arXiv subjects

You Wu

Publications and source records attributed to You Wu.

At least 55 records · Page 3Linked to original sources

arXiv2025

Reasoning RAG via System 1 or System 2: A Survey on Reasoning Agentic Retrieval-Augmented Generation for Industry Challenges

Retrieval-Augmented Generation (RAG) has emerged as a powerful framework to overcome the knowledge limitations of Large Language Models (LLMs) by integrating external retrieval with language generation. While early RAG systems based on static pipelines have shown effectiveness in well-structured tasks, they struggle in real-world scenarios requiring complex reasoning, dynamic retrieval, and multi-modal integration. To address these challenges, the field has shifted toward Reasoning Agentic RAG, a paradigm that embeds decision-making and adaptive tool use directly into the retrieval process. In this paper, we present a comprehensive review of Reasoning Agentic RAG methods, categorizing them into two primary systems: predefined reasoning, which follows fixed modular pipelines to boost reasoning, and agentic reasoning, where the model autonomously orchestrates tool interaction during inference. We analyze representative techniques under both paradigms, covering architectural design, reasoning strategies, and tool coordination. Finally, we discuss key research challenges and propose future directions to advance the flexibility, robustness, and applicability of reasoning agentic RAG systems. Our collection of the relevant research has been organized into a https://github.com/ByebyeMonica/Reasoning-Agentic-RAG.

cs.AI↗

arXiv2025

Pre-Big-Bang Cosmology Cannot Explain NANOGrav 15-year Signal

We investigate whether the Pre-Big Bang (PBB) scenario from string cosmology can explain the stochastic gravitational wave background signal reported in the NANOGrav 15-year dataset. Using Bayesian analysis techniques, we constrain the key parameters of the PBB model by comparing its theoretical predictions with the observed data. Our analysis yields $β= -0.12^{+0.06}_{-0.21}$ ($90\%$ credible interval) for the dilaton-dynamics parameter, which lies outside the theoretically allowed range $0 \leq β< 3$ with more than $5σ$ confidence. Additionally, model comparison strongly favors a simple power-law spectrum over the PBB scenario, with a Bayes factor of approximately $468$. These results demonstrate that the PBB scenario, in its current formulation, cannot adequately explain the NANOGrav observations, highlighting the need for either significant modifications to the model or alternative explanations for the observed signal.

astro-ph.CO↗

arXiv2025

In-Context Brush: Zero-shot Customized Subject Insertion with Context-Aware Latent Space Manipulation

Recent advances in diffusion models have enhanced multimodal-guided visual generation, enabling customized subject insertion that seamlessly "brushes" user-specified objects into a given image guided by textual prompts. However, existing methods often struggle to insert customized subjects with high fidelity and align results with the user's intent through textual prompts. In this work, we propose "In-Context Brush", a zero-shot framework for customized subject insertion by reformulating the task within the paradigm of in-context learning. Without loss of generality, we formulate the object image and the textual prompts as cross-modal demonstrations, and the target image with the masked region as the query. The goal is to inpaint the target image with the subject aligning textual prompts without model tuning. Building upon a pretrained MMDiT-based inpainting network, we perform test-time enhancement via dual-level latent space manipulation: intra-head "latent feature shifting" within each attention head that dynamically shifts attention outputs to reflect the desired subject semantics and inter-head "attention reweighting" across different heads that amplifies prompt controllability through differential attention prioritization. Extensive experiments and applications demonstrate that our approach achieves superior identity preservation, text alignment, and image quality compared to existing state-of-the-art methods, without requiring dedicated training or additional data collection.

cs.CV↗

arXiv2025

MambaNUT: Nighttime UAV Tracking via Mamba-based Adaptive Curriculum Learning

Harnessing low-light enhancement and domain adaptation, nighttime UAV tracking has made substantial strides. However, over-reliance on image enhancement, limited high-quality nighttime data, and a lack of integration between daytime and nighttime trackers hinder the development of an end-to-end trainable framework. Additionally, current ViT-based trackers demand heavy computational resources due to their reliance on the self-attention mechanism. In this paper, we propose a novel pure Mamba-based tracking framework (MambaNUT) that employs a state space model with linear complexity as its backbone, incorporating a single-stream architecture that integrates feature learning and template-search coupling within Vision Mamba. We introduce an adaptive curriculum learning (ACL) approach that dynamically adjusts sampling strategies and loss weights, thereby improving the model's ability of generalization. Our ACL is composed of two levels of curriculum schedulers: (1) sampling scheduler that transforms the data distribution from imbalanced to balanced, as well as from easier (daytime) to harder (nighttime) samples; (2) loss scheduler that dynamically assigns weights based on the size of the training set and IoU of individual instances. Exhaustive experiments on multiple nighttime UAV tracking benchmarks demonstrate that the proposed MambaNUT achieves state-of-the-art performance while requiring lower computational costs. The code will be available at https://github.com/wuyou3474/MambaNUT.

cs.CV↗

arXiv2025

Wan: Open and Advanced Large-Scale Video Generative Models

This report presents Wan, a comprehensive and open suite of video foundation models designed to push the boundaries of video generation. Built upon the mainstream diffusion transformer paradigm, Wan achieves significant advancements in generative capabilities through a series of innovations, including our novel VAE, scalable pre-training strategies, large-scale data curation, and automated evaluation metrics. These contributions collectively enhance the model's performance and versatility. Specifically, Wan is characterized by four key features: Leading Performance: The 14B model of Wan, trained on a vast dataset comprising billions of images and videos, demonstrates the scaling laws of video generation with respect to both data and model size. It consistently outperforms the existing open-source models as well as state-of-the-art commercial solutions across multiple internal and external benchmarks, demonstrating a clear and significant performance superiority. Comprehensiveness: Wan offers two capable models, i.e., 1.3B and 14B parameters, for efficiency and effectiveness respectively. It also covers multiple downstream applications, including image-to-video, instruction-guided video editing, and personal video generation, encompassing up to eight tasks. Consumer-Grade Efficiency: The 1.3B model demonstrates exceptional resource efficiency, requiring only 8.19 GB VRAM, making it compatible with a wide range of consumer-grade GPUs. Openness: We open-source the entire series of Wan, including source code and all models, with the goal of fostering the growth of the video generation community. This openness seeks to significantly expand the creative possibilities of video production in the industry and provide academia with high-quality video foundation models. All the code and models are available at https://github.com/Wan-Video/Wan2.1.

cs.CV↗

arXiv2025

Implications for the non-Gaussianity of primordial gravitational waves from pulsar timing arrays

The detection of a stochastic signal by recent pulsar timing array (PTA) collaborations, including NANOGrav, PPTA, EPTA+InPTA, CPTA and MPTA, has opened a new window to explore gravitational waves (GWs) at nanohertz frequencies. Motivated by the possibility that such a signal could arise from primordial gravitational waves (PGWs), we investigate the implications of tensor non-Gaussianity for the PGW power spectrum. Utilizing PTA data sets, we provide constraints on local-type tensor non-Gaussianity parameter ${F}_{\mathrm{NL}}$. We find $|{F}_{\mathrm{NL}}|\lesssim 7.97$ for a log-normal PGW power spectrum. Our analysis reveals that even moderate tensor non-Gaussianity can lead to significant deviations from standard predictions, thereby offering a novel means to test inflationary scenarios and probe the underlying dynamics of the early Universe. Future multi-band GW observatories, such as LISA, Taiji, and TianQin, will be instrumental in complementing these efforts and further refining our understanding of tensor non-Gaussianity.

astro-ph.CO↗

arXiv2025

Learning Occlusion-Robust Vision Transformers for Real-Time UAV Tracking

Single-stream architectures using Vision Transformer (ViT) backbones show great potential for real-time UAV tracking recently. However, frequent occlusions from obstacles like buildings and trees expose a major drawback: these models often lack strategies to handle occlusions effectively. New methods are needed to enhance the occlusion resilience of single-stream ViT models in aerial tracking. In this work, we propose to learn Occlusion-Robust Representations (ORR) based on ViTs for UAV tracking by enforcing an invariance of the feature representation of a target with respect to random masking operations modeled by a spatial Cox process. Hopefully, this random masking approximately simulates target occlusions, thereby enabling us to learn ViTs that are robust to target occlusion for UAV tracking. This framework is termed ORTrack. Additionally, to facilitate real-time applications, we propose an Adaptive Feature-Based Knowledge Distillation (AFKD) method to create a more compact tracker, which adaptively mimics the behavior of the teacher model ORTrack according to the task's difficulty. This student model, dubbed ORTrack-D, retains much of ORTrack's performance while offering higher efficiency. Extensive experiments on multiple benchmarks validate the effectiveness of our method, demonstrating its state-of-the-art performance. Codes is available at https://github.com/wuyou3474/ORTrack.

cs.CV↗

arXiv2025

A Systematic Study of Cross-Layer KV Sharing for Efficient LLM Inference

Recently, sharing key-value (KV) cache across layers has been found effective in efficient inference of large language models (LLMs). To systematically investigate different techniques of cross-layer KV sharing, we propose a unified framework that covers several recent methods and their novel variants. We conduct comprehensive experiments on all the configurations of the framework, evaluating their generation throughput and performance in language modeling and downstream tasks. We find that when reducing the size of the KV cache by 2$\times$, most configurations can achieve higher throughput than standard transformers while maintaining competitive performance. When further reducing the size of the KV cache, however, pairing queries of all layers with KVs of upper layers performs better, at the expense of additional training cost and prefilling latency. We hope that this work will help users make more informed choices of cross-layer KV sharing approaches and facilitate future research on efficient LLM inference.

cs.CL↗

arXiv2025

Search for a Gravitational-Wave Background from Sound Speed Resonance from Advanced LIGO and Advanced Virgo's First Three Observing Runs

We search for a stochastic gravitational-wave background (SGWB) originating from scalar-induced gravitational waves (SIGWs) with the sound speed resonance (SSR) effect using data from Advanced LIGO and Advanced Virgo's first three observing runs. The SSR mechanism, characterized by an oscillating sound speed squared term, can induce a nonperturbative parametric amplification of specific perturbation modes during inflation, leading to enhanced primordial curvature perturbations and a significant SIGW signal. We perform a Bayesian analysis to constrain the model parameters describing the SGWB spectrum from the SSR effect. Our results show no statistically significant evidence for the presence of such a signal in the current data. Consequently, we place an upper limit of $|τ_0| \lesssim 5.9 \times 10^3\,\mathrm{s}$ at $95\%$ confidence level on the start time of the oscillation in the SSR model. These results demonstrate the capability of current gravitational wave detectors to probe inflation models through the SSR mechanism and paves the way for future searches with improved sensitivity.

astro-ph.CO↗

arXiv2024

Integrative Analysis of Financial Market Sentiment Using CNN and GRU for Risk Prediction and Alert Systems

This document presents an in-depth examination of stock market sentiment through the integration of Convolutional Neural Networks (CNN) and Gated Recurrent Units (GRU), enabling precise risk alerts. The robust feature extraction capability of CNN is utilized to preprocess and analyze extensive network text data, identifying local features and patterns. The extracted feature sequences are then input into the GRU model to understand the progression of emotional states over time and their potential impact on future market sentiment and risk. This approach addresses the order dependence and long-term dependencies inherent in time series data, resulting in a detailed analysis of stock market sentiment and effective early warnings of future risks.

cs.LG↗

arXiv2024

Advanced Risk Prediction and Stability Assessment of Banks Using Time Series Transformer Models

This paper aims to study the prediction of the bank stability index based on the Time Series Transformer model. The bank stability index is an important indicator to measure the health status and risk resistance of financial institutions. Traditional prediction methods are difficult to adapt to complex market changes because they rely on single-dimensional macroeconomic data. This paper proposes a prediction framework based on the Time Series Transformer, which uses the self-attention mechanism of the model to capture the complex temporal dependencies and nonlinear relationships in financial data. Through experiments, we compare the model with LSTM, GRU, CNN, TCN and RNN-Transformer models. The experimental results show that the Time Series Transformer model outperforms other models in both mean square error (MSE) and mean absolute error (MAE) evaluation indicators, showing strong prediction ability. This shows that the Time Series Transformer model can better handle multidimensional time series data in bank stability prediction, providing new technical approaches and solutions for financial risk management.

q-fin.RM↗

arXiv2024

Integrating Planning into Single-Turn Long-Form Text Generation

Generating high-quality, in-depth textual documents, such as academic papers, news articles, Wikipedia entries, and books, remains a significant challenge for Large Language Models (LLMs). In this paper, we propose to use planning to generate long form content. To achieve our goal, we generate intermediate steps via an auxiliary task that teaches the LLM to plan, reason and structure before generating the final text. Our main novelty lies in a single auxiliary task that does not require multiple rounds of prompting or planning. To overcome the scarcity of training data for these intermediate steps, we leverage LLMs to generate synthetic intermediate writing data such as outlines, key information and summaries from existing full articles. Our experiments demonstrate on two datasets from different domains, namely the scientific news dataset SciNews and Wikipedia datasets in KILT-Wiki and FreshWiki, that LLMs fine-tuned with the auxiliary task generate higher quality documents. We observed +2.5% improvement in ROUGE-Lsum, and a strong 3.60 overall win/loss ratio via human SxS evaluation, with clear wins in organization, relevance, and verifiability.

cs.CL↗

arXiv2024

Online Multi-Label Classification under Noisy and Changing Label Distribution

Multi-label data stream usually contains noisy labels in the real-world applications, namely occuring in both relevant and irrelevant labels. However, existing online multi-label classification methods are mostly limited in terms of label quality and fail to deal with the case of noisy labels. On the other hand, the ground-truth label distribution may vary with the time changing, which is hidden in the observed noisy label distribution and difficult to track, posing a major challenge for concept drift adaptation. Motivated by this, we propose an online multi-label classification algorithm under Noisy and Changing Label Distribution (NCLD). The convex objective is designed to simultaneously model the label scoring and the label ranking for high accuracy, whose robustness to NCLD benefits from three novel works: 1) The local feature graph is used to reconstruct the label scores jointly with the observed labels, and an unbiased ranking loss is derived and applied to learn reliable ranking information. 2) By detecting the difference between two adjacent chunks with the unbiased label cardinality, we identify the change in the ground-truth label distribution and reset the ranking or all information learned from the past to match the new distribution. 3) Efficient and accurate updating is achieved based on the updating rule derived from the closed-form optimal model solution. Finally, empirical experimental results validate the effectiveness of our method in classifying instances under NCLD.

cs.LG↗

arXiv2024

Constraining string cosmology with the gravitational-wave background using the NANOGrav 15-year data set

The North American Nanohertz Observatory for Gravitational Waves (NANOGrav) collaboration has recently reported strong evidence for a signal at nanohertz, potentially the first detection of the stochastic gravitational-wave background (SGWB). We investigate whether the NANOGrav signal is consistent with the SGWB predicted by string cosmology models. By performing Bayesian parameter estimation on the NANOGrav 15-year data set, we constrain the key parameters of a string cosmology model: the frequency $f_s$ and the fractional energy density $Ω_\mathrm{gw}^{s}$ of gravitational waves at the end of the dilaton-driven stage, and the Hubble parameter $H_r$ at the end of the string phase. Our analysis yields constraints of $f_s = 1.2^{+0.6}_{-0.6}\times 10^{-8} \mathrm{Hz}$ and $Ω_\mathrm{gw}^{s} = 2.9^{+5.4}_{-2.3}\times 10^{-8}$, consistent with theoretical predictions from string cosmology. However, the current NANOGrav data is not sensitive to the $H_r$ parameter. We also compare the string cosmology model to a simple power-law model using Bayesian model selection, finding a Bayes factor of $2.2$ in favor of the string cosmology model. Our results demonstrate the potential of pulsar timing arrays to constrain cosmological models and study the early Universe.

gr-qc↗

arXiv2024

The Exploration-Exploitation Dilemma Revisited: An Entropy Perspective

The imbalance of exploration and exploitation has long been a significant challenge in reinforcement learning. In policy optimization, excessive reliance on exploration reduces learning efficiency, while over-dependence on exploitation might trap agents in local optima. This paper revisits the exploration-exploitation dilemma from the perspective of entropy by revealing the relationship between entropy and the dynamic adaptive process of exploration and exploitation. Based on this theoretical insight, we establish an end-to-end adaptive framework called AdaZero, which automatically determines whether to explore or to exploit as well as their balance of strength. Experiments show that AdaZero significantly outperforms baseline models across various Atari and MuJoCo environments with only a single setting. Especially in the challenging environment of Montezuma, AdaZero boosts the final returns by up to fifteen times. Moreover, we conduct a series of visualization analyses to reveal the dynamics of our self-adaptive mechanism, demonstrating how entropy reflects and changes with respect to the agent's performance and adaptive process.

cs.LG↗

arXiv2024

Advanced Financial Fraud Detection Using GNN-CL Model

The innovative GNN-CL model proposed in this paper marks a breakthrough in the field of financial fraud detection by synergistically combining the advantages of graph neural networks (gnn), convolutional neural networks (cnn) and long short-term memory (LSTM) networks. This convergence enables multifaceted analysis of complex transaction patterns, improving detection accuracy and resilience against complex fraudulent activities. A key novelty of this paper is the use of multilayer perceptrons (MLPS) to estimate node similarity, effectively filtering out neighborhood noise that can lead to false positives. This intelligent purification mechanism ensures that only the most relevant information is considered, thereby improving the model's understanding of the network structure. Feature weakening often plagues graph-based models due to the dilution of key signals. In order to further address the challenge of feature weakening, GNN-CL adopts reinforcement learning strategies. By dynamically adjusting the weights assigned to central nodes, it reinforces the importance of these influential entities to retain important clues of fraud even in less informative data. Experimental evaluations on Yelp datasets show that the results highlight the superior performance of GNN-CL compared to existing methods.

cs.LG↗

arXiv2024

Tracking Transforming Objects: A Benchmark

Tracking transforming objects holds significant importance in various fields due to the dynamic nature of many real-world scenarios. By enabling systems accurately represent transforming objects over time, tracking transforming objects facilitates advancements in areas such as autonomous systems, human-computer interaction, and security applications. Moreover, understanding the behavior of transforming objects provides valuable insights into complex interactions or processes, contributing to the development of intelligent systems capable of robust and adaptive perception in dynamic environments. However, current research in the field mainly focuses on tracking generic objects. In this study, we bridge this gap by collecting a novel dedicated Dataset for Tracking Transforming Objects, called DTTO, which contains 100 sequences, amounting to approximately 9.3K frames. We provide carefully hand-annotated bounding boxes for each frame within these sequences, making DTTO the pioneering benchmark dedicated to tracking transforming objects. We thoroughly evaluate 20 state-of-the-art trackers on the benchmark, aiming to comprehend the performance of existing methods and provide a comparison for future research on DTTO. With the release of DTTO, our goal is to facilitate further research and applications related to tracking transforming objects.

cs.CV↗

arXiv2024

AI-driven multi-omics integration for multi-scale predictive modeling of causal genotype-environment-phenotype relationships

Despite the wealth of single-cell multi-omics data, it remains challenging to predict the consequences of novel genetic and chemical perturbations in the human body. It requires knowledge of molecular interactions at all biological levels, encompassing disease models and humans. Current machine learning methods primarily establish statistical correlations between genotypes and phenotypes but struggle to identify physiologically significant causal factors, limiting their predictive power. Key challenges in predictive modeling include scarcity of labeled data, generalization across different domains, and disentangling causation from correlation. In light of recent advances in multi-omics data integration, we propose a new artificial intelligence (AI)-powered biology-inspired multi-scale modeling framework to tackle these issues. This framework will integrate multi-omics data across biological levels, organism hierarchies, and species to predict causal genotype-environment-phenotype relationships under various conditions. AI models inspired by biology may identify novel molecular targets, biomarkers, pharmaceutical agents, and personalized medicines for presently unmet medical needs.

cs.AI↗