arXiv subjects

Qiang Mao

Publications and source records attributed to Qiang Mao.

14 recordsLinked to original sources

arXiv2025

A Comparative Study of Retrieval Methods in Azure AI Search

Increasingly, attorneys are interested in moving beyond keyword and semantic search to improve the efficiency of how they find key information during a document review task. Large language models (LLMs) are now seen as tools that attorneys can use to ask natural language questions of their data during document review to receive accurate and concise answers. This study evaluates retrieval strategies within Microsoft Azure's Retrieval-Augmented Generation (RAG) framework to identify effective approaches for Early Case Assessment (ECA) in eDiscovery. During ECA, legal teams analyze data at the outset of a matter to gain a general understanding of the data and attempt to determine key facts and risks before beginning full-scale review. In this paper, we compare the performance of Azure AI Search's keyword, semantic, vector, hybrid, and hybrid-semantic retrieval methods. We then present the accuracy, relevance, and consistency of each method's AI-generated responses. Legal practitioners can use the results of this study to enhance how they select RAG configurations in the future.

cs.IR↗

arXiv2025

Leveraging Machine Learning and Large Language Models for Automated Image Clustering and Description in Legal Discovery

The rapid increase in digital image creation and retention presents substantial challenges during legal discovery, digital archive, and content management. Corporations and legal teams must organize, analyze, and extract meaningful insights from large image collections under strict time pressures, making manual review impractical and costly. These demands have intensified interest in automated methods that can efficiently organize and describe large-scale image datasets. This paper presents a systematic investigation of automated cluster description generation through the integration of image clustering, image captioning, and large language models (LLMs). We apply K-means clustering to group images into 20 visually coherent clusters and generate base captions using the Azure AI Vision API. We then evaluate three critical dimensions of the cluster description process: (1) image sampling strategies, comparing random, centroid-based, stratified, hybrid, and density-based sampling against using all cluster images; (2) prompting techniques, contrasting standard prompting with chain-of-thought prompting; and (3) description generation methods, comparing LLM-based generation with traditional TF-IDF and template-based approaches. We assess description quality using semantic similarity and coverage metrics. Results show that strategic sampling with 20 images per cluster performs comparably to exhaustive inclusion while significantly reducing computational cost, with only stratified sampling showing modest degradation. LLM-based methods consistently outperform TF-IDF baselines, and standard prompts outperform chain-of-thought prompts for this task. These findings provide practical guidance for deploying scalable, accurate cluster description systems that support high-volume workflows in legal discovery and other domains requiring automated organization of large image collections.

cs.IR↗

arXiv2025

Empirical Evaluation of Embedding Models in the Context of Text Classification in Document Review in Construction Delay Disputes

Text embeddings are numerical representations of text data, where words, phrases, or entire documents are converted into vectors of real numbers. These embeddings capture semantic meanings and relationships between text elements in a continuous vector space. The primary goal of text embeddings is to enable the processing of text data by machine learning models, which require numerical input. Numerous embedding models have been developed for various applications. This paper presents our work in evaluating different embeddings through a comprehensive comparative analysis of four distinct models, focusing on their text classification efficacy. We employ both K-Nearest Neighbors (KNN) and Logistic Regression (LR) to perform binary classification tasks, specifically determining whether a text snippet is associated with 'delay' or 'not delay' within a labeled dataset. Our research explores the use of text snippet embeddings for training supervised text classification models to identify delay-related statements during the document review process of construction delay disputes. The results of this study highlight the potential of embedding models to enhance the efficiency and accuracy of document analysis in legal contexts, paving the way for more informed decision-making in complex investigative scenarios.

cs.IR↗

arXiv2022

Identifying the $Ξ_b(6100)$ as the $P$-wave bottom baryon of $J^P=3/2^-$

We study the $Ξ_b(6100)$ using the methods of QCD sum rules and light-cone sum rules within the framework of heavy quark effective theory. Our results suggest that the $Ξ_b(6100)$ can be well interpreted as the $P$-wave bottom baryon of $J^P=3/2^-$, belonging to the $SU(3)$ flavor $\mathbf{\bar 3}_F$ representation. It has a partner state of $J^P=1/2^-$, labelled as $Ξ_b(1/2^-)$, whose mass and width are predicted to be $m_{Ξ_b(1/2^-)}=6.08^{+0.13}_{-0.11}$~GeV and $Γ_{Ξ_b(1/2^-)}=4^{+29}_{-~4}$~MeV, with the mass splitting $ΔM=m_{Ξ_b(6100)}-m_{Ξ_b(1/2^-)}=9\pm3$~MeV. We propose to search for it in the $Ξ_c({1/2}^-)\to Ξ_b^{\prime}π$ decay channel. Our results also suggest that the $Λ_b(5912)$ and $Λ_b(5920)$ are their partner states with $J^P=1/2^-$ and $3/2^-$ respectively, and moreover, the $Λ_c(2595)$, $Λ_c(2625)$, $Ξ_c(2790)$, and $Ξ_c(2815)$ are their charmed partner states.

hep-ph↗

arXiv2020

Identifying the $Ξ_c^0$ baryons observed by LHCb as $P$-wave $Ξ_c^\prime$ baryons

We systematically study mass spectra and decay properties of $P$-wave $Ξ_c^\prime$ baryons of the $SU(3)$ flavor $\mathbf{6}_F$, using the methods of QCD sum rules and light-cone sum rules within the framework of heavy quark effective theory. Our results suggest that the three excited $Ξ_c^0$ baryons recently observed by LHCb can be well explained as $P$-wave $Ξ_c^\prime$ baryons: the $Ξ_c(2923)^0$ and $Ξ_c(2939)^0$ are partner states of $J^P = 1/2^-$ and $3/2^-$ respectively, both of which contain one $λ$-mode orbital excitation; the $Ξ_c(2965)^0$ has $J^P = 3/2^-$, and also contains one $λ$-mode orbital excitation. We propose to search for another $P$-wave $Ξ_c^\prime$ state of $J^P = 5/2^-$ in the $Λ_c K/Ξ_c π$ mass spectral in future experiments. Its mass is about $56^{+30}_{-35}$ MeV larger than the $Ξ_c(2965)^0$, and its width is about $18.1^{+19.7}_{-~8.3}$ MeV.

hep-ph↗

arXiv2020

Identifying the $Λ_b(6146)^0$ and $Λ_b(6152)^0$ as $D$-wave bottom baryons

We study the $Λ_b(6146)^0$ and $Λ_b(6152)^0$ recently observed by LHCb using the method of QCD sum rules within the framework of heavy quark effective theory. Our results suggest that they can be interpreted as $D$-wave bottom baryons of $J^P = 3/2^+$ and $5/2^+$ respectively, both of which contain two $λ$-mode excitations. We also investigate other possible assignments containing $ρ$-mode excitations. We predict masses of their strangeness partners to be $m_{Ξ_b(3/2^+)} = 6.26^{+0.11}_{-0.14}$ GeV and $m_{Ξ_b(5/2^+)} = 6.26^{+0.11}_{-0.14}$ GeV with the mass splitting $ΔM = 4.5^{+1.9}_{-1.5}$ MeV, and propose to search for them in future LHCb and CMS experiments.

hep-ph↗

arXiv2020

Excited $Ω_b$ baryons and fine structure of strong interaction

The heavy baryon system bounded by the strong interaction has a rich internal structure, so its mass spectra can have the fine structure similar to the line spectra of atom bounded by the electromagnetic interaction. We systematically study the internal structure of $P$-wave $Ω_b$ baryons and calculate their $D$-wave decay properties. The present study, together with our previous studies on their mass spectra and $S$-wave decay properties, suggest that all the four excited $Ω_b$ baryons recently discovered by LHCb can be well explained as $P$-wave $Ω_b$ baryons, and their beautiful fine structure is directly related to the rich internal structure of $P$-wave $Ω_b$ baryons.

hep-ph↗

arXiv2020

Decay properties of $P$-wave heavy baryons accompanied by vector mesons within light-cone sum rules

We use the method of light-cone sum rules to study decay properties of $P$-wave bottom baryons belonging to the $SU(3)$ flavor $\mathbf{6}_F$ representation. In Ref.~\cite{Cui:2019dzj} we have studied their mass spectrum and pionic decays, and found that the $Σ_{b}(6097)$ and $Ξ_{b}(6227)$ can be well interpreted as $P$-wave bottom baryons of $J^P = 3/2^-$. In this paper we further study their decays into ground-state bottom baryons and vector mesons. We propose to search for a new state $Ξ_b({5/2}^-)$, that is the $J^P = 5/2^-$ partner state of the $Ξ_{b}(6227)$, in the $Ξ_b({5/2}^-) \to Ξ_b^{*}ρ\to Ξ_b^{*}ππ$ decay process. Its mass is $12 \pm 5$~MeV larger than that of the $Ξ_{b}(6227)$.

hep-ph↗

arXiv2017

Establishing low-lying doubly charmed baryons

We systematically study the $S$-wave doubly charmed baryons using the method of QCD sum rules. Our results suggest that the $Ξ_{cc}^{++}$ recently observed by LHCb can be well identified as the $S$-wave $Ξ_{cc}$ state of $J^P = 1/2^+$. We study its relevant $Ω_{cc}$ state, whose mass is predicted to be around 3.7 GeV. We also systematically study the $P$-wave doubly charmed baryons, whose masses are predicted to be around 4.1 GeV. Especially, there can be several excited doubly charmed baryons in this energy region, and we suggest to search for them in order to study the fine structure of the strong interaction.

hep-ph↗

arXiv2017

$D$-wave heavy baryons of the $SU(3)$ flavor $\mathbf{6}_F$

We use the method of QCD sum rules to study the $D$-wave charmed and bottom baryons of the $SU(3)$ flavor $\mathbf{6}_F$, and calculate their masses up to the order $\mathcal{O}(1/m_Q)$ with $m_Q$ the heavy quark mass. Our results suggest that the $Ξ_c(3123)$ can be well interpreted as a $D$-wave $Ξ_c^\prime(\mathbf{6}_F)$ state, and it probably has a partner state close to it. Our results also suggest that there may exist as many as four $D$-wave $Ω_c$ states in the energy region $3.3\sim3.5$ GeV, and we propose to search for them in the further LHCb and BelleII experiments.

hep-ph↗

arXiv2017

Decay properties of $P$-wave charmed baryons from light-cone QCD sum rules

We study decay properties of the $P$-wave charmed baryons using the method of light-cone QCD sum rules, including the $S$-wave decays of the flavor $\mathbf{\bar 3}_F$ $P$-wave charmed baryons into ground-state charmed baryons accompanied by a pseudoscalar meson ($π$ or $K$) or a vector meson ($ρ$ or $K^*$), and the $S$-wave decays of the flavor $\mathbf{6}_F$ $P$-wave charmed baryons into ground-state charmed baryons accompanied by a pseudoscalar meson ($π$ or $K$). We study both two-body and three-body decays which are kinematically allowed. We find two mixing solutions from internal $ρ$- and $λ$-mode excitations, which can well describe both masses and decay properties of the $Λ_c(2595)$, $Λ_c(2625)$, $Ξ_c(2790)$ and $Ξ_c(2815)$. We also discuss the possible interpretations of $P$-wave charmed baryons for the $Σ_c(2800)$, $Ξ_c(2930)$, $Ξ_c(2980)$, and the recently observed $Ω_c(3000)$, $Ω_c(3050)$, $Ω_c(3066)$, $Ω_c(3090)$, and $Ω_c(3119)$.

hep-ph↗

arXiv2016

D-wave charmed and bottomed baryons from QCD sum rules

We study the $D$-wave charmed baryons of $SU(3)$ flavor $\mathbf{\bar3}_F$ using the method of QCD sum rule in the framework of heavy quark effective theory (HQET). We find that the $Λ_c(2880)$, $Ξ_c(3055)$ and $Ξ_c(3080)$ can be well described by the $D$-wave $SU(3)$ $\mathbf{\bar 3}_F$ charmed baryon multiplets of $J^P=3/2^+$ and $5/2^+$, which contain two $λ$-mode orbital excitations, i.e., the $Λ_c(2880)$ has $J^P=5/2^+$, and the $Ξ_c(3055)$ and $Ξ_c(3080)$ have $J^P = 3/2^+$ and $5/2^+$, respectively. Our results also suggest that the $Λ_c(2880)$ has a partner state, the $Λ_c(3/2^+)$ of $J^P=3/2^+$. Its mass is around $2.81 ^{+0.33}_{-0.18}$ GeV, and the mass difference between it and the $Λ_c(2880)$ is $28^{+45}_{-24}$ MeV. We also evaluate the masses of their bottom partners.

hep-ph↗

arXiv2015

QCD sum rule calculation for P-wave bottom baryons

We study the P-wave bottom baryons using the method of QCD sum rule and heavy quark effective theory. Our results suggest that Lambda_b(5912) and Lambda_b(5920) can be well described by the baryon doublet [3F_bar, 1, 1, rho], and they belong to the SU(3) 3F_bar multiplets of J(P)=1/2(-) and 3/2(-). Their SU(3) flavor partners, Xi_b(1/2-) and Xi_b(3/2-), have masses 6.06 \pm 0.13 GeV and 6.07 \pm 0.13 GeV, respectively, with mass splitting 9 \pm 4 MeV. The results obtained using baryon doublet [3F_bar, 1, 0, lambda] are similar and also consistent with the experimental data. We also study the SU(3) 6F multiplets by using the baryon multiplets [6F, 0, 1, lambda], [6_F, 1, 0, rho] and [6F, 2, 1, lambda], and our results suggest that the P-wave bottom baryons Sigma_b, Xi'_b and Omega_b have (averaged) masses about 6.0 GeV, 6.2 GeV and 6.4 GeV, respectively.

hep-ph↗

arXiv2015

P-wave charmed baryons from QCD sum rules

We study the P-wave charmed baryons using the method of QCD sum rule in the framework of heavy quark effective theory (HQET). We consider systematically all possible baryon currents with a derivative for internal rho- and lambda-mode excitations. We have found a good working window for the currents corresponding to the rho-mode excitations for Lambda_c(2595), Lambda_c(2625), Xi_c(2790) and Xi_c(2815) which complete two SU(3) 3F_bar multiplets of J(P)=1/2(-) and 3/2(-), while the currents corresponding to the lambda-mode excitations seem also consistent with the data. Our results also suggest that there are two Sigma_c(2800) states of J(P)=1/2(-) and 3/2(-) whose mass splitting is 14 \pm 7 MeV, and two Xi_c(2980) states whose mass splitting is 12 \pm 7 MeV. They have two Omega_c partners of J(P) = 1/2(-) and 3/2(-), whose masses are around 3.25 \pm 0.20 GeV with mass splitting 10 \pm 6 MeV. All of them together complete two SU(3) 6F multiplets of J(P)=1/2(-) and 3/2(-). They may also have J(P)=5/2(-) partners. Xi_c(3080) may be one of them, and the other two are Sigma_c(5/2(-)) and Omega_c(5/2(-)), whose masses are 85 \pm 23 and 50 \pm 27 MeV larger.

hep-ph↗