arXiv subjects

Lijie Liu

Publications and source records attributed to Lijie Liu.

At least 19 recordsLinked to original sources

arXiv2026

Spin relaxation in $X$-wave magnets with $X=p, d, f, g, i$

Spin relaxation results in the spin decoherence and a finite spin lifetime, which are detrimental to spintronic devices. To achieve a long spin lifetime desirable for spintronic devices, elucidating the spin relaxation mechanism and factors influencing the spin lifetime is of vital importance. Here, we investigate the spin relaxation in $X$-wave magnets ($X=p, d, f, g, i$) with Rashba spin-orbit coupling within the framework of D'yakonov-Perel' mechanism. We calculate the general matrix of the spin relaxation time for an arbitrary N\'eel vector direction of the $X$-wave magnet. As an illustration, we study the spin relaxation for the N\'eel vector along the $[001]$ direction. It is found that the reciprocal spin-relaxation-time matrices are anisotropic and diagonal for the $d$-, $f$-, $g$- and $i$-wave magnets. For the $p$-wave magnet, we derive the analytical expressions for the temporal evolution of spins. Moreover, the spin relaxation rate is proportional to the momentum relaxation time, Rashba and altermagnetic spin-split strengths for all $X$-wave magnets. Our results shine more light on the fundamental understanding of the spin relaxation mechanism in $X$-wave magnets.

cond-mat.other↗

arXiv2026

Speed by Simplicity: A Single-Stream Architecture for Fast Audio-Video Generative Foundation Model

We present daVinci-MagiHuman, an open-source audio-video generative foundation model for human-centric generation. daVinci-MagiHuman jointly generates synchronized video and audio using a single-stream Transformer that processes text, video, and audio within a unified token sequence via self-attention only. This single-stream design avoids the complexity of multi-stream or cross-attention architectures while remaining easy to optimize with standard training and inference infrastructure. The model is particularly strong in human-centric scenarios, producing expressive facial performance, natural speech-expression coordination, realistic body motion, and precise audio-video synchronization. It supports multilingual spoken generation across Chinese (Mandarin and Cantonese), English, Japanese, Korean, German, and French. For efficient inference, we combine the single-stream backbone with model distillation, latent-space super-resolution, and a Turbo VAE decoder, enabling generation of a 5-second 256p video in 2 seconds on a single H100 GPU. In automatic evaluation, daVinci-MagiHuman achieves the highest visual quality and text alignment among leading open models, along with the lowest word error rate (14.60%) for speech intelligibility. In pairwise human evaluation, it achieves win rates of 80.0% against Ovi 1.1 and 60.9% against LTX 2.3 over 2000 comparisons. We open-source the complete model stack, including the base model, the distilled model, the super-resolution model, and the inference codebase.

cs.CV↗

arXiv2026

LibraGen: Playing a Balance Game in Subject-Driven Video Generation

With the advancement of video generation foundation models (VGFMs), customized generation, particularly subject-to-video (S2V), has attracted growing attention. However, a key challenge lies in balancing the intrinsic priors of a VGFM, such as motion coherence, visual aesthetics, and prompt alignment, with its newly derived S2V capability. Existing methods often neglect this balance by enhancing one aspect at the expense of others. To address this, we propose LibraGen, a novel framework that views extending foundation models for S2V generation as a balance game between intrinsic VGFM strengths and S2V capability. Specifically, guided by the core philosophy of "Raising the Fulcrum, Tuning to Balance," we identify data quality as the fulcrum and advocate a quality-over-quantity approach. We construct a hybrid pipeline that combines automated and manual data filtering to improve overall data quality. To further harmonize the VGFM's native capabilities with its S2V extension, we introduce a Tune-to-Balance post-training paradigm. During supervised fine-tuning, both cross-pair and in-pair data are incorporated, and model merging is employed to achieve an effective trade-off. Subsequently, two tailored direct preference optimization (DPO) pipelines, namely Consis-DPO and Real-Fake DPO, are designed and merged to consolidate this balance. During inference, we introduce a time-dependent dynamic classifier-free guidance scheme to enable flexible and fine-grained control. Experimental results demonstrate that LibraGen outperforms both open-source and commercial S2V models using only thousand-scale training data.

cs.CV↗

arXiv2026

WISDOM Project - XXVII. Giant molecular clouds of the lenticular galaxy NGC 1387: similarities with spiral galaxy clouds

Molecular gas is crucial to understanding star formation and galaxy evolution, but the giant molecular clouds (GMCs) of early-type galaxies (ETGs) have rarely been studied. Here, we present analyses of the spatially resolved GMCs of the lenticular galaxy NGC 1387, exploiting high spatial resolution (0.15" or 14 pc) 12CO(2-1) line observations from the Atacama Large Millimeter/submillimeter Array. We identify 1285 individual GMCs and measure the fundamental properties (radius, velocity dispersion, and molecular gas mass) of each with a modified version of the CPROPStoo package. Unusually for an ETG, the GMCs of NGC 1387 follow scaling relations very similar to those of the Milky Way disc and Local Group galaxy clouds, and most are virialised. GMCs with large masses and radii and/or small galactocentric distances have their angular momenta aligned with the large-scale galactic rotation, while other GMCs do not. These results show that ETGs have more diversified GMC properties than previously thought. We discuss potential reasons for such diversity, and viewing-angle dependency is a plausible candidate.

astro-ph.GA↗

arXiv2025

OmniInsert: Mask-Free Video Insertion of Any Reference via Diffusion Transformer Models

Recent advances in video insertion based on diffusion models are impressive. However, existing methods rely on complex control signals but struggle with subject consistency, limiting their practical applicability. In this paper, we focus on the task of Mask-free Video Insertion and aim to resolve three key challenges: data scarcity, subject-scene equilibrium, and insertion harmonization. To address the data scarcity, we propose a new data pipeline InsertPipe, constructing diverse cross-pair data automatically. Building upon our data pipeline, we develop OmniInsert, a novel unified framework for mask-free video insertion from both single and multiple subject references. Specifically, to maintain subject-scene equilibrium, we introduce a simple yet effective Condition-Specific Feature Injection mechanism to distinctly inject multi-source conditions and propose a novel Progressive Training strategy that enables the model to balance feature injection from subjects and source video. Meanwhile, we design the Subject-Focused Loss to improve the detailed appearance of the subjects. To further enhance insertion harmonization, we propose an Insertive Preference Optimization methodology to optimize the model by simulating human preferences, and incorporate a Context-Aware Rephraser module during reference to seamlessly integrate the subject into the original scenes. To address the lack of a benchmark for the field, we introduce InsertBench, a comprehensive benchmark comprising diverse scenes with meticulously selected subjects. Evaluation on InsertBench indicates OmniInsert outperforms state-of-the-art closed-source commercial solutions. The code will be released.

cs.CV↗

arXiv2025

HuMo: Human-Centric Video Generation via Collaborative Multi-Modal Conditioning

Human-Centric Video Generation (HCVG) methods seek to synthesize human videos from multimodal inputs, including text, image, and audio. Existing methods struggle to effectively coordinate these heterogeneous modalities due to two challenges: the scarcity of training data with paired triplet conditions and the difficulty of collaborating the sub-tasks of subject preservation and audio-visual sync with multimodal inputs. In this work, we present HuMo, a unified HCVG framework for collaborative multimodal control. For the first challenge, we construct a high-quality dataset with diverse and paired text, reference images, and audio. For the second challenge, we propose a two-stage progressive multimodal training paradigm with task-specific strategies. For the subject preservation task, to maintain the prompt following and visual generation abilities of the foundation model, we adopt the minimal-invasive image injection strategy. For the audio-visual sync task, besides the commonly adopted audio cross-attention layer, we propose a focus-by-predicting strategy that implicitly guides the model to associate audio with facial regions. For joint learning of controllabilities across multimodal inputs, building on previously acquired capabilities, we progressively incorporate the audio-visual sync task. During inference, for flexible and fine-grained multimodal control, we design a time-adaptive Classifier-Free Guidance strategy that dynamically adjusts guidance weights across denoising steps. Extensive experimental results demonstrate that HuMo surpasses specialized state-of-the-art methods in sub-tasks, establishing a unified framework for collaborative multimodal-conditioned HCVG. Project Page: https://phantom-video.github.io/HuMo.

cs.CV↗

arXiv2025

Phantom-Data : Towards a General Subject-Consistent Video Generation Dataset

Subject-to-video generation has witnessed substantial progress in recent years. However, existing models still face significant challenges in faithfully following textual instructions. This limitation, commonly known as the copy-paste problem, arises from the widely used in-pair training paradigm. This approach inherently entangles subject identity with background and contextual attributes by sampling reference images from the same scene as the target video. To address this issue, we introduce \textbf{Phantom-Data, the first general-purpose cross-pair subject-to-video consistency dataset}, containing approximately one million identity-consistent pairs across diverse categories. Our dataset is constructed via a three-stage pipeline: (1) a general and input-aligned subject detection module, (2) large-scale cross-context subject retrieval from more than 53 million videos and 3 billion images, and (3) prior-guided identity verification to ensure visual consistency under contextual variation. Comprehensive experiments show that training with Phantom-Data significantly improves prompt alignment and visual quality while preserving identity consistency on par with in-pair baselines.

cs.CV↗

arXiv2025

Phantom: Subject-consistent video generation via cross-modal alignment

The continuous development of foundational models for video generation is evolving into various applications, with subject-consistent video generation still in the exploratory stage. We refer to this as Subject-to-Video, which extracts subject elements from reference images and generates subject-consistent videos following textual instructions. We believe that the essence of subject-to-video lies in balancing the dual-modal prompts of text and image, thereby deeply and simultaneously aligning both text and visual content. To this end, we propose Phantom, a unified video generation framework for both single- and multi-subject references. Building on existing text-to-video and image-to-video architectures, we redesign the joint text-image injection model and drive it to learn cross-modal alignment via text-image-video triplet data. The proposed method achieves high-fidelity subject-consistent video generation while addressing issues of image content leakage and multi-subject confusion. Evaluation results indicate that our method outperforms other state-of-the-art closed-source commercial solutions. In particular, we emphasize subject consistency in human generation, covering existing ID-preserving video generation while offering enhanced advantages.

cs.CV↗

arXiv2025

The HASHTAG project II. Giant molecular cloud properties across the M31 disc

We present a study of giant molecular cloud (GMC) properties in the Andromeda galaxy (M31) using CO(3-2) data from the James Clerk Maxwell Telescope (JCMT) in selected regions across the disc and in the nuclear ring, and comparing them with CO(1-0) observations from the IRAM 30m telescope in the same regions. We find that GMCs in the centre of M31 generally exhibit larger velocity dispersions ($σ$) and sizes ($R$) compared to those in the disc, while their average surface density ($Σ$) and turbulent pressure ($P_{\rm turb}$) are lower. This low turbulent pressure in the central region is primarily due to the low density of molecular gas. The estimated GMC properties depend on the choice of CO transitions. Compared to CO(1-0), CO(3-2) exhibits smaller velocity dispersion and equivalent radius but higher surface density. These differences highlight the distinct physical conditions probed by different molecular gas tracers. We estimate the virial parameter $α_{\rm vir}\propto σ^2 R/Σ$ and find that most molecular clouds exhibit high values ($α_{\rm vir} \sim 4-6$) for both CO transitions, indicating that they are unbound. Furthermore, clouds in the nuclear ring display even larger $α_{\rm vir}$ values of $\lesssim 100$, suggesting that they may be highly dynamic, short-lived structures, although they could potentially achieve equilibrium under the external pressure exerted by the surrounding interstellar medium.

astro-ph.GA↗

arXiv2025

Inadequate turbulent support in low-metallicity molecular clouds

The dynamic properties of molecular clouds are set by the interplay of their self-gravity, turbulence, external pressure and magnetic fields. Extended surveys of Galactic molecular clouds typically find that their kinetic energy ($E_{\rm k}$) counterbalances their self-gravitational energy ($E_{\rm g}$), setting their virial parameter $α_{\rm vir}=2E_{\rm k}/|E_{\rm g}|\approx1$. However, past studies either have been biased by the use of optically-thick lines or have been limited within the solar neighborhood and the inner Galaxy (Galactocentric radius $R_{\rm gc}<R_{\rm gc,\odot} \approx 8$ kpc). Here we present sensitive mapping observations of optically thin $^{13}$CO lines towards molecular clouds in the low-metallicity Galactic outer disk ($R_{\rm gc}\sim9-24$ kpc). By combining archival data from the inner Galaxy and four nearby metal-poor dwarf galaxies, we reveal a systematic trend of $α_{\rm vir}$, which declines from supervirial dynamic states in metal-rich clouds to extremely subvirial dynamic states in metal-poor clouds. In these metal-poor environments, turbulence alone is insufficient to counterbalance the self-gravity of a cloud. A cloud-volumetric magnetic field may replace turbulence as the dominant cloud-supporting mechanism in low-metallicity conditions, for example, the outermost galactic disks, dwarf galaxies and galaxies in the early Universe, which would then inevitably impact the initial conditions for star formation in such environments.

astro-ph.GA↗

arXiv2024

WISDOM project XX -- Strong shear tearing molecular clouds apart in NGC 524

Early-type galaxies (ETGs) are known to harbour dense spheroids of stars but scarce star formation (SF). Approximately a quarter of these galaxies have rich molecular gas reservoirs yet do not form stars efficiently. We study here the ETG NGC~524, with strong shear suspected to result in a smooth molecular gas disc and low star-formation efficiency (SFE). We present new spatially-resolved observations of the \textsuperscript{12}CO(2-1)-emitting cold molecular gas from the Atacama Large Millimeter/sub-millimeter Array (ALMA) and of the warm ionised-gas emission lines from SITELLE at the Canada-France-Hawaii Telescope. Although constrained by the resolution of the ALMA observations ($\approx37$~pc), we identify only $52$ GMCs with radii ranging from $30$ to $140$~pc, a low mean molecular gas mass surface density $\langleΣ_{\rm gas}\rangle\approx125$~M$_\odot$~pc$^{-2}$ and a high mean virial parameter $\langleα_{\rm obs,vir}\rangle\approx5.3$. We measure spatially-resolved molecular gas depletion times ($τ_{\rm dep}\equiv1/{\rm SFE}$) with a spatial resolution of $\approx100$~pc within a galactocentric distance of $1.5$~kpc. The global depletion time is $\approx2.0$~Gyr but $τ_{\rm dep}$ increases toward the galaxy centre, with a maximum $τ_{\rm dep,max}\approx5.2$~Gyr. However, no pure \ion{H}{II} region is identified in NGC~524 using ionised-gas emission-line ratio diagnostics, so the $τ_{\rm dep}$ inferred are in fact lower limits. Measuring the GMC properties and dynamical states, we conclude that shear is the dominant mechanism shaping the molecular gas properties and regulating SF in NGC~524. This is supported by analogous analyses of the GMCs in a simulated ETG similar to NGC~524.

astro-ph.GA↗

arXiv2024

WISDOM Project -- XXI. Giant molecular clouds in the central region of the barred spiral galaxy NGC 613: a steep size -- linewidth relation

NGC~613 is a nearby barred spiral galaxy with a nuclear ring. Exploiting high spatial resolution ($\approx20$ pc) Atacama Large Millimeter/sub-millimeter Array $^{12}$CO(1-0) observations, we study the giant molecular clouds (GMCs) in the nuclear ring and its vicinity, identifying $158$ spatially- and spectrally-resolved GMCs. The GMC sizes ($R_{\mathrm{c}}$) are comparable to those of the clouds in the Milky Way (MW) disc, but their gas masses, observed linewidths ($σ_{\mathrm{obs,los}}$) and gas mass surface densities are larger. The GMC size -- linewidth relation ($σ_{\mathrm{obs,los}}\propto R_{\mathrm{c}}^{0.77}$) is steeper than that of the clouds of the MW disc and centre, and the GMCs are on average only marginally gravitationally bound (with a mean virial parameter $\langleα_{\mathrm{obs,vir}}\rangle\approx1.7$). We discuss the possible origins of the steep size -- linewidth relation and enhanced observed linewidths of the clouds and suggest that a combination of mechanisms such as stellar feedback, gas accretion and cloud-cloud collisions, as well as the gas inflows driven by the large-scale bar, may play a role.

astro-ph.GA↗

arXiv2023

WISDOM Project -- XVI. The link between circumnuclear molecular gas reservoirs and active galactic nucleus fuelling

We use high-resolution data from the millimetre-Wave Interferometric Survey of Dark Object Masses (WISDOM) project to investigate the connection between circumnuclear gas reservoirs and nuclear activity in a sample of nearby galaxies. Our sample spans a wide range of nuclear activity types including radio galaxies, Seyfert galaxies, low-luminosity active galactic nuclei (AGN) and inactive galaxies. We use measurements of nuclear millimetre continuum emission along with other archival tracers of AGN accretion/activity to investigate previous claims that at, circumnuclear scales (<100 pc), these should correlate with the mass of the cold molecular gas. We find that the molecular gas mass does not correlate with any tracer of nuclear activity. This suggests the level of nuclear activity cannot solely be regulated by the amount of cold gas around the supermassive black hole (SMBH). This indicates that AGN fuelling, that drives gas from the large scale galaxy to the nuclear regions, is not a ubiquitous process and may vary between AGN type, with timescale variations likely to be very important. By studying the structure of the central molecular gas reservoirs, we find our galaxies have a range of nuclear molecular gas concentrations. This could indicate that some of our galaxies may have had their circumnuclear regions impacted by AGN feedback, even though they currently have low nuclear activity. On the other hand, the nuclear molecular gas concentrations in our galaxies could instead be set by secular processes.

astro-ph.GA↗

arXiv2023

A fundamental plane of black hole accretion at millimetre wavelengths

We report the discovery of the ``mm fundamental plane of black-hole accretion'', which is a tight correlation between the nuclear 1 mm luminosity ($L_{\rm ν, mm}$), the intrinsic $2$ -- $10$~keV X-ray luminosity ($L_{\rm X,2-10}$) and the supermassive black hole (SMBH) mass ($M_{\rm BH}$) with an intrinsic scatter ($σ_{\rm int}$) of $0.40$ dex. The plane is found for a sample of 48 nearby galaxies, most of which are low-luminosity active galactic nuclei (LLAGN). Combining these sources with a sample of high-luminosity (quasar-like) nearby AGN, we find that the plane still holds. We also find that $M_{\rm BH}$ correlates with $L_{\rm ν, mm}$ at a highly significant level, although such correlation is less tight than the mm fundamental plane ($σ_{\rm int}=0.51$ dex). Crucially, we show that spectral energy distribution (SED) models for both advection-dominated accretion flows (ADAFs) and compact jets can explain the existence of these relations, which are not reproduced by the standard torus-thin accretion disc models usually associated to quasar-like AGN. The ADAF models reproduces the observed relations somewhat better than those for compact jets, although neither provides a perfect prediction. Our findings thus suggest that radiatively-inefficient accretion processes such as those in ADAFs or compact (and thus possibly young) jets may play a key role in both low- and high-luminosity AGN. This mm fundamental plane also offers a new, rapid method to (indirectly) estimate SMBH masses.

astro-ph.GA↗

arXiv2023

WISDOM project -- XIV. SMBH mass in the early-type galaxies NGC0612, NGC1574, and NGC4261 from CO dynamical modelling

We present a CO dynamical estimate of the mass of the super-massive black hole (SMBH) in three nearby early-type galaxies: NGC0612, NGC1574 and NGC4261. Our analysis is based on Atacama Large Millimeter/submillimeter Array (ALMA) Cycle 3-6 observations of the $^{12}$CO(2-1) emission line with spatial resolutions of $14-58$ pc ($0.01"-0.26"$). We detect disc-like CO distributions on scales from $\lesssim200$ pc (NGC1574 and NGC4261) to $\approx10$ kpc (NGC0612). In NGC0612 and NGC1574 the bulk of the gas is regularly rotating. The data also provide evidence for the presence of a massive dark object at the centre of NGC1574, allowing us to obtain the first measure of its mass, $M_{\rm BH}=(1.0\pm0.2)\times10^{8}$ M$_{\odot}$ (1$σ$ uncertainty). In NGC4261, the CO kinematics is clearly dominated by the SMBH gravitational influence, allowing us to determine an accurate black hole mass of $(1.62{\pm 0.04})\times10^{9}$ M$_{\odot}$ ($1σ$ uncertainty). This is fully consistent with a previous CO dynamical estimate obtained using a different modelling technique. Signs of non-circular gas motions (likely outflow) are also identified in the inner regions of NGC4261. In NGC0612, we are only able to obtain a (conservative) upper limit of $M_{\rm BH}\lesssim3.2\times10^{9}$ M$_{\odot}$. This has likely to be ascribed to the presence of a central CO hole (with a radius much larger than that of the SMBH sphere of influence), combined with the inability of obtaining a robust prediction for the CO velocity curve. The three SMBH mass estimates are overall in agreement with predictions from the $M_{\rm BH}-σ_{\star}$ relation.

astro-ph.GA↗

arXiv2023

The first ground-based detection of the 752 GHz water line in local ultra-luminous infrared galaxies using APEX-SEPIA

We report the first ground-based detection of the water line p-H2O (211-202) at 752.033 GHz in three z < 0.08 ultra-luminous infrared galaxies (ULIRGs): IRAS 06035-7102, IRAS 17207-0014 and IRAS 09022-3615. Using the Atacama Pathfinder EXperiment (APEX), with its Swedish-ESO PI Instrument for APEX (SEPIA) band-9 receiver, we detect this H2O line with overall signal-to-noise ratios of 8-10 in all three galaxies. Notably, this is the first detection of this line in IRAS 06035-7102. Our new APEX-measured fluxes, between 145 to 705 Jy km s-1, are compared with previous values taken from Herschel SPIRE FTS. We highlight the great capabilities of APEX for resolving the H2O line profiles with high spectral resolutions while also improving by a factor of two the significance of the detection within moderate integration times. While exploring the correlation between the p-H2O(211-202) and the total infrared luminosity, our galaxies are found to follow the trend at the bright end of the local ULIRG's distribution. The p-H2O(211-202) line spectra are compared to the mid-J CO and HCN spectra, and dust continuum previously observed with ALMA. In the complex interacting system IRAS 09022-3615, the profile of the water emission line is offset in velocity with respect to the ALMA CO(J = 4 - 3) emission. For IRAS 17207-0014 and IRAS 06035-7102, the profiles between the water line and the CO lines are spectroscopically aligned. This pilot study demonstrates the feasibility of directly conducting ground-based high-frequency observations of this key water line, opening the possibility of detailed follow-up campaigns to tackle its nature.

astro-ph.GA↗

arXiv2023

HS-Diffusion: Semantic-Mixing Diffusion for Head Swapping

Image-based head swapping task aims to stitch a source head to another source body flawlessly. This seldom-studied task faces two major challenges: 1) Preserving the head and body from various sources while generating a seamless transition region. 2) No paired head swapping dataset and benchmark so far. In this paper, we propose a semantic-mixing diffusion model for head swapping (HS-Diffusion) which consists of a latent diffusion model (LDM) and a semantic layout generator. We blend the semantic layouts of source head and source body, and then inpaint the transition region by the semantic layout generator, achieving a coarse-grained head swapping. Semantic-mixing LDM can further implement a fine-grained head swapping with the inpainted layout as condition by a progressive fusion process, while preserving head and body with high-quality reconstruction. To this end, we propose a semantic calibration strategy for natural inpainting and a neck alignment for geometric realism. Importantly, we construct a new image-based head swapping benchmark and design two tailor-designed metrics (Mask-FID and Focal-FID). Extensive experiments demonstrate the superiority of our framework. The code will be available: https://github.com/qinghew/HS-Diffusion.

cs.CV↗

arXiv2023

Energy-Efficient WiFi Backscatter Communication for Green IoTs

The boom of the Internet of Things has revolutionized people's lives, but it has also resulted in massive resource consumption and environmental pollution. Recently, Green IoT (GIoT) has become a worldwide consensus to address this issue. In this paper, we propose EEWScatter, an energy-efficient WiFi backscatter communication system to pursue the goal of GIoT. Unlike previous backscatter systems that solely focus on tags, our approach offers a comprehensive system-wide view on energy conservation. Specifically, we reuse ambient signals as carriers and utilize an ultra-low-power and battery-free design for tag nodes by backscatter. Further, we design a new CRC-based algorithm that enables the demodulation of both ambient and tag data by only a single receiver while using ambient carriers. Such a design eliminates system reliance on redundant transceivers with high power consumption. Results demonstrate that EEWScatter achieves the lowest overall system power consumption and saves at least half of the energy. What's more, the power consumption of our tag is only 1/1000 of that of active radio. We believe that EEWScatter is a critical step towards a sustainable future.

cs.NI↗