arXiv subjects

Jun Gao

Publications and source records attributed to Jun Gao.

At least 145 records · Page 8Linked to original sources

arXiv2025

ER-RAG: Enhance RAG with ER-Based Unified Modeling of Heterogeneous Data Sources

Large language models (LLMs) excel in question-answering (QA) tasks, and retrieval-augmented generation (RAG) enhances their precision by incorporating external evidence from diverse sources like web pages, databases, and knowledge graphs. However, current RAG methods rely on agent-specific strategies for individual data sources, posing challenges low-resource or black-box environments and complicates operations when evidence is fragmented across sources. To address these limitations, we propose ER-RAG, a framework that unifies evidence integration across heterogeneous data sources using the Entity-Relationship (ER) model. ER-RAG standardizes entity retrieval and relationship querying through ER-based APIs with GET and JOIN operations. It employs a two-stage generation process: first, a preference optimization module selects optimal sources; second, another module constructs API chains based on source schemas. This unified approach allows efficient fine-tuning and seamless integration across diverse data sources. ER-RAG demonstrated its effectiveness by winning all three tracks of the 2024 KDDCup CRAG Challenge, achieving performance on par with commercial RAG pipelines using an 8B LLM backbone. It outperformed hybrid competitors by 3.1% in LLM score and accelerated retrieval by 5.5X.

cs.IR↗

arXiv2025

The impact of data from future lepton colliders on light hadrons fragmentation functions

In this work, we study the constraining power of future lepton colliders on fragmentation functions (FFs) to light charged hadrons from quarks and gluon in the framework of QCD collinear factorization. We perform analyses of FFs at NLO by including a wide range of pseudo--data from future lepton colliders, such as measurements on hadron multiplicities in the production of two jets and $W$ boson pairs, at various center of mass energies, and from hadronic decays of the Higgs boson, including both to heavy quarks and to gluons. The high luminosity and high energies of future lepton colliders allow for quark flavor separations and ensure a precise determination of FFs based solely on data from electron-positron collisions. We find that either the CEPC, FCC-$ee$ or ILC can significantly reduce the uncertainties of FFs in a wide kinematic range, compared to the NPC23 set obtained with a global analysis to current world data. We also discuss the impact of higher-order QCD corrections, and the potential constraints from measurements of three-jet production. Furthermore, we describe an update of the FMNLO program allowing for calculating hadron production cross sections at next-to-next-to-leading order in QCD, which is used in this study.

hep-ph↗

arXiv2025

SpaceMesh: A Continuous Representation for Learning Manifold Surface Meshes

Meshes are ubiquitous in visual computing and simulation, yet most existing machine learning techniques represent meshes only indirectly, e.g. as the level set of a scalar field or deformation of a template, or as a disordered triangle soup lacking local structure. This work presents a scheme to directly generate manifold, polygonal meshes of complex connectivity as the output of a neural network. Our key innovation is to define a continuous latent connectivity space at each mesh vertex, which implies the discrete mesh. In particular, our vertex embeddings generate cyclic neighbor relationships in a halfedge mesh representation, which gives a guarantee of edge-manifoldness and the ability to represent general polygonal meshes. This representation is well-suited to machine learning and stochastic optimization, without restriction on connectivity or topology. We first explore the basic properties of this representation, then use it to fit distributions of meshes from large datasets. The resulting models generate diverse meshes with tessellation structure learned from the dataset population, with concise details and high-quality mesh elements. In applications, this approach not only yields high-quality outputs from generative models, but also enables directly learning challenging geometry processing tasks such as mesh repair.

cs.CV↗

arXiv2025

Strong Ramsey game on two boards

The strong Ramsey game $R(\mathcal{B}, H)$ is a two-player game played on a graph $\mathcal{B}$, referred to as the board, with a target graph $H$. In this game, two players, $P_1$ and $P_2$, alternately claim unclaimed edges of $\mathcal{B}$, starting with $P_1$. The goal is to claim a subgraph isomorphic to $H$, with the first player achieving this declared the winner. A fundamental open question, persisting for over three decades, asks whether there exists a graph $H$ such that in the game $R(K_n, H)$, $P_1$ does not have a winning strategy in a bounded number of moves as $n \to \infty$. In this paper, we shift the focus to the variant $R(K_n \sqcup K_n, H)$, introduced by David, Hartarsky, and Tiba, where the board $K_n \sqcup K_n$ consists of two disjoint copies of $K_n$. We prove that there exist infinitely many graphs $H$ such that $P_1$ cannot win in $R(K_n \sqcup K_n, H)$ within a bounded number of moves through a concise proof. This perhaps provides evidence for the existence of examples to the above longstanding open problem.

math.CO↗

arXiv2024

Global analysis of fragmentation functions to charged hadrons with high-precision data from the LHC

Fragmentation functions (FFs) are essential non-perturbative QCD inputs for predicting hadron production cross sections in high energy scatterings. In this study, we present a joint determination of FFs for light charged hadrons through a global analysis at next-to-leading order (NLO) in QCD. Our analysis incorporates a wide range of precision measurements from the LHC, as well as data from electron-positron collisions and semi-inclusive deep inelastic scatterings. By including measurements of jet fragmentation at the LHC in our global analysis, we are able to impose strong constraints on the gluon FFs. A careful selection of hadron kinematics is applied to ensure the validity of factorization and perturbative calculations of QCD. In addition, we introduce several methodological advances in fitting, resulting in a flexible parametrization form and the inclusion of theoretical uncertainties from perturbative calculations. Our best-fit predictions show very good agreement with the global data, with $χ^2/N_{pt}\sim 0.90$. We also generate a large number of Hessian error sets to estimate uncertainties and correlations of the extracted FFs. FFs to charged pions (kaons and protons) are well constrained for momentum fractions down to 0.01 (0.1). Total momentum of partons carried by light charged hadrons are determined precisely. Their values for $u$, $d$ quarks and gluon saturate at about 50\% for a lower cut of the momentum fraction of 0.01. Pulls from individual datasets and impact of various choices of the analysis are also studied in details. Additionally, we present an update of the FMNLO program used for calculating hadron production cross sections. Our FFs, including the error sets (denoted as NPC23), are publicly available in the form of LHAPDF6 grids.

hep-ph↗

arXiv2024

A framework for simultaneous fit of QCD and BSM parameters with xFitter

An extension of the xFitter open-source program for QCD analyses is presented, allowing for a polynomial parameterization of the dependence of physical observables on theoretical parameters. This extension enables simultaneous determination of parton distribution functions (PDFs) and new physics parameters within the framework of the Standard Model Effective Field Theory (SMEFT). The functionalities of the code are illustrated in a sensitivity study, where a simultaneous determination of the PDFs, top quark mass and the couplings of selected dimension-6 SMEFT operators is addressed using projections for measurements of top quark-antiquark pair production at the High-Luminosity LHC. The importance of considering all the correlations of the parton distributions, top quark mass and the EFT parameters in a simultaneous QCD+SMEFT analysis is demonstrated. This work serves as a new platform for simultaneous extraction of the PDFs and the SM/SMEFT parameters based on xFitter.

hep-ph↗

arXiv2024

Induced even cycles in locally sparse graphs

A graph $G$ is $(c,t)$-sparse if for every pair of vertex subsets $A,B\subset V(G)$ with $|A|,|B|\geq t$, $e(A,B)\leq (1-c)|A||B|$. In this paper we prove that for every $c>0$ and integer $\ell$, there exists $C>1$ such that if an $n$-vertex graph $G$ is $(c,t)$-sparse for some $t$, and has at least $C t^{1-1/\ell}n^{1+1/\ell}$ edges, then $G$ contains an induced copy of $C_{2\ell}$. This resolves a conjecture of Fox, Nenadov and Pham.

math.CO↗

arXiv2024

Angularity in Higgs boson decays via $\boldsymbol{H\to gg}$ at NNLL$'$ accuracy

We present improved predictions of a class of event-shape distributions called angularity for a contribution from an effective operator $H\to gg$ in Higgs hadronic decay that suffers from large perturbative uncertainties. In the frame of Soft-Collinear Effective Theory, logarithmic terms of the distribution are resummed at NNLL$'$ accuracy, for which 2-loop constant of gluon-jet function for angularity is independently determined by a fit to fixed-order distribution at NLO corresponding to ${\mathcal{O}}(α_s^2)$ relative to the Born rate. Our determination shows reasonable agreement with the value in a thesis recently released. In the fit, we use an asymptotic form with a fractional power conjectured from recoil corrections at one-loop order and it improves the accuracy of determination in positive values of angularity parameter $a$. The resummed distribution is matched to the NLO fixed-order results to make our predictions valid at all angularity values. We also discuss the first moment and subtracted moment of angularity as a function of $a$ that allow to extract information on leading and subleading nonperturbative corrections associated with gluons.

hep-ph↗

arXiv2024

HelloMeme: Integrating Spatial Knitting Attentions to Embed High-Level and Fidelity-Rich Conditions in Diffusion Models

We propose an effective method for inserting adapters into text-to-image foundation models, which enables the execution of complex downstream tasks while preserving the generalization ability of the base model. The core idea of this method is to optimize the attention mechanism related to 2D feature maps, which enhances the performance of the adapter. This approach was validated on the task of meme video generation and achieved significant results. We hope this work can provide insights for post-training tasks of large text-to-image models. Additionally, as this method demonstrates good compatibility with SD1.5 derivative models, it holds certain value for the open-source community. Therefore, we will release the related code (\url{https://songkey.github.io/hellomeme}).

cs.CV↗

arXiv2024

TV-SAM: Increasing Zero-Shot Segmentation Performance on Multimodal Medical Images Using GPT-4 Generated Descriptive Prompts Without Human Annotation

This study presents a novel multimodal medical image zero-shot segmentation algorithm named the text-visual-prompt segment anything model (TV-SAM) without any manual annotations. The TV-SAM incorporates and integrates the large language model GPT-4, the vision language model GLIP, and the SAM to autonomously generate descriptive text prompts and visual bounding box prompts from medical images, thereby enhancing the SAM's capability for zero-shot segmentation. Comprehensive evaluations are implemented on seven public datasets encompassing eight imaging modalities to demonstrate that TV-SAM can effectively segment unseen targets across various modalities without additional training. TV-SAM significantly outperforms SAM AUTO and GSAM, closely matching the performance of SAM BBOX with gold standard bounding box prompts and surpasses the state-of-the-art methods on specific datasets such as ISIC and WBC. The study indicates that TV-SAM serves as an effective multimodal medical image zero-shot segmentation algorithm, highlighting the significant contribution of GPT-4 to zero-shot segmentation. By integrating foundational models such as GPT-4, GLIP, and SAM, the ability to address complex problems in specialized domains can be enhanced.

cs.CV↗

arXiv2024

BSG4Bot: Efficient Bot Detection based on Biased Heterogeneous Subgraphs

The detection of malicious social bots has become a crucial task, as bots can be easily deployed and manipulated to spread disinformation, promote conspiracy messages, and more. Most existing approaches utilize graph neural networks (GNNs)to capture both user profle and structural features,achieving promising progress. However, they still face limitations including the expensive training on large underlying graph, the performance degration when similar neighborhood patterns' assumption preferred by GNNs is not satisfied, and the dynamic features of bots in a highly adversarial context. Motivated by these limitations, this paper proposes a method named BSG4Bot with an intuition that GNNs training on Biased SubGraphs can improve both performance and time/space efficiency in bot detection. Specifically, BSG4Bot first pre-trains a classifier on node features efficiently to define the node similarities, and constructs biased subgraphs by combining the similarities computed by the pre-trained classifier and the node importances computed by Personalized PageRank (PPR scores). BSG4Bot then introduces a heterogeneous GNN over the constructed subgraphs to detect bots effectively and efficiently. The relatively stable features, including the content category and temporal activity features, are explored and incorporated into BSG4Bot after preliminary verification on sample data. The extensive experimental studies show that BSG4Bot outperforms the state-of-the-art bot detection methods, while only needing nearly 1/5 training time.

cs.LG↗

arXiv2024

On demand single photon generation and coherent control of excitons from resonantly driven nanowire quantum dots

Coherent control of single photon sources is a key requirement for the advancement of photonic quantum technologies. Among them, nanowire-based quantum dot sources are popular due to their potential for on-chip hybrid integration. Here we demonstrate on-demand single-photon generation ($g^{(2)}(0)(X^{*}) =0.078$ and $g^{(2)}(0)(X)= 0.03$) from resonantly excited InAsP/InP nanowire quantum dots and observe Rabi oscillations in the dot emission, indicating successful coherent manipulation of the excitonic states in the nanowire. We also measure a low emission time jitter for resonant excitation as compared to above-band excitation. This work addresses the long-standing challenge of resonantly exciting nanowire-quantum dots. It paves the way for hybrid quantum photonic integration, enabling spin-photon entanglement and matter memories on-chip.

cond-mat.mes-hall↗

arXiv2024

Probing multi-mobility edges in quasiperiodic mosaic lattices

The mobility edge (ME) is a crucial concept in understanding localization physics, marking the critical transition between extended and localized states in the energy spectrum. Anderson localization scaling theory predicts the absence of ME in lower dimensional systems. Hence, the search for exact MEs, particularly for single particles in lower dimensions, has recently garnered significant interest in both theoretical and experimental studies, resulting in notable progress. However, several open questions remain, including the possibility of a single system exhibiting multiple MEs and the continual existence of extended states, even within the strong disorder domain. Here, we provide experimental evidence to address these questions by utilizing a quasiperiodic mosaic lattice with meticulously designed nanophotonic circuits. Our observations demonstrate the coexistence of both extended and localized states in lattices with broken duality symmetry and varying modulation periods. By single site injection and scanning the disorder level, we could approximately probe the ME of the modulated lattice. These results corroborate recent theoretical predictions, introduce a new avenue for investigating ME physics, and offer inspiration for further exploration of ME physics in the quantum regime using hybrid integrated photonic devices.

cond-mat.dis-nn↗

arXiv2024

Winning Solution For Meta KDD Cup' 24

This paper describes the winning solutions of all tasks in Meta KDD Cup 24 from db3 team. The challenge is to build a RAG system from web sources and knowledge graphs. We are given multiple sources for each query to help us answer the question. The CRAG challenge involves three tasks: (1) condensing information from web pages into accurate answers, (2) integrating structured data from mock knowledge graphs, and (3) selecting and integrating critical data from extensive web pages and APIs to reflect real-world retrieval challenges. Our solution for Task #1 is a framework of web or open-data retrieval and answering. The large language model (LLM) is tuned for better RAG performance and less hallucination. Task #2 and Task #3 solutions are based on a regularized API set for domain questions and the API generation method using tuned LLM. Our knowledge graph API interface extracts directly relevant information to help LLMs answer correctly. Our solution achieves 1st place on all three tasks, achieving a score of 28.4%, 42.7%, and 47.8%, respectively.

cs.IR↗

arXiv2024

Evolving Diverse Red-team Language Models in Multi-round Multi-agent Games

The primary challenge in deploying Large Language Model (LLM) is ensuring its harmlessness. Red team can identify vulnerabilities by attacking LLM to attain safety. However, current efforts heavily rely on single-round prompt designs and unilateral red team optimizations against fixed blue teams. These static approaches lead to significant reductions in generation diversity, known as the mode collapse, which makes it difficult to discover the potential risks in the increasingly complex human-LLM interactions. Here we introduce dynamic Red Team Game (RTG) to comprehensively analyze the multi-round offensive and defensive interactions between red team and blue team. Furthermore, we develop a Gamified Red Team Solver (GRTS) with diversity measures to mitigate mode collapse and theoretically guarantee the convergence of approximate Nash equilibrium which results in better strategies for both teams. Empirical results demonstrate that GRTS explore diverse and implicit attacks to adaptively exploit various LLMs, surpassing the constraints of specific modes. Insightfully, the geometrical structure we unveil of the red team task aligns with the spinning top hypothesis, confirming the necessity of constructing a diverse LLM population as a promising proxy for heterogeneous human expert red-teamers. This paves the way for scalable toxicity detection and safe alignment for LLMs.

cs.CL↗

arXiv2024

Acoustic modulation of individual nanowire quantum dots integrated into a hybrid thin-film lithium niobate photonic platform

Surface acoustic waves (SAWs) are a powerful tool for controlling a wide range of quantum systems, particularly quantum dots (QDs) via their oscillating strain fields. The resulting energy modulation of these single photon sources can be harnessed to achieve spectral overlap between two QDs otherwise emitting at different wavelengths. In this study, we integrate InAsP/InP nanowire quantum dots onto a thin-film lithium niobate platform, a strong piezoelectric material, and embed them within Si$_{3}$N$_{4}$-loaded waveguides. We demonstrate emission wavelength modulation of 0.70 nm at 13 dBm with a single focused interdigital transducer (FIDT) operating at 400 MHz, and achieve twice this modulation by using two FIDTs as an acoustic cavity. Additionally, we bring two QDs with an initial wavelength difference of 0.5 nm into resonance using SAWs. This scalable strain-tuning approach represents a significant step towards producing indistinguishable single photons from remote emitters heterogeneously integrated on a single photonic chip, and paves the way for large scale on-chip quantum information processing using photonic platforms.

cond-mat.mes-hall↗

arXiv2024

AIM: Let Any Multi-modal Large Language Models Embrace Efficient In-Context Learning

In-context learning (ICL) facilitates Large Language Models (LLMs) exhibiting emergent ability on downstream tasks without updating billions of parameters. However, in the area of multi-modal Large Language Models (MLLMs), two problems hinder the application of multi-modal ICL: (1) Most primary MLLMs are only trained on single-image datasets, making them unable to read multi-modal demonstrations. (2) With the demonstrations increasing, thousands of visual tokens highly challenge hardware and degrade ICL performance. During preliminary explorations, we discovered that the inner LLM tends to focus more on the linguistic modality within multi-modal demonstrations to generate responses. Therefore, we propose a general and light-weighted framework \textbf{AIM} to tackle the mentioned problems through \textbf{A}ggregating \textbf{I}mage information of \textbf{M}ultimodal demonstrations to the dense latent space of the corresponding linguistic part. Specifically, AIM first uses the frozen backbone MLLM to read each image-text demonstration and extracts the vector representations on top of the text. These vectors naturally fuse the information of the image-text pair, and AIM transforms them into fused virtual tokens acceptable for the inner LLM via a trainable projection layer. Ultimately, these fused tokens function as variants of multi-modal demonstrations, fed into the MLLM to direct its response to the current query as usual. Because these fused tokens stem from the textual component of the image-text pair, a multi-modal demonstration is nearly reduced to a pure textual demonstration, thus seamlessly applying to any MLLMs. With its de facto MLLM frozen, AIM is parameter-efficient and we train it on public multi-modal web corpora which have nothing to do with downstream test tasks.

cs.MM↗

arXiv2024

Simultaneous Determination of Fragmentation Functions and Test on Momentum Sum Rule

We perform a simultaneous global analysis of hadron fragmentation functions (FFs) to various charged hadrons at next-to-leading order in QCD. The world data set includes results from electron-positron single-inclusive annihilation, semi-inclusive deep inelastic scattering, as well as proton-proton collisions including jet fragmentation measurements which lead to strong constraints on the gluon fragmentations. By carefully selecting hadron kinematics to ensure the validity of QCD factorization and the convergence of perturbative calculations, we achieve a satisfying best fit with $χ^2/$d.o.f.$=0.90$, in the simultaneous extraction of FFs for light charged hadrons ($π^{\pm}$, $K^{\pm}$ and $p/\bar{p}$). The total momentum of $u$, $d$ quarks and gluon carried by light charged hadrons have been determined precisely. That urges future precision measurements on fragmentation to neutral hadrons, which are crucial for the test of fundamental sum rules in QCD fragmentation.

hep-ph↗