Search

Lifelong learning in artificial intelligence (AI) aims to mimic the biological brain's ability to continuously learn and retain knowledge, yet it faces challenges such as catastrophic forgetting. Recent neuroscience research suggests that neural activity in biological systems undergoes representational drift, where neural responses evolve over time, even with consistent inputs and tasks. We hypothesize that representational drift can alleviate catastrophic forgetting in AI during new task acquisition. To test this, we introduce DriftNet, a network designed to constantly explore various local minima in the loss landscape while dynamically retrieving relevant tasks. This approach ensures efficient integration of new information and preserves existing knowledge. Experimental studies in image classification and natural language processing demonstrate that DriftNet outperforms existing models in lifelong learning. Importantly, DriftNet is scalable in handling a sequence of tasks such as sentiment analysis and question answering using large language models (LLMs) with billions of parameters on a single Nvidia A100 GPU. DriftNet efficiently updates LLMs using only new data, avoiding the need for full dataset retraining. Tested on GPT-2 and RoBERTa, DriftNet is a robust, cost-effective solution for lifelong learning in LLMs. This study not only advances AI systems to emulate biological learning, but also provides insights into the adaptive mechanisms of biological neural systems, deepening our understanding of lifelong learning in nature.

Neutron star (NS) has many extreme physical conditions, and one may obtain some important informations about NS via accreting neutron star binary (ANSB) systems. The upcoming Chinese Space Station Telescope (CSST) provides an opportunity to search for a large sample of ANSB candidates. Our goal is to check the completeness of the potential ANSB samples from CSST data. In this paper, we generate some ANSBs and normal binaries under CSST photometric system by binary evolution and binary population synthesis method and use a machine learning method to train a classification model. Although the Precision ($94.56~ \%$) of our machine learning model is as high as before study, the Recall is only about $63.29~ \%$. The Precision/Recall is mainly determined by the mass transfer rate between the NSs and their companions. In addition, we also find that the completeness of ANSB samples from CSST photometric data by the machine learning method also depends on the companion mass and the age of the system. ANSB candidates with low initial mass companion star ($0.1~ {\rm M}_\odot$ to $1~ {\rm M}_\odot$) have a relatively high Precision ($94.94~ \%$) and high Recall ($86.32~ \%$), whereas ANSB candidates with higher initial mass companion star ($1.1~ {\rm M}_\odot$ to $3~ {\rm M}_\odot$) have similar Precision ($93.88~ \%$) and quite low Recall ($42.67~ \%$). Our results indicate that although the machine learning method may obtain a relative pure sample of ANSBs, a completeness correction is necessary for one to obtain a complete sample., Comment: 12 pages, 5 figures

In contemporary society, the issue of psychological health has become increasingly prominent, characterized by the diversification, complexity, and universality of mental disorders. Cognitive Behavioral Therapy (CBT), currently the most influential and clinically effective psychological treatment method with no side effects, has limited coverage and poor quality in most countries. In recent years, researches on the recognition and intervention of emotional disorders using large language models (LLMs) have been validated, providing new possibilities for psychological assistance therapy. However, are LLMs truly possible to conduct cognitive behavioral therapy? Many concerns have been raised by mental health experts regarding the use of LLMs for therapy. Seeking to answer this question, we collected real CBT corpus from online video websites, designed and conducted a targeted automatic evaluation framework involving the evaluation of emotion tendency of generated text, structured dialogue pattern and proactive inquiry ability. For emotion tendency, we calculate the emotion tendency score of the CBT dialogue text generated by each model. For structured dialogue pattern, we use a diverse range of automatic evaluation metrics to compare speaking style, the ability to maintain consistency of topic and the use of technology in CBT between different models . As for inquiring to guide the patient, we utilize PQA (Proactive Questioning Ability) metric. We also evaluated the CBT ability of the LLM after integrating a CBT knowledge base to explore the help of introducing additional knowledge to enhance the model's CBT counseling ability. Four LLM variants with excellent performance on natural language processing are evaluated, and the experimental result shows the great potential of LLMs in psychological counseling realm, especially after combining with other technological means.

In real-world applications of human pose estimation, low-resolution input images are frequently encountered when the performance of the image acquisition equipment is limited or the shooting distance is too far. However, existing state-of-the-art models for human pose estimation perform poorly on low-resolution images. One key reason is the presence of downsampling layers in these models, e.g., strided convolutions and pooling layers. It further reduces the already insufficient image information. Another key reason is that the body skeleton and human kinematic information are not fully utilized. In this work, we propose a Multi-Granular Information-Lossless (MGIL) model to replace the downsampling layers to address the above issues. Specifically, MGIL employs a Fine-grained Lossless Information Extraction (FLIE) module, which can prevent the loss of local information. Furthermore, we design a Coarse-grained Information Interaction (CII) module to adequately leverage human body structural information. To efficiently fuse cross-granular information and thoroughly exploit the relationships among keypoints, we further introduce a Multi-Granular Adaptive Fusion (MGAF) mechanism. The mechanism assigns weights to features of different granularities based on the content of the image. The model is effective, flexible, and universal. We show its potential in various vision tasks with comprehensive experiments. It outperforms the SOTA methods by 7.7 mAP on COCO and performs well with different input resolutions, different backbones, and different vision tasks. The code is provided in supplementary material., Comment: 8 pages, 5 figures, conference

In practical applications of human pose estimation, low-resolution inputs frequently occur, and existing state-of-the-art models perform poorly with low-resolution images. This work focuses on boosting the performance of low-resolution models by distilling knowledge from a high-resolution model. However, we face the challenge of feature size mismatch and class number mismatch when applying knowledge distillation to networks with different input resolutions. To address this issue, we propose a novel cross-domain knowledge distillation (CDKD) framework. In this framework, we construct a scale-adaptive projector ensemble (SAPE) module to spatially align feature maps between models of varying input resolutions. It adopts a projector ensemble to map low-resolution features into multiple common spaces and adaptively merges them based on multi-scale information to match high-resolution features. Additionally, we construct a cross-class alignment (CCA) module to solve the problem of the mismatch of class numbers. By combining an easy-to-hard training (ETHT) strategy, the CCA module further enhances the distillation performance. The effectiveness and efficiency of our approach are demonstrated by extensive experiments on two common benchmark datasets: MPII and COCO. The code is made available in supplementary material., Comment: 11 pages, 5 figures

Video Anomaly Detection (VAD), aiming to identify abnormalities within a specific context and timeframe, is crucial for intelligent Video Surveillance Systems. While recent deep learning-based VAD models have shown promising results by generating high-resolution frames, they often lack competence in preserving detailed spatial and temporal coherence in video frames. To tackle this issue, we propose a self-supervised learning approach for VAD through an inter-patch relationship prediction task. Specifically, we introduce a two-branch vision transformer network designed to capture deep visual features of video frames, addressing spatial and temporal dimensions responsible for modeling appearance and motion patterns, respectively. The inter-patch relationship in each dimension is decoupled into inter-patch similarity and the order information of each patch. To mitigate memory consumption, we convert the order information prediction task into a multi-label learning problem, and the inter-patch similarity prediction task into a distance matrix regression problem. Comprehensive experiments demonstrate the effectiveness of our method, surpassing pixel-generation-based methods by a significant margin across three public benchmarks. Additionally, our approach outperforms other self-supervised learning-based methods.

Existing research based on deep learning has extensively explored the problem of daytime image dehazing. However, few studies have considered the characteristics of nighttime hazy scenes. There are two distinctions between nighttime and daytime haze. First, there may be multiple active colored light sources with lower illumination intensity in nighttime scenes, which may cause haze, glow and noise with localized, coupled and frequency inconsistent characteristics. Second, due to the domain discrepancy between simulated and real-world data, unrealistic brightness may occur when applying a dehazing model trained on simulated data to real-world data. To address the above two issues, we propose a semi-supervised model for real-world nighttime dehazing. First, the spatial attention and frequency spectrum filtering are implemented as a spatial-frequency domain information interaction module to handle the first issue. Second, a pseudo-label-based retraining strategy and a local window-based brightness loss for semi-supervised training process is designed to suppress haze and glow while achieving realistic brightness. Experiments on public benchmarks validate the effectiveness of the proposed method and its superiority over state-of-the-art methods. The source code and Supplementary Materials are placed in the https://github.com/Xiaofeng-life/SFSNiD., Comment: This paper is accepted by CVPR2024

Language significantly influences the formation and evolution of Human emergent behavior, which is crucial in understanding collective intelligence within human societies. Considering that the study of how language affects human behavior needs to put it into the dynamic scenarios in which it is used, we introduce AgentGroupChat in this paper, a simulation that delves into the complex role of language in shaping collective behavior through interactive debate scenarios. Central to this simulation are characters engaging in dynamic conversation interactions. To enable simulation, we introduce the Verbal Strategist Agent, utilizing large language models to enhance interaction strategies by incorporating elements of persona and action. We set four narrative scenarios based on AgentGroupChat to demonstrate the simulation's capacity to mimic complex language use in group dynamics. Evaluations focus on aligning agent behaviors with human expectations and the emergence of collective behaviors within the simulation. Results reveal that emergent behaviors materialize from a confluence of factors: a conducive environment for extensive information exchange, characters with diverse traits, high linguistic comprehension, and strategic adaptability. During discussions on ``the impact of AI on humanity'' in AgentGroupChat simulation, philosophers commonly agreed that ``AI could enhance societal welfare with judicious limitations'' and even come to a conclusion that ``the essence of true intelligence encompasses understanding the necessity to constrain self abilities''. Additionally, in the competitive domain of casting for primary roles in films in AgentGroupChat, certain actors were ready to reduce their remuneration or accept lesser roles, motivated by their deep-seated desire to contribute to the project.

This article revisits the problem of global well-posedness for the generalized parabolic Anderson model on $\mathbb{R}^+\times \mathbb{T}^2$ within the framework of paracontrolled calculus \cite{GIP15}. The model is given by the equation: \begin{equation*} (\partial_t-\Delta) u=F(u)\eta \end{equation*} where $\eta\in C^{-1-\kappa}$ with $1/6>\kappa>0$, and $F\in C_b^2(\mathbb{R})$. Assume that $\eta\in C^{-1-\kappa}$ and can be lifted to enhanced noise, we derive new a priori bounds. The key idea follows from the recent work \cite{CFW24} by A.Chandra, G.L. Feltes and H.Weber to represent the leading error term as a transport type term, and our techniques encompass the paracontrolled calculus, the maximum principle, and the localization approach (i.e. high-low frequency argument)., Comment: 19 pages

We investigate the Langevin dynamics of various lattice formulations of the Yang-Mills-Higgs model, where the Higgs component takes values in $\mathbb{R}^N$, $\mathbb{S}^{N-1}$ or a Lie group. We prove the exponential ergodicity of the dynamics on the whole lattice via functional inequalities. As an application, we establish that correlations for a broad range of observables decay exponentially. Specifically, the infinite volume measure exhibits a strictly positive mass gap under strong coupling conditions. Moreover, appropriately rescaled observables exhibit factorized correlations in the large $N$ limit when the state space is compact. Our approach involves disintegration and a nuanced analysis of correlations to effectively control the unbounded Higgs component., Comment: 54 pages

Contrastive learning has recently emerged as a promising approach for learning data representations that discover and disentangle the explanatory factors of the data. Previous analyses of such approaches have largely focused on individual contrastive losses, such as noise-contrastive estimation (NCE) and InfoNCE, and rely on specific assumptions about the data generating process. This paper extends the theoretical guarantees for disentanglement to a broader family of contrastive methods, while also relaxing the assumptions about the data distribution. Specifically, we prove identifiability of the true latents for four contrastive losses studied in this paper, without imposing common independence assumptions. The theoretical findings are validated on several benchmark datasets. Finally, practical limitations of these methods are also investigated.

Multi-stage architectures have exhibited efficacy in image dehazing, which usually decomposes a challenging task into multiple more tractable sub-tasks and progressively estimates latent hazy-free images. Despite the remarkable progress, existing methods still suffer from the following shortcomings: (1) limited exploration of frequency domain information; (2) insufficient information interaction; (3) severe feature redundancy. To remedy these issues, we propose a novel Mutual Information-driven Triple interaction Network (MITNet) based on spatial-frequency dual domain information and two-stage architecture. To be specific, the first stage, named amplitude-guided haze removal, aims to recover the amplitude spectrum of the hazy images for haze removal. And the second stage, named phase-guided structure refined, devotes to learning the transformation and refinement of the phase spectrum. To facilitate the information exchange between two stages, an Adaptive Triple Interaction Module (ATIM) is developed to simultaneously aggregate cross-domain, cross-scale, and cross-stage features, where the fused features are further used to generate content-adaptive dynamic filters so that applying them to enhance global context representation. In addition, we impose the mutual information minimization constraint on paired scale encoder and decoder features from both stages. Such an operation can effectively reduce information redundancy and enhance cross-stage feature complementarity. Extensive experiments on multiple public datasets exhibit that our MITNet performs superior performance with lower model complexity.The code and models are available at https://github.com/it-hao/MITNet., Comment: Accepted in ACM MM 2023

In this paper, we continue the study of large $N$ problems for the Wick renormalized linear sigma model, i.e. $N$-component $\Phi^4$ model, in two spatial dimensions, using stochastic quantization methods and Dyson--Schwinger equations. We identify the large $N$ limiting law of a collection of Wick renormalized $O(N)$ invariant observables. In particular, under a suitable scaling, the quadratic observables converge in the large $N$ limit to a mean-zero (singular) Gaussian field denoted by $\mathcal{Q}$ with an explicit covariance; and the observables which are renormalized powers of order $2n$ converge in the large $N$ limit to suitably renormalized $n$-th powers of $\mathcal{Q}$. The quartic interaction term of the model has no effect on the large $N$ limit of the field, but has nontrivial contributions to the limiting law of the observables, and the renormalization of the $n$-th powers of $\mathcal{Q}$ in the limit has an interesting finite shift from the standard one. Furthermore, we derive the $1/N$ asymtotic expansion for the $k$-point functions of the quadratic observables by employing graph representations and analyzing the order of each graph from Dyson--Schwinger equations. Finally, turning to the stationary solutions to the stochastic quantization equations, with the Ornstein--Uhlenbeck process being the large $N$ limiting dynamic, we derive here its next order correction in stationarity, as described by an SPDE with the right-hand side having explicit fixed-time marginal law which involves the above field $\mathcal{Q}$., Comment: 54 pages