose a novel phoneme-based noise with the idea of informational masking, which can distract both machines and humans and is resistant to denoising techniques. Besides, we optimize the noise transmission strategy for broader coverage and implement a hardware prototype of our system. Experimental results show that our system can reduce the recognition accuracy of recordings to below 50\% under all tested speech recognition systems, which is much better than existing solutions.

cs.CR↗

arXiv2023

Is Imitation All You Need? Generalized Decision-Making with Dual-Phase Training

We introduce DualMind, a generalist agent designed to tackle various decision-making tasks that addresses challenges posed by current methods, such as overfitting behaviors and dependence on task-specific fine-tuning. DualMind uses a novel "Dual-phase" training strategy that emulates how humans learn to act in the world. The model first learns fundamental common knowledge through a self-supervised objective tailored for control tasks and then learns how to make decisions based on different contexts through imitating behaviors conditioned on given prompts. DualMind can handle tasks across domains, scenes, and embodiments using just a single set of model weights and can execute zero-shot prompting without requiring task-specific fine-tuning. We evaluate DualMind on MetaWorld and Habitat through extensive experiments and demonstrate its superior generalizability compared to previous techniques, outperforming other generalist agents by over 50$\%$ and 70$\%$ on Habitat and MetaWorld, respectively. On the 45 tasks in MetaWorld, DualMind achieves over 30 tasks at a 90$\%$ success rate.

cs.AI↗

arXiv2023

A comprehensive exploration of structural and electronic properties of Molybdenum clusters

Molybdenum clusters, characterised by their unique structure and intriguing catalytic properties, have gained significant attention in recent years. In several existing studies density functional theory (DFT) methods have been used to find the lowest energy Mo clusters and explore their electronic and magnetic structure. In all cases, with the exception of a single recent study, where a genetic algorithm was employed, initial geometries of the clusters, prior to geometry optimisation, were chosen using heuristic approaches based on symmetry considerations and known structures. DFT calculations were performed using different types of pseudopotentials, from hard to soft, and different types of basis sets. However, no comprehensive study has yet been done in which a DFT method with the best control on its precision would be complemented by a reliable global minimum search method to find the lowest energy Mo clusters. In this work, we employ a combination of a plane wave-based DFT method and \emph{ab initio} random structure searching (AIRSS) technique to find the lowest energy clusters of up to 10 Mo atoms. In each case, the search has been performed for clusters with different spin multiplicities, which enabled us to explore their magnetic structure. The results are compared for both hard and soft pseudopotentials stressing the importance of treating more electrons explicitly, in agreement with some of the previous studies. For most of the low-energy magnetic structures found, we investigate the distribution of their spin densities, and for all low energy clusters, we confirm their stability by calculating their phonon structure. Finally, free energies of the Mo clusters, within the quasi-harmonic approximation, are also calculated and discussed.

cond-mat.mtrl-sci↗

arXiv2023

DFIL: Deepfake Incremental Learning by Exploiting Domain-invariant Forgery Clues

The malicious use and widespread dissemination of deepfake pose a significant crisis of trust. Current deepfake detection models can generally recognize forgery images by training on a large dataset. However, the accuracy of detection models degrades significantly on images generated by new deepfake methods due to the difference in data distribution. To tackle this issue, we present a novel incremental learning framework that improves the generalization of deepfake detection models by continual learning from a small number of new samples. To cope with different data distributions, we propose to learn a domain-invariant representation based on supervised contrastive learning, preventing overfit to the insufficient new data. To mitigate catastrophic forgetting, we regularize our model in both feature-level and label-level based on a multi-perspective knowledge distillation approach. Finally, we propose to select both central and hard representative samples to update the replay set, which is beneficial for both domain-invariant representation learning and rehearsal-based knowledge preserving. We conduct extensive experiments on four benchmark datasets, obtaining the new state-of-the-art average forgetting rate of 7.01 and average accuracy of 85.49 on FF++, DFDC-P, DFD, and CDF2. Our code is released at https://github.com/DeepFakeIL/DFIL.

cs.CV↗

arXiv2023

BuilDiff: 3D Building Shape Generation using Single-Image Conditional Point Cloud Diffusion Models

3D building generation with low data acquisition costs, such as single image-to-3D, becomes increasingly important. However, most of the existing single image-to-3D building creation works are restricted to those images with specific viewing angles, hence they are difficult to scale to general-view images that commonly appear in practical cases. To fill this gap, we propose a novel 3D building shape generation method exploiting point cloud diffusion models with image conditioning schemes, which demonstrates flexibility to the input images. By cooperating two conditional diffusion models and introducing a regularization strategy during denoising process, our method is able to synthesize building roofs while maintaining the overall structures. We validate our framework on two newly built datasets and extensive experiments show that our method outperforms previous works in terms of building generation quality.

cs.CV↗

arXiv2022

Attention Based Relation Network for Facial Action Units Recognition

Facial action unit (AU) recognition is essential to facial expression analysis. Since there are highly positive or negative correlations between AUs, some existing AU recognition works have focused on modeling AU relations. However, previous relationship-based approaches typically embed predefined rules into their models and ignore the impact of various AU relations in different crowds. In this paper, we propose a novel Attention Based Relation Network (ABRNet) for AU recognition, which can automatically capture AU relations without unnecessary or even disturbing predefined rules. ABRNet uses several relation learning layers to automatically capture different AU relations. The learned AU relation features are then fed into a self-attention fusion module, which aims to refine individual AU features with attention weights to enhance the feature robustness. Furthermore, we propose an AU relation dropout strategy and AU relation loss (AUR-Loss) to better model AU relations, which can further improve AU recognition. Extensive experiments show that our approach achieves state-of-the-art performance on the DISFA and DISFA+ datasets.

cs.CV↗