arXiv subjects

Feng Liu

Publications and source records attributed to Feng Liu.

At least 343 records · Page 19Linked to original sources

arXiv2024

SongCreator: Lyrics-based Universal Song Generation

Music is an integral part of human culture, embodying human intelligence and creativity, of which songs compose an essential part. While various aspects of song generation have been explored by previous works, such as singing voice, vocal composition and instrumental arrangement, etc., generating songs with both vocals and accompaniment given lyrics remains a significant challenge, hindering the application of music generation models in the real world. In this light, we propose SongCreator, a song-generation system designed to tackle this challenge. The model features two novel designs: a meticulously designed dual-sequence language model (DSLM) to capture the information of vocals and accompaniment for song generation, and a series of attention mask strategies for DSLM, which allows our model to understand, generate and edit songs, making it suitable for various songrelated generation tasks by utilizing specific attention masks. Extensive experiments demonstrate the effectiveness of SongCreator by achieving state-of-the-art or competitive performances on all eight tasks. Notably, it surpasses previous works by a large margin in lyrics-to-song and lyrics-to-vocals. Additionally, it is able to independently control the acoustic conditions of the vocals and accompaniment in the generated song through different audio prompts, exhibiting its potential applicability. Our samples are available at https://thuhcsi.github.io/SongCreator/.

cs.SD↗

arXiv2024

Predicting the Encoding Error of SIRENs

Implicit Neural Representations (INRs), which encode signals such as images, videos, and 3D shapes in the weights of neural networks, are becoming increasingly popular. Among their many applications is signal compression, for which there is great interest in achieving the highest possible fidelity to the original signal subject to constraints such as neural network size, training (encoding) and inference (decoding) time. But training INRs can be a computationally expensive process, making it challenging to determine the best possible tradeoff under such constraints. Towards this goal, we present a method which predicts the encoding error that a popular INR network (SIREN) will reach, given its network hyperparameters and the signal to encode. This method is trained on a unique dataset of 300,000 SIRENs, trained across a variety of images and hyperparameters. (Dataset available here: https://huggingface.co/datasets/predict-SIREN-PSNR/COIN-collection.) Our predictive method demonstrates the feasibility of this regression problem, and allows users to anticipate the encoding error that a SIREN network will reach in milliseconds instead of minutes or longer. We also provide insights into the behavior of SIREN networks, such as why narrow SIRENs can have very high random variation in encoding error, and how the performance of SIRENs relates to JPEG compression.

cs.CV↗

arXiv2024

Fingerprints of Super Resolution Networks

Several recent studies have demonstrated that deep-learning based image generation models, such as GANs, can be uniquely identified, and possibly even reverse-engineered, by the fingerprints they leave on their output images. We extend this research to single image super-resolution (SISR) networks. Compared to previously studied models, SISR networks are a uniquely challenging class of image generation model from which to extract and analyze fingerprints, as they can often generate images that closely match the corresponding ground truth and thus likely leave little flexibility to embed signatures. We take SISR models as examples to investigate if the findings from the previous work on fingerprints of GAN-based networks are valid for general image generation models. We show that SISR networks with a high upscaling factor or trained using adversarial loss leave highly distinctive fingerprints, and that under certain conditions, some SISR network hyperparameters can be reverse-engineered from these fingerprints.

cs.CV↗

arXiv2024

Longest cycles and longest chordless cycles in $2$-connected graphs

Thomassen's chord conjecture from 1976 states that every longest cycle in a $3$-connected graph has a chord. The circumference $c(G)$ and induced circumference $c'(G)$ of a graph $G$ are the length of its longest cycles and the length of its longest chordless cycles, respectively. In $2017$, Harvey proposed a stronger conjecture: Every $2$-connected graph $G$ with minimum degree at least $3$ has $c(G)\geq c'(G)+2$. This conjecture implies Thomassen's chord conjecture. We observe that wheels are the unique hamiltonian graphs for which the circumference and the induced circumference differ by exactly one. Thus we need only consider non-hamiltonian graphs for Harvey's conjecture. In this paper, we propose a conjecture involving wheels that is equivalent to Harvey's conjecture on non-hamiltonian graphs. A graph is $\ell$-holed if its all holes have length exactly $\ell$. Furthermore, we prove that Harvey's conjecture holds for $\ell$-holed graphs and graphs with a small induced circumference. Consequently, Thomassen's conjecture also holds for this two classes of graphs.

math.CO↗

arXiv2024

Mind the Gap Between Prototypes and Images in Cross-domain Finetuning

In cross-domain few-shot classification (CFC), recent works mainly focus on adapting a simple transformation head on top of a frozen pre-trained backbone with few labeled data to project embeddings into a task-specific metric space where classification can be performed by measuring similarities between image instance and prototype representations. Technically, an assumption implicitly adopted in such a framework is that the prototype and image instance embeddings share the same representation transformation. However, in this paper, we find that there naturally exists a gap, which resembles the modality gap, between the prototype and image instance embeddings extracted from the frozen pre-trained backbone, and simply applying the same transformation during the adaptation phase constrains exploring the optimal representations and shrinks the gap between prototype and image representations. To solve this problem, we propose a simple yet effective method, contrastive prototype-image adaptation (CoPA), to adapt different transformations respectively for prototypes and images similarly to CLIP by treating prototypes as text prompts. Extensive experiments on Meta-Dataset demonstrate that CoPA achieves the state-of-the-art performance more efficiently. Meanwhile, further analyses also indicate that CoPA can learn better representation clusters, enlarge the gap, and achieve minimal validation loss at the enlarged gap.

cs.CV↗

arXiv2024

Measurement of gas properties for the ion-TPC of N$ν$DEx experiment

In the N$ν$DEx collaboration, a high-pressure gas TPC is being developed to search for the neutrinoless double beta decay. The use of electronegative $\mathrm{^{82}SeF_{6}}$ gas mandates an ion-TPC. The reconstruction of $z$ coordinate is to be realized exploiting the feature of multiple species of charge carriers. As the initial stage of the development, we studied the properties of the $\mathrm{SF_{6}}$ gas, which is non-toxic and has similar molecular structure to $\mathrm{SeF_{6}}$. In the paper we present the measurement of drift velocities and mobilities of the majority and minority negative charge carriers found in $\mathrm{SF_{6}}$ at a pressure of 750 Torr, slightly higher than the local atmospheric pressure. The reduced fields range between 3.0 and 5.5 Td. It was performed using a laser beam to ionize the gas inside a small TPC, with a drift length of 3.7 cm. A customized charge sensitive amplifier was developed to read out the anode signals induced by the slowly drifting ions. The reconstruction of $z$ coordinate using the difference in the velocities of the two carriers was also demonstrated.

physics.ins-det↗

arXiv2024

La$_2$O$_3$Mn$_2$Se$_2$: a correlated insulating layered d-wave altermagnet

Altermagnets represent a new class of magnetic phases without net magnetization that are invariant under a combination of rotation and time reversal. Unlike conventional collinear antiferromagnets (AFM), altermagnets could lead to new correlated states and important material properties deriving from their non-relativistic spin-split band structure. Indeed, they are the magnetic analogue of unconventional superconductors and can yield spin polarized electrical currents in the absence of external magnetic fields, making them promising candidates for next-generation spintronics. Here, we report altermagnetism in the correlated insulator, magnetically-ordered tetragonal oxychalcogenide, La$_2$O$_3$Mn$_2$Se$_2$. Symmetry analysis reveals a $\mathit{d}_{x^2 - y^2}$-wave type spin momentum locking, which is supported by density functional theory (DFT) calculations. Magnetic measurements confirm the AFM transition below $\sim$166 K while neutron pair distribution function analysis reveals a 2D short-range magnetic order that persists above the Néel temperature. Single crystals are grown and characterized using X-ray diffraction, optical and electron microscopy, and microRaman spectroscopy to confirm the crystal structure, stoichiometry, and uniformity.

cond-mat.mtrl-sci↗

arXiv2024

DreamVideo-2: Zero-Shot Subject-Driven Video Customization with Precise Motion Control

Recent advances in customized video generation have enabled users to create videos tailored to both specific subjects and motion trajectories. However, existing methods often require complicated test-time fine-tuning and struggle with balancing subject learning and motion control, limiting their real-world applications. In this paper, we present DreamVideo-2, a zero-shot video customization framework capable of generating videos with a specific subject and motion trajectory, guided by a single image and a bounding box sequence, respectively, and without the need for test-time fine-tuning. Specifically, we introduce reference attention, which leverages the model's inherent capabilities for subject learning, and devise a mask-guided motion module to achieve precise motion control by fully utilizing the robust motion signal of box masks derived from bounding boxes. While these two components achieve their intended functions, we empirically observe that motion control tends to dominate over subject learning. To address this, we propose two key designs: 1) the masked reference attention, which integrates a blended latent mask modeling scheme into reference attention to enhance subject representations at the desired positions, and 2) a reweighted diffusion loss, which differentiates the contributions of regions inside and outside the bounding boxes to ensure a balance between subject and motion control. Extensive experimental results on a newly curated dataset demonstrate that DreamVideo-2 outperforms state-of-the-art methods in both subject customization and motion control. The dataset, code, and models will be made publicly available.

cs.CV↗

arXiv2024

A fluctuation-sensitivity-timescale trade-off in feedback-controlled dynamics

Feedback control is a renowned mechanism for attenuating intrinsic fluctuation in regulatory networks. However, its impact on the response sensitivity to external signals and the response timescale, which are also critical for signal transmission, has yet to be understood. In this letter, we study a general feedback-controlled network in which the feedback is achieved by a complex interactive module. By comparing the solution of Langevin equations with and without feedback, we analytically derive a fundamental trade-off between fluctuation, sensitivity, and timescale altered by the feedback. We show that feedback control cannot infinitely suppress fluctuation without the cost of reducing sensitivity or response speed. Furthermore, the lower bound for this trade-off can be reduced up to half in non-gradient dynamical systems compared to gradient systems. We validate this trade-off as a tight bound for high-dimensional systems in nonlinear regime through numerical simulations. These results elucidate the fundamental limitation of feedback control in enhancing the information transmission capacity of regulatory networks.

physics.bio-ph↗

arXiv2024

The minimum size and maximum diameter of an edge-pancyclic graph of a given order

A $k$-cycle in a graph is a cycle of length $k.$ A graph $G$ of order $n$ is called edge-pancyclic if for every integer $k$ with $3\le k\le n,$ every edge of $G$ lies in a $k$-cycle. It seems difficult to determine the minimum size $f(n)$ of a simple edge-pancyclic graph of order $n.$ We give lower and upper bounds on $f(n),$ and determine the maximum diameter of such a graph. In the $3$-connected case, the precise value of $f(n)$ is determined. We also determine the minimum size of a graph of a given order with connectivity conditions in which every edge lies in a triangle.

math.CO↗

arXiv2024

HARIVO: Harnessing Text-to-Image Models for Video Generation

We present a method to create diffusion-based video models from pretrained Text-to-Image (T2I) models. Recently, AnimateDiff proposed freezing the T2I model while only training temporal layers. We advance this method by proposing a unique architecture, incorporating a mapping network and frame-wise tokens, tailored for video generation while maintaining the diversity and creativity of the original T2I model. Key innovations include novel loss functions for temporal smoothness and a mitigating gradient sampling technique, ensuring realistic and temporally consistent video generation despite limited public video data. We have successfully integrated video-specific inductive biases into the architecture and loss functions. Our method, built on the frozen StableDiffusion model, simplifies training processes and allows for seamless integration with off-the-shelf models like ControlNet and DreamBooth. project page: https://kwonminki.github.io/HARIVO

cs.CV↗

arXiv2024

Optimal Control in Both Steady State and Transient Process with Unknown Disturbances

The scheme of online optimization as a feedback controller is widely used to steer the states of a physical system to the optimal solution of a predefined optimization problem. Such methods focus on regulating the physical states to the optimal solution in the steady state, without considering the performance during the transient process. In this paper, we simultaneously consider the performance in both the steady state and the transient process of a linear time-invariant system with unknown disturbances. The performance of the transient process is illustrated by the concept of overtaking optimality. An overtaking optimal controller with known disturbances is derived to achieve the transient overtaking optimality while guaranteeing steady-state performance. Then, we propose a disturbance independent near-optimal controller, which can achieve optimal steady-state performance and approach the overtaking optimal performance in the transient process. The system performance gap between the overtaking optimal controller and the proposed controller proves to be inversely proportional to the control gains. A case study on a power system with four buses is used to validate the effectiveness of the two controllers.

eess.SY↗

arXiv2024

All-optical ultrafast arbitrary rotation of hole orbital qubits with direct phase control

Complete quantum control of a stationary quantum bit embedded in a quantum emitter is crucial for photonic quantum information technologies. Recently, the orbital degree of freedom in optically active quantum dots has emerged as a promising candidate. However, the essential ability to perform arbitrary rotations on orbital qubits remains elusive. Here, we demonstrate arbitrary rotation of a hole orbital qubit with direct phase control using picosecond optical pulses. This is achieved by successfully inducing stimulated Raman transitions within $Λ$ systems coupled via radiative Auger processes. The new capability enables direct control of polar and azimuth angles of the Bloch vector without requiring timed precession. Our results establish orbital states in solid-state quantum emitters as a viable resource for applications in high-speed quantum information processing.

quant-ph↗

arXiv2024

Deep Learning Enhanced Quantum Holography with Undetected Photons

Holography is an essential technique of generating three-dimensional images. Recently, quantum holography with undetected photons (QHUP) has emerged as a groundbreaking method capable of capturing complex amplitude images. Despite its potential, the practical application of QHUP has been limited by susceptibility to phase disturbances, low interference visibility, and limited spatial resolution. Deep learning, recognized for its ability in processing complex data, holds significant promise in addressing these challenges. In this report, we present an ample advancement in QHUP achieved by harnessing the power of deep learning to extract images from single-shot holograms, resulting in vastly reduced noise and distortion, alongside a notable enhancement in spatial resolution. The proposed and demonstrated deep learning QHUP (DL-QHUP) methodology offers a transformative solution by delivering high-speed imaging, improved spatial resolution, and superior noise resilience, making it suitable for diverse applications across an array of research fields stretching from biomedical imaging to remote sensing. DL-QHUP signifies a crucial leap forward in the realm of holography, demonstrating its immense potential to revolutionize imaging capabilities and pave the way for advancements in various scientific disciplines. The integration of DL-QHUP promises to unlock new possibilities in imaging applications, transcending existing limitations and offering unparalleled performance in challenging environments.

physics.optics↗

arXiv2024

Turbulent Accelerating Combusting Flows with a Methane-Vitiated Air Flamelet Model

This work presents a numerical study of a diffusion flame in a reacting, two-dimensional, turbulent, viscous, multi-component, compressible mixing layer subject to a large favorable streamwise pressure gradient. The boundary-layer equations are solved coupled with both the $k$-$ω$ and SST turbulence models. A compressible extension of the flamelet progress variable method has been proposed and tested for use with large eddy simulations or Reynolds-averaged Navier-Stokes analyses of the burning of methane in pure air and vitiated air; the latter being particularly relevant in turbine burner scenarios. Effects of the level of detail of the reaction mechanism on the sub-grid and resolved-scale computations are studied. A comparison is made with results obtained using a simplified one-step reaction. The numerical results employing the flamelet model with the more detailed reaction mechanism show faster chemistry, significantly reduced peak temperatures and stronger sensitivity to pressure. Vitiated air flames are found to be dominated by unstable solutions, resulting in a weak flame with substantially lower peak temperature and impeded development, struggling to persist without quenching.

physics.flu-dyn↗

arXiv2024

AdaPPA: Adaptive Position Pre-Fill Jailbreak Attack Approach Targeting LLMs

Jailbreak vulnerabilities in Large Language Models (LLMs) refer to methods that extract malicious content from the model by carefully crafting prompts or suffixes, which has garnered significant attention from the research community. However, traditional attack methods, which primarily focus on the semantic level, are easily detected by the model. These methods overlook the difference in the model's alignment protection capabilities at different output stages. To address this issue, we propose an adaptive position pre-fill jailbreak attack approach for executing jailbreak attacks on LLMs. Our method leverages the model's instruction-following capabilities to first output pre-filled safe content, then exploits its narrative-shifting abilities to generate harmful content. Extensive black-box experiments demonstrate our method can improve the attack success rate by 47% on the widely recognized secure model (Llama2) compared to existing approaches. Our code can be found at: https://github.com/Yummy416/AdaPPA.

cs.CR↗

arXiv2024

An End-to-End Approach for Chord-Conditioned Song Generation

The Song Generation task aims to synthesize music composed of vocals and accompaniment from given lyrics. While the existing method, Jukebox, has explored this task, its constrained control over the generations often leads to deficiency in music performance. To mitigate the issue, we introduce an important concept from music composition, namely chords, to song generation networks. Chords form the foundation of accompaniment and provide vocal melody with associated harmony. Given the inaccuracy of automatic chord extractors, we devise a robust cross-attention mechanism augmented with dynamic weight sequence to integrate extracted chord information into song generations and reduce frame-level flaws, and propose a novel model termed Chord-Conditioned Song Generator (CSG) based on it. Experimental evidence demonstrates our proposed method outperforms other approaches in terms of musical performance and control precision of generated songs.

cs.SD↗

arXiv2024

Spatial Craving Patterns in Marijuana Users: Insights from fMRI Brain Connectivity Analysis with High-Order Graph Attention Neural Networks

The excessive consumption of marijuana can induce substantial psychological and social consequences. In this investigation, we propose an elucidative framework termed high-order graph attention neural networks (HOGANN) for the classification of Marijuana addiction, coupled with an analysis of localized brain network communities exhibiting abnormal activities among chronic marijuana users. HOGANN integrates dynamic intrinsic functional brain networks, estimated from functional magnetic resonance imaging (fMRI), using graph attention-based long short-term memory (GAT-LSTM) to capture temporal network dynamics. We employ a high-order attention module for information fusion and message passing among neighboring nodes, enhancing the network community analysis. Our model is validated across two distinct data cohorts, yielding substantially higher classification accuracy than benchmark algorithms. Furthermore, we discern the most pertinent subnetworks and cognitive regions affected by persistent marijuana consumption, indicating adverse effects on functional brain networks, particularly within the dorsal attention and frontoparietal networks. Intriguingly, our model demonstrates superior performance in cohorts exhibiting prolonged dependence, implying that prolonged marijuana usage induces more pronounced alterations in brain networks. The model proficiently identifies craving brain maps, thereby delineating critical brain regions for analysis

q-bio.NC↗