arXiv subjects

Pengcheng Li

Publications and source records attributed to Pengcheng Li.

At least 55 records · Page 3Linked to original sources

arXiv2024

Predicting Artificial Neural Network Representations to Learn Recognition Model for Music Identification from Brain Recordings

Recent studies have demonstrated that the representations of artificial neural networks (ANNs) can exhibit notable similarities to cortical representations when subjected to identical auditory sensory inputs. In these studies, the ability to predict cortical representations is probed by regressing from ANN representations to cortical representations. Building upon this concept, our approach reverses the direction of prediction: we utilize ANN representations as a supervisory signal to train recognition models using noisy brain recordings obtained through non-invasive measurements. Specifically, we focus on constructing a recognition model for music identification, where electroencephalography (EEG) brain recordings collected during music listening serve as input. By training an EEG recognition model to predict ANN representations-representations associated with music identification-we observed a substantial improvement in classification accuracy. This study introduces a novel approach to developing recognition models for brain recordings in response to external auditory stimuli. It holds promise for advancing brain-computer interfaces (BCI), neural decoding techniques, and our understanding of music cognition. Furthermore, it provides new insights into the relationship between auditory brain activity and ANN representations.

q-bio.NC↗

arXiv2024

A Point-Neighborhood Learning Framework for Nasal Endoscope Image Segmentation

Lesion segmentation on nasal endoscopic images is challenging due to its complex lesion features. Fully-supervised deep learning methods achieve promising performance with pixel-level annotations but impose a significant annotation burden on experts. Although weakly supervised or semi-supervised methods can reduce the labelling burden, their performance is still limited. Some weakly semi-supervised methods employ a novel annotation strategy that labels weak single-point annotations for the entire training set while providing pixel-level annotations for a small subset of the data. However, the relevant weakly semi-supervised methods only mine the limited information of the point itself, while ignoring its label property and surrounding reliable information. This paper proposes a simple yet efficient weakly semi-supervised method called the Point-Neighborhood Learning (PNL) framework. PNL incorporates the surrounding area of the point, referred to as the point-neighborhood, into the learning process. In PNL, we propose a point-neighborhood supervision loss and a pseudo-label scoring mechanism to explicitly guide the model's training. Meanwhile, we proposed a more reliable data augmentation scheme. The proposed method significantly improves performance without increasing the parameters of the segmentation neural network. Extensive experiments on the NPC-LES dataset demonstrate that PNL outperforms existing methods by a significant margin. Additional validation on colonoscopic polyp segmentation datasets confirms the generalizability of the proposed PNL.

cs.CV↗

arXiv2024

MAIN-VC: Lightweight Speech Representation Disentanglement for One-shot Voice Conversion

One-shot voice conversion aims to change the timbre of any source speech to match that of the unseen target speaker with only one speech sample. Existing methods face difficulties in satisfactory speech representation disentanglement and suffer from sizable networks as some of them leverage numerous complex modules for disentanglement. In this paper, we propose a model named MAIN-VC to effectively disentangle via a concise neural network. The proposed model utilizes Siamese encoders to learn clean representations, further enhanced by the designed mutual information estimator. The Siamese structure and the newly designed convolution module contribute to the lightweight of our model while ensuring performance in diverse voice conversion tasks. The experimental results show that the proposed model achieves comparable subjective scores and exhibits improvements in objective metrics compared to existing methods in a one-shot voice conversion scenario.

cs.SD↗

arXiv2024

IDEAW: Robust Neural Audio Watermarking with Invertible Dual-Embedding

The audio watermarking technique embeds messages into audio and accurately extracts messages from the watermarked audio. Traditional methods develop algorithms based on expert experience to embed watermarks into the time-domain or transform-domain of signals. With the development of deep neural networks, deep learning-based neural audio watermarking has emerged. Compared to traditional algorithms, neural audio watermarking achieves better robustness by considering various attacks during training. However, current neural watermarking methods suffer from low capacity and unsatisfactory imperceptibility. Additionally, the issue of watermark locating, which is extremely important and even more pronounced in neural audio watermarking, has not been adequately studied. In this paper, we design a dual-embedding watermarking model for efficient locating. We also consider the impact of the attack layer on the invertible neural network in robustness training, improving the model to enhance both its reasonableness and stability. Experiments show that the proposed model, IDEAW, can withstand various attacks with higher capacity and more efficient locating ability compared to existing methods.

cs.MM↗

arXiv2024

TG-LLaVA: Text Guided LLaVA via Learnable Latent Embeddings

Currently, inspired by the success of vision-language models (VLMs), an increasing number of researchers are focusing on improving VLMs and have achieved promising results. However, most existing methods concentrate on optimizing the connector and enhancing the language model component, while neglecting improvements to the vision encoder itself. In contrast, we propose Text Guided LLaVA (TG-LLaVA) in this paper, which optimizes VLMs by guiding the vision encoder with text, offering a new and orthogonal optimization direction. Specifically, inspired by the purpose-driven logic inherent in human behavior, we use learnable latent embeddings as a bridge to analyze textual instruction and add the analysis results to the vision encoder as guidance, refining it. Subsequently, another set of latent embeddings extracts additional detailed text-guided information from high-resolution local patches as auxiliary information. Finally, with the guidance of text, the vision encoder can extract text-related features, similar to how humans focus on the most relevant parts of an image when considering a question. This results in generating better answers. Experiments on various datasets validate the effectiveness of the proposed method. Remarkably, without the need for additional training data, our propsoed method can bring more benefits to the baseline (LLaVA-1.5) compared with other concurrent methods. Furthermore, the proposed method consistently brings improvement in different settings.

cs.CV↗

arXiv2024

IVGF: The Fusion-Guided Infrared and Visible General Framework

Infrared and visible dual-modality tasks such as semantic segmentation and object detection can achieve robust performance even in extreme scenes by fusing complementary information. Most current methods design task-specific frameworks, which are limited in generalization across multiple tasks. In this paper, we propose a fusion-guided infrared and visible general framework, IVGF, which can be easily extended to many high-level vision tasks. Firstly, we adopt the SOTA infrared and visible foundation models to extract the general representations. Then, to enrich the semantics information of these general representations for high-level vision tasks, we design the feature enhancement module and token enhancement module for feature maps and tokens, respectively. Besides, the attention-guided fusion module is proposed for effectively fusing by exploring the complementary information of two modalities. Moreover, we also adopt the cutout&mix augmentation strategy to conduct the data augmentation, which further improves the ability of the model to mine the regional complementary between the two modalities. Extensive experiments show that the IVGF outperforms state-of-the-art dual-modality methods in the semantic segmentation and object detection tasks. The detailed ablation studies demonstrate the effectiveness of each module, and another experiment explores the anti-missing modality ability of the proposed method in the dual-modality semantic segmentation task.

cs.CV↗

arXiv2024

Envelopes created by pseudo-circle families in the Minkowski plane

In this paper, we address the topic of envelopes created by pseudo-circle families in the Minkowski plane, which exhibit some different properties when compared with the Euclidean case. We provide solutions to all four basic problems associated with these envelopes, namely the existence problem, representation problem, problem on the number of envelopes, and problem on the relationships of definitions.

math.DG↗

arXiv2024

Descent of splendid Rickard equivalences in ${\rm GL}_n(q)$

Let $n$ be a positive integer and $q$ a prime power. We prove that a refined version of Broué's abelian defect group conjecture holds for unipotent $\ell$-blocks of ${\rm GL}_n(q)$, where $\ell\nmid q$. We also give a sufficient condition on general $\ell$-blocks of ${\rm GL}_n(q)$ to satisfy the refined abelian defect group conjecture. We explain by an example that this sufficient condition does not hold in general.

math.RT↗

arXiv2024

Isospin-dependent in-medium nucleon-Delta elastic cross section

In heavy-ion collisions at intermediate energies, the production and propagation of $Δ$ particles are crucial to understanding the nuclear equation of state and inferring the properties of nuclear matter at high densities. Based on the self-consistent relativistic Boltzmann-Uehling-Uhlenbeck (RBUU) transport theory, by introducing the isovector $ρ$ meson exchange into the effective Lagrangian and adopting the density-dependent coupling constants, the detailed expressions for isospin-dependent in-medium $NΔ\rightarrow NΔ$ elastic cross sections $σ_{NΔ}^{*}$ have been calculated. The energy and density dependence of the isospin-related $σ_{NΔ}^{*}$ as well as the total contributions of $σ$, $ω$ and $ρ$ meson fields are analyzed. It is found that the total $σ_{NΔ}^{*}$ has a sensitive center-of-mass energy dependence at lower energies while exhibiting a slight increase as the center-of-mass energy increases. The isospin effect between different isospin-separated channels weakens as the energy and/or density increases. The isospin effect on the density- and energy-dependent in-medium $NΔ$ elastic cross sections is dominantly caused by the delicate balance of the isovector $ρ$ meson exchange.

nucl-th↗

arXiv2024

Medical Speech Symptoms Classification via Disentangled Representation

Intent is defined for understanding spoken language in existing works. Both textual features and acoustic features involved in medical speech contain intent, which is important for symptomatic diagnosis. In this paper, we propose a medical speech classification model named DRSC that automatically learns to disentangle intent and content representations from textual-acoustic data for classification. The intent representations of the text domain and the Mel-spectrogram domain are extracted via intent encoders, and then the reconstructed text feature and the Mel-spectrogram feature are obtained through two exchanges. After combining the intent from two domains into a joint representation, the integrated intent representation is fed into a decision layer for classification. Experimental results show that our model obtains an average accuracy rate of 95% in detecting 25 different medical symptoms.

cs.AI↗

arXiv2024

CONTUNER: Singing Voice Beautifying with Pitch and Expressiveness Condition

Singing voice beautifying is a novel task that has application value in people's daily life, aiming to correct the pitch of the singing voice and improve the expressiveness without changing the original timbre and content. Existing methods rely on paired data or only concentrate on the correction of pitch. However, professional songs and amateur songs from the same person are hard to obtain, and singing voice beautifying doesn't only contain pitch correction but other aspects like emotion and rhythm. Since we propose a fast and high-fidelity singing voice beautifying system called ConTuner, a diffusion model combined with the modified condition to generate the beautified Mel-spectrogram, where the modified condition is composed of optimized pitch and expressiveness. For pitch correction, we establish a mapping relationship from MIDI, spectrum envelope to pitch. To make amateur singing more expressive, we propose the expressiveness enhancer in the latent space to convert amateur vocal tone to professional. ConTuner achieves a satisfactory beautification effect on both Mandarin and English songs. Ablation study demonstrates that the expressiveness enhancer and generator-based accelerate method in ConTuner are effective.

cs.SD↗

arXiv2024

The homotopy decomposition of the suspension of a non-simply-connected $5$-manifold

In this paper we determine the homotopy types of the reduced suspension space of certain connected orientable closed smooth $5$-manifolds. As applications, we compute the reduced $K$-groups of $M$ and show that the suspension map between the third cohomotopy set $π^3(M)$ and the fourth cohomotopy set $π^4(ΣM)$ is a bijection.

math.AT↗

arXiv2024

The Homotopy Types of Suspended Simply-connected $6$-manifolds

Under the assumption that certain Adem cohomology operation acts trivially on $H^2(M;\mathbb{Z}/2)$, we determine the homotopy types of the triple suspension $Σ^3M$ of a simply-connected oriented closed topological(or smooth) $6$-manifold $M$, whose integral homology groups can have $2$-torsion.

math.AT↗

arXiv2024

DQR-TTS: Semi-supervised Text-to-speech Synthesis with Dynamic Quantized Representation

Most existing neural-based text-to-speech methods rely on extensive datasets and face challenges under low-resource condition. In this paper, we introduce a novel semi-supervised text-to-speech synthesis model that learns from both paired and unpaired data to address this challenge. The key component of the proposed model is a dynamic quantized representation module, which is integrated into a sequential autoencoder. When given paired data, the module incorporates a trainable codebook that learns quantized representations under the supervision of the paired data. However, due to the limited paired data in low-resource scenario, these paired data are difficult to cover all phonemes. Then unpaired data is fed to expand the dynamic codebook by adding quantized representation vectors that are sufficiently distant from the existing ones during training. Experiments show that with less than 120 minutes of paired data, the proposed method outperforms existing methods in both subjective and objective metrics.

cs.SD↗

arXiv2024

Homotopy classification of based maps between $\mathbf{A}_n^2$-complexes

Let $X,Y$ be $(n-1)$-connected finite pointed CW-complexes of dimension at most $n+2$, $n\geq 3$. In this paper we give elementary proofs of the abelian group structure of $[X,Y]$ of homotopy classes of based maps from $X$ to $Y$, which was due to Baues and Schmidt. Furthermore, we determine the explicit generators associated to $[X,Y]$. As an application, we compute certain (sub)groups of self-homotopy equivalences of certain Chang complexes.

math.AT↗

arXiv2024

Pedal and contrapedal curves of mixed-type Minkowski plane curves

Pedal and contrapedal curves are important study objects of plane curves. As for a mixed-type Minkowski plane curve, since the definitions of the pedal and contrapedal curves at lightlike points can not always be given, the investigation of them is difficult. We have done some research on the pedal curves of a mixed-type curve. In this paper, we discuss when the contrapedal curves of a mixed-type curve exist and give the definition of them when they exist. Then, we study when the contrapedal curves of the mixed-type curve have singular points. Meanwhile, we consider the types of the points on the contrapedal curves. Moreover, we investigate the relationship between the pedal and contrapedal curves of a mixed-type curve, as well as the relationship among them and the evolute of the base curve.

math-ph↗

arXiv2023

Visualization of Oxygen Vacancies and Self-doped Ligand Holes in La3Ni2O7-δ

The recent discovery of superconductivity in La3Ni2O7-δ under high pressure with a transition temperature around 80 K has sparked extensive experimental and theoretical efforts. Several key questions regarding the pairing mechanism remain to be answered, such as the most relevant atomic orbitals and the role of atomic deficiencies. Here, we develop a new energy-filtered multislice electron ptychography technique, assisted with electron energy loss spectroscopy, to address these critical issues. Oxygen vacancies are directly visualized and are found to primarily occupy the inner apical sites, which have been proposed to be crucial to superconductivity. We precisely determine the nanoscale stoichiometry and its correlation to the oxygen K edge spectra, which reveals a significant inhomogeneity in the oxygen content and electronic structure within the sample. The spectroscopic results also unveil that stoichiometric La3Ni2O7 is in the charge-transfer regime, with holes that are self-doped from Ni sites into O sites. The outer apical oxygen is found to be less relevant to the low-energy physics and can be safely disregarded in theoretical models. These observations will assist in further development and understanding of superconducting nickelate materials. Our imaging technique for quantifying atomic deficiencies can also be widely applied in materials science and condensed matter physics.

cond-mat.supr-con↗

arXiv2023

Categorical actions and derived equivalences for finite odd-dimensional orthogonal groups

In this paper we prove that Broué's abelian defect group conjecture is true for the finite odd-dimensional orthogonal groups $\SO_{2n+1}(q)$ at linear primes with $q$ odd. We first make use of the reduction theorem of Bonnafé-Dat-Rouquier to reduce the problem to isolated blocks. Then we construct a categorical action of a Kac-Moody algebra on the category of quadratic unipotent representations of the various groups $\SO_{2n+1}(q)$ in non-defining characteristic, by generalizing the corresponding work of Dudas-Varagnolo-Vasserot for unipotent representations. This is one of the main ingredients of our work which may be of independent interest. To obtain derived equivalences of blocks and their Brauer correspondents, we define and investigate isolated RoCK blocks. Finally, we establish the desired derived equivalence based on the work of Chuang-Rouquier that categorical actions provide derived equivalences between certain weight spaces.

math.RT↗