Neurodegenerative diseases are a group of disorders characterized by progressive degeneration or death of neurons. The complexity of clinical symptoms and irreversibility of disease progression significantly affects individual lives, leading to premature mortality. The prevalence of neurodegenerative diseases keeps increasing, yet the specific pathogenic mechanisms remain incompletely understood and effective treatment strategies are lacking. In recent years, convergent experimental evidence supports the "prion-like transmission" assumption that abnormal proteins induce misfolding of normal proteins, and these misfolded proteins propagate throughout the neural networks to cause neuronal death. To elucidate this dynamic process in vivo from a computational perspective, researchers have proposed three connectome-based biophysical models to simulate the spread of pathological proteins: the Network Diffusion Model, the Epidemic Spreading Model, and the agent-based Susceptible-Infectious-Removed model. These models have demonstrated promising predictive capabilities. This review focuses on the explanations of their fundamental principles and applications. Then, we compare the strengths and weaknesses of the models. Building upon this foundation, we introduce new directions for model optimization and propose a unified framework for the evaluation of connectome-based biophysical models. We expect that this review could lower the entry barrier for researchers in this field, accelerate model optimization, and thereby advance the clinical translation of connectome-based biophysical models.

Deep learning (DL) is increasingly viewed as a foundational methodology for advancing Artificial Intelligence (AI). However, its interpretability remains limited, and it often underperforms in certain fields due to its lack of human-like characteristics. Consequently, leveraging insights from Brain and Cognitive Science (BCS) to understand and advance DL has become a focal point for researchers in the DL community. However, BCS is a diverse discipline where existing studies often concentrate on cognitive theories within their respective domains. These theories are typically grounded in certain assumptions, complicating comparisons between different approaches. Therefore, this review is intended to provide a comprehensive landscape of more than 300 papers on the intersection of DL and BCS grounded in DL community. Unlike previous reviews that based on sub-disciplines of Cognitive Science, this article aims to establish a unified framework encompassing all aspects of DL inspired by BCS, offering insights into the symbiotic relationship between DL and BCS. Additionally, we present a forward-looking perspective on future research directions, with the intention of inspiring further advancements in AI research.

Query by Humming (QBH) is a practically meaningful task, while most existing methods struggle to scale to real-life applications due to the complex preprocessing for building the database and the limited search speed. In this paper, we propose the ByteHum system, a fast and efficient humming retrieval system which is capable of searching against large-scale databases built on raw song audios without the need for extensive preprocessing. ByteHum employs a convolutional neural network to extract features from raw audio, and utilizes a source-separated cover song identification dataset for weakly supervised training of the feature extractor. We explore the use of unsupervised domain adaptation techniques to enhance the performance of our weakly supervised model on the QBH task. Furthermore, to evaluate QBH systems’ performance on non-manually processed databases in the wild, we annotate original recordings for three existing QBH benchmark sets. Our experimental results demonstrate that ByteHum significantly outperforms existing QBH systems in terms of speed and accuracy under both classical and unconstrained settings.

When dealing with large-scale group decision-making problems, the central emphasis lies in the objective and rational acquisition of a collective opinion acceptable to the majority of decision makers . Manipulative and overconfident behaviours are two common behaviours that make the decision results deviate from the objective facts in the decision-making process. To manage manipulative and overconfident behaviours in decision-making, this paper investigated a novel consensus model based on social networks. A novel efficient clustering model is first proposed in the model, in which the subgroups' combined cohesion is considered. Furthermore, for the manipulative behaviour of decision makers, we proposed an improved method for the identification and management based on trust relationships. In the consensus reaching process, we proposed a new identification mechanism to promote consensus reaching effectively. In the feedback mechanism, the social network DeGroot model is employed to adjust the opinions of decision makers. Moreover, a management approach is proposed for the overconfident behaviour of decision makers in the social network DeGroot model. Lastly, the feasibility and applicability of the proposed model are verified by an illustrative example. Simulation experiments and comparative analysis demonstrate the effectiveness of the model in facilitating consensus reaching.

Deep learning based methods have become a paradigm for cover song identification (CSI) in recent years, where the ByteCover systems have achieved state-of-the-art results on all the mainstream datasets of CSI. However, with the burgeon of short videos, many real-world applications require matching short music excerpts to full-length music tracks in the database, which is still under-explored and waiting for an industrial-level solution. In this paper, we upgrade the previous ByteCover systems to ByteCover3 that utilizes local features to further improve the identification performance of short music queries. ByteCover3 is designed with a local alignment loss (LAL) module and a two-stage feature retrieval pipeline, allowing the system to perform CSI in a more precise and efficient way. We evaluated ByteCover3 on multiple datasets with different benchmark settings, where ByteCover3 beat all the compared methods including its previous versions.

As one of the fundamental tasks in music and speech signal processing, pitch tracking has been attracting attention for decades. While a human can focus on the voiced pitch even in highly noisy environments, most existing automatic pitch tracking systems show unsatisfactory performance encountering noise. To mimic human auditory, a data-driven model named GIO is proposed in this paper, in which timbre information is introduced to guide pitch tracking. The proposed model takes two inputs: a short audio segment to extract pitch from and a timbre embedding derived from the speaker's or singer's voice. In experiments, we use a music artist classification model to extract timbre embedding vectors. A dual-branch structure and a two-step training method are designed to enable the model to predict voice presence. The experimental results show that the proposed model gains a significant improvement in noise robustness and outperforms existing state-of-the-art methods with fewer parameters.

Many previous audio-visual voice-related works focus on speech, ignoring the singing voice in the growing number of musical video streams on the Internet. For processing diverse musical video data, voice activity detection is a necessary step. This paper attempts to detect the speech and singing voices of target performers in musical video streams using audio-visual information. To integrate information of audio and visual modalities, a multi-branch network is proposed to learn audio and image representations, and the representations are fused by attention based on semantic similarity to shape the acoustic representations through the probability of anchor vocalization. Experiments show the proposed audio-visual multi-branch network far outperforms the audio-only model in challenging acoustic environments, indicating the cross-modal information fusion based on semantic correlation is sensible and successful.

In the field of social information dissemination, current studies are mainly based on the network topology with an explicit user link. Implicit link in this study indicates users who often participate in the same topic due to certain interests or game relationships, but no explicit “following-followed” link exists among these users. In this article, the implicit link is regarded as one of the driving factors of users to participate in information dissemination, and then the network topology with the implicit link is established. First, we excavate implicit relationships among users to strengthen friends' influence in driving information dissemination and then establish a more accurate network topology. Second, we extract the individual and friend driving mechanisms based on the accurate network topology, analyze the causes of two information dynamics, and measure social influence based on a multiple linear regression model. Finally, considering the timeliness and uncertainty of information dissemination in an infectious disease model, we introduce the mean-field theory and obtain an information dissemination model based on social influence. The experimental results show that the implicit link plays an important role in driving user behavior, and the model can well explain the process of information transmission and explore the dynamic factors of information dissemination.

When you first encounter a person, a mental image of that person is formed. First impression, an interactive art, is proposed to let AI understand human personality at first glance. The mental image is demonstrated by Beijing opera facial makeups, which shows the character personality with a combination of realism and symbolism. We build Beijing opera facial makeup dataset and semantic dataset of facial features to establish relationships among real faces, personalities and facial makeups. First impression detects faces, recognizes personality from facial appearance and finds the matching Beijing opera facial makeup. Finally, the morphing process from real face to facial makeup is shown to let users enjoy the process of AI understanding personality.

A Q-rung orthopair fuzzy set (q-ROFS) originally proposed by Yager (2017) is a new generalization of orthopair fuzzy sets, which has a larger representation space of acceptable membership grades and gives decision makers more flexibility to express their real preferences. In this paper, for multiple attribute decision-making problems with q-rung orthopair fuzzy information, we propose a new method for dealing with heterogeneous relationship among attributes and unknown attribute weight information. First, we present two novel q-rung orthopair fuzzy extended Bonferroni mean (q-ROFEBM) operator and its weighted form (q-ROFEWEBM). A comparative example is provided to illustrate the advantages of the new operators, that is, they can effectively model the heterogeneous relationship among attributes. We prove that some existing known intuitionistic fuzzy aggregation operators and Pythagorean fuzzy aggregation operators are special cases of the proposed q-ROFEBM and q-ROFEWEBM operators. Meanwhile, several desirable properties are also investigated. Then, a new knowledge-based entropy measure for q-ROFSs is also proposed to obtain the attribute weights. Based on the proposed q-ROFWEBM and the new entropy measure, a new method is developed to solve multiple attribute decision making problems with q-ROFSs. Finally, an illustrative example is given to demonstrate the application process of the proposed method, and a comparison analysis with other existing representative methods is also conducted to show its validity and superiority.