Search

Large Language Models (LLMs) hold promise for advancing legal practice by automating complex tasks and improving access to justice. However, their adoption is limited by concerns over client confidentiality, especially when lawyers include sensitive Personally Identifiable Information (PII) in prompts, risking unauthorized data exposure. To mitigate this, we introduce LegalGuardian, a lightweight, privacy-preserving framework tailored for lawyers using LLM-based tools. LegalGuardian employs Named Entity Recognition (NER) techniques and local LLMs to mask and unmask confidential PII within prompts, safeguarding sensitive data before any external interaction. We detail its development and assess its effectiveness using a synthetic prompt library in immigration law scenarios. Comparing traditional NER models with one-shot prompted local LLM, we find that LegalGuardian achieves a F1-score of 93% with GLiNER and 97% with Qwen2.5-14B in PII detection. Semantic similarity analysis confirms that the framework maintains high fidelity in outputs, ensuring robust utility of LLM-based tools. Our findings indicate that legal professionals can harness advanced AI technologies without compromising client confidentiality or the quality of legal documents., Comment: 10 pages, 3 figures

Metal-organic frameworks (MOFs), with their high porosity and large internal surface area, provide versatile platforms for integrating spin centers with potential applications in catalysis and quantum sensing. Here, we identify a stable NH$\cdot$ radical spin center in UiO-66-NH$_2$, a zirconium-based MOF with aminoterephthalic acid (TPA-NH$_2$) linkers. Using electron paramagnetic resonance spectroscopy and density functional theory calculations, we determine the nature and spin Hamiltonian parameters of this radical. We also demonstrate that the NH$\cdot$ spin center exhibits a relatively long coherence time, making it a strong candidate for quantum sensing. Since this spin center is intrinsic to TPA-NH$_2$ linkers, our findings open new directions for leveraging organic radicals in MOFs for quantum technologies beyond the UiO-66-NH$_2$ family.

Unraveling the complexities of Gene Regulatory Networks (GRNs) is crucial for understanding cellular processes and disease mechanisms. Traditional computational methods often struggle with the dynamic nature of these networks. This study explores the use of Graph Neural Networks (GNNs), a powerful approach for modeling graph-structured data like GRNs. Utilizing a Graph Attention Network v2 (GATv2), our study presents a novel approach to the construction and interrogation of GRNs, informed by gene expression data and Boolean models derived from literature. The model's adeptness in accurately predicting regulatory interactions and pinpointing key regulators is attributed to advanced attention mechanisms, a hallmark of the GNN framework. These insights suggest that GNNs are primed to revolutionize GRN analysis, addressing traditional limitations and offering richer biological insights. The success of GNNs, as highlighted by our model's reliance on high-quality data, calls for enhanced data collection methods to sustain progress. The integration of GNNs in GRN research is set to pioneer developments in personalized medicine, drug discovery, and our grasp of biological systems, bolstered by the structural analysis of networks for improved node and edge prediction., Comment: 24 Pages, 6 Figures

Attention-Deficit/Hyperactivity Disorder (ADHD) is a challenging disorder to study due to its complex symptomatology and diverse contributing factors. To explore how we can gain deeper insights on this topic, we performed a network analysis on a comprehensive knowledge graph (KG) of ADHD, constructed by integrating scientific literature and clinical data with the help of cutting-edge large language models. The analysis, including k-core techniques, identified critical nodes and relationships that are central to understanding the disorder. Building on these findings, we curated a knowledge graph that is usable in a context-aware chatbot (Graph-RAG) with Large Language Models (LLMs), enabling accurate and informed interactions. Our knowledge graph not only advances the understanding of ADHD but also provides a powerful tool for research and clinical applications., Comment: 12 pages, 2 figures

This paper presents the design and implementation of a Federated Learning (FL) testbed, focusing on its application in cybersecurity and evaluating its resilience against poisoning attacks. Federated Learning allows multiple clients to collaboratively train a global model while keeping their data decentralized, addressing critical needs for data privacy and security, particularly in sensitive fields like cybersecurity. Our testbed, built using the Flower framework, facilitates experimentation with various FL frameworks, assessing their performance, scalability, and ease of integration. Through a case study on federated intrusion detection systems, we demonstrate the testbed's capabilities in detecting anomalies and securing critical infrastructure without exposing sensitive network data. Comprehensive poisoning tests, targeting both model and data integrity, evaluate the system's robustness under adversarial conditions. Our results show that while federated learning enhances data privacy and distributed learning, it remains vulnerable to poisoning attacks, which must be mitigated to ensure its reliability in real-world applications., Comment: 7 pages, 4 figures

The rapid evolution of cyber threats necessitates innovative solutions for detecting and analyzing malicious activity. Honeypots, which are decoy systems designed to lure and interact with attackers, have emerged as a critical component in cybersecurity. In this paper, we present a novel approach to creating realistic and interactive honeypot systems using Large Language Models (LLMs). By fine-tuning a pre-trained open-source language model on a diverse dataset of attacker-generated commands and responses, we developed a honeypot capable of sophisticated engagement with attackers. Our methodology involved several key steps: data collection and processing, prompt engineering, model selection, and supervised fine-tuning to optimize the model's performance. Evaluation through similarity metrics and live deployment demonstrated that our approach effectively generates accurate and informative responses. The results highlight the potential of LLMs to revolutionize honeypot technology, providing cybersecurity professionals with a powerful tool to detect and analyze malicious activity, thereby enhancing overall security infrastructure., Comment: 6 pages, 5 figures

Metal-Organic frameworks (MOFs) are promising candidates for advanced photocatalytically active materials. These porous crystalline compounds have large active surface areas and structural tunability and are thus highly competitive with oxides, the well-established material class for photocatalysis. However, due to their complex organic and coordination chemistry composition, photophysical mechanisms involved in the photocatalytic processes in MOFs are still not well understood. Employing electron paramagnetic resonance (EPR) spectroscopy and time-resolved photoluminescence spectroscopy (trPL), the fundamental processes of electron and hole generation are investigated, as well as capture events that lead to the formation of various radical species in UiO-66, an archetypical MOF photocatalyst. A manifold of photoinduced electron spin centers is detected, which is subsequently analyzed and identified with the help of density-functional theory (DFT) calculations. Under UV illumination, the symmetry, g-tensors and lifetimes of three distinct contributions are revealed: a surface O2-radical, a light-induced electron-hole pair, and a triplet exciton. Notably, the latter was found to emit (delayed) fluorescence. Our findings provide new insights into the photoinduced charge transfer processes, which are the basis of photocatalytic activity in UiO-66. This sets the stage for further studies on photogenerated spin centers in this and similar MOF materials., Comment: 23 pages, 12 figures

Environmental disasters such as floods, hurricanes, and wildfires have increasingly threatened communities worldwide, prompting various mitigation strategies. Among these, property buyouts have emerged as a prominent approach to reducing vulnerability to future disasters. This strategy involves governments purchasing at-risk properties from willing sellers and converting the land into open space, ostensibly reducing future disaster risk and impact. However, the aftermath of these buyouts, particularly concerning land-use patterns and community impacts, remains under-explored. This research aims to fill this gap by employing innovative techniques like satellite imagery analysis and deep learning to study these patterns. To achieve this goal, we employed FEMA's Hazard Mitigation Grant Program (HMGP) buyout dataset, encompassing over 41,004 addresses of these buyout properties from 1989 to 2017. Leveraging Google's Maps Static API, we gathered 40,053 satellite images corresponding to these buyout lands. Subsequently, we implemented five cutting-edge machine learning models to evaluate their performance in classifying land cover types. Notably, this task involved multi-class classification, and our model achieved an outstanding ROC-AUC score of 98.86%

Emergencies and critical incidents often unfold rapidly, necessitating a swift and effective response. In this research, we introduce a novel approach to identify and classify emergency situations from social media posts and direct emergency messages using an open source Large Language Model, LLAMA2. The goal is to harness the power of natural language processing and machine learning to assist public safety telecommunicators and huge crowds during countrywide emergencies. Our research focuses on developing a language model that can understand users describe their situation in the 911 call, enabling LLAMA2 to analyze the content and offer relevant instructions to the telecommunicator, while also creating workflows to notify government agencies with the caller's information when necessary. Another benefit this language model provides is its ability to assist people during a significant emergency incident when the 911 system is overwhelmed, by assisting the users with simple instructions and informing authorities with their location and emergency information.

We classify all the $6$-dimensional unimodular Lie algebras $\mathfrak{g}$ admitting a complex structure with non-zero closed $(3,0)$-form. This gives rise to $6$-dimensional compact homogeneous spaces $M=\Gamma\backslash G$, where $\Gamma$ is a lattice, admitting an invariant complex structure with holomorphically trivial canonical bundle. As an application, in the balanced Hermitian case, we study the instanton condition for any metric connection $\nabla^{\varepsilon,\rho}$ in the plane generated by the Levi-Civita connection and the Gauduchon line of Hermitian connections. In the setting of the Hull-Strominger system with connection on the tangent bundle being Hermitian-Yang-Mills, we prove that if a compact non-K\"ahler homogeneous space $M=\Gamma\backslash G$ admits an invariant solution with respect to some non-flat connection $\nabla$ in the family $\nabla^{\varepsilon,\rho}$, then $M$ is a nilmanifold with underlying Lie algebra $\mathfrak{h}_3$, a solvmanifold with underlying algebra $\mathfrak{g}_7$, or a quotient of the semisimple group SL(2,$\mathbb{C}$). Since it is known that the system can be solved on these spaces, our result implies that they are the unique compact non-K\"ahler balanced homogeneous spaces admitting such invariant solutions. As another application, on the compact solvmanifold underlying the Nakamura manifold, we construct solutions, on any given balanced Bott-Chern class, to the heterotic equations of motion taking the Chern connection as (flat) instanton.

In the current paper we study the groups, whose subnormal abelian subgroups are normal. We obtained a quite detailed description of such hyperabelian groups with a periodic Baer radical. The description of hyperabelian Lie algebras, whose abelian subideals are ideals, is also obtained.

The Grassmannian $\mathcal{G}_q(n,k)$ is the set of all $k$-dimensional subspaces of the vector space $\mathbb{F}_q^n$. K\"{o}tter and Kschischang showed that codes in Grassmannian space can be used for error-correction in random network coding. On the other hand, these codes are $q$-analogs of codes in the Johnson scheme, i.e., constant dimension codes. These codes of the Grassmannian $\mathcal{G}_q(n,k)$ also form a family of $q$-analogs of block designs and they are called subspace designs. In this paper, we examine one of the last families of $q$-analogs of block designs which was not considered before. This family, called subspace packings, is the $q$-analog of packings, and was considered recently for network coding solution for a family of multicast networks called the generalized combination networks. A subspace packing $t$-$(n,k,\lambda)_q$ is a set $\mathcal{S}$ of $k$-subspaces from $\mathcal{G}_q(n,k)$ such that each $t$-subspace of $\mathcal{G}_q(n,t)$ is contained in at most $\lambda$ elements of $\mathcal{S}$. The goal of this work is to consider the largest size of such subspace packings. We derive a sequence of lower and upper bounds on the maximum size of such packings, analyse these bounds, and identify the important problems for further research in this area., Comment: 30 pages, 27 tables, continuation of arXiv:1811.04611, typos corrected

We consider the Laplacian coflow of a $\mathrm{G}_2$-structure on warped products of the form $M^7= M^6 \times_f S^1$ with $M^6$ a compact 6-manifold endowed with an $\mathrm{SU}(3)$-structure. We give an explicit reinterpretation of this flow as a set of evolution equations of the differential forms defining the $\mathrm{SU}(3)$-structure on $M^6$ and the warping function $f$. Necessary and sufficient conditions for the existence of solution for this flow are given. Finally we describe new long time solutions for this flow where the $\mathrm{SU}(3)$-structure on $M^6$ is nearly K\"ahler, symplectic half-flat or balanced.

The Grassmannian ${\mathcal G}_q(n,k)$ is the set of all $k$-dimensional subspaces of the vector space $\mathbb{F}_q^n$. It is well known that codes in the Grassmannian space can be used for error-correction in random network coding. On the other hand, these codes are $q$-analogs of codes in the Johnson scheme, i.e. constant dimension codes. These codes of the Grassmannian ${\mathcal G}_q(n,k)$ also form a family of $q$-analogs of block designs and they are called \emph{subspace designs}. The application of subspace codes has motivated extensive work on the $q$-analogs of block designs. In this paper, we examine one of the last families of $q$-analogs of block designs which was not considered before. This family called \emph{subspace packings} is the $q$-analog of packings. This family of designs was considered recently for network coding solution for a family of multicast networks called the generalized combination networks. A \emph{subspace packing} $t$-$(n,k,\lambda)^m_q$ is a set $\mathcal{S}$ of $k$-subspaces from ${\mathcal G}_q(n,k)$ such that each $t$-subspace of ${\mathcal G}_q(n,t)$ is contained in at most $\lambda$ elements of $\mathcal{S}$. The goal of this work is to consider the largest size of such subspace packings., Comment: 10 pages, 3 tables, typos corrected

We study Hermitian metrics with a Gauduchon connection being "K\"ahler-like", namely, satisfying the same symmetries for curvature as the Levi-Civita and Chern connections. In particular, we investigate $6$-dimensional solvmanifolds with invariant complex structures with trivial canonical bundle and with invariant Hermitian metrics. The results for this case give evidence for two conjectures that are expected to hold in more generality: first, if the Strominger-Bismut connection is K\"ahler-like, then the metric is pluriclosed; second, if another Gauduchon connection, different from Chern or Strominger-Bismut, is K\"ahler-like, then the metric is K\"ahler. As a further motivation, we show that the K\"ahler-like condition for the Levi-Civita connection assures that the Ricci flow preserves the Hermitian condition along analytic solutions., Comment: to appear in Commun. Anal. Geom

We study the Laplacian flow of a $\mathrm{G}_2$-structure where this latter structure is claimed to be Locally Conformal Parallel. The first examples of long time solutions of this flow with the Locally Conformal Parallel condition are given. All of the solutions are ancient and Laplacian soliton of shrinking type. These examples are one-parameter families of Locally Conformal Parallel $\mathrm{G}_2$-structures on rank-one solvable extensions of six-dimensional nilpotent Lie groups. The found solutions are used to construct long time solutions to the Laplacian coflow starting from a Locally Conformal Parallel structure. We also study the behavior of the curvature of the solutions obtaining that for one of the examples the induced metric is Einstein along all the flow (resp. coflow)., Comment: 20 pages, some new results added and new presentation

Additive cyclic codes over Galois rings were investigated in previous works. In this paper, we investigate the same problem but over a more general ring family, finite commutative chain rings. When we focus on non-Galois finite commutative chain rings, we observe two different kinds of additivity. One of them is a natural generalization of the previous studies, whereas the other one has some unusual properties especially while constructing dual codes. We interpret the reasons of such properties and illustrate our results giving concrete examples.

We construct many new invariant solutions to the Strominger system with respect to a 2-parameter family of metric connections $\nabla^{\varepsilon,\rho}$ in the anomaly cancellation equation. The ansatz $\nabla^{\varepsilon,\rho}$ is a natural extension of the canonical 1-parameter family of Hermitian connections found by Gauduchon, as one recovers the Chern connection $\nabla^{c}$ for $({\varepsilon,\rho})=(0,\frac12)$, and the Bismut connection $\nabla^{+}$ for $({\varepsilon,\rho})=(\frac12,0)$. In particular, explicit invariant solutions to the Strominger system with respect to the Chern connection, with non-flat instanton and positive $\alpha'$ are obtained. Furthermore, we give invariant solutions to the heterotic equations of motion with respect to the Bismut connection. Our solutions live on three different compact non-K\"ahler homogeneous spaces, obtained as the quotient by a lattice of maximal rank of a nilpotent Lie group, the semisimple group SL(2,$\mathbb{C}$) and a solvable Lie group. To our knowledge, these are the only known invariant solutions to the heterotic equations of motion, and we conjecture that there is no other such homogeneous space admitting an invariant solution to the heterotic equations of motion with respect to a connection in the ansatz $\nabla^{\varepsilon,\rho}$., Comment: 27 pages, 3 figures

In this paper we present a method which aims to improve the spelling of children with dyslexia through playful and targeted exercises. In contrast to previous approaches, our method does not use correct words or positive examples to follow, but presents the child a misspelled word as an exercise to solve. We created these training exercises on the basis of the linguistic knowledge extracted from the errors found in texts written by children with dyslexia. To test the effectiveness of this method in Spanish, we integrated the exercises in a game for iPad, DysEggxia (Piruletras in Spanish), and carried out a within-subject experiment. During eight weeks, 48 children played either DysEggxia or Word Search, which is another word game. We conducted tests and questionnaires at the beginning of the study, after four weeks when the games were switched, and at the end of the study. The children who played DysEggxia for four weeks in a row had significantly less writing errors in the tests that after playing Word Search for the same time. This provides evidence that error-based exercises presented in a tablet help children with dyslexia improve their spelling skills., Comment: 8 pages, ASSETS'14, October 20-22, 2014, Rochester, NY, USA

We study the six-dimensional solvmanifolds that admit complex structures of splitting type classifying the underlying solvable Lie algebras. In particular, many complex structures of this type exist on the Nakamura manifold $X$, and they allow us to construct a countable family of compact complex non-$\partial\overline\partial$ manifolds $X_k$, $k\in\mathbb{Z}$, that admit a small holomorphic deformation $\{(X_{k})_{t}\}_{t\in\Delta_k}$ satisfying the $\partial\overline\partial$-Lemma for any $t\in\Delta_k$ except for the central fibre. Moreover, a study of the existence of special Hermitian metrics is also carried out on six-dimensional solvmanifolds with splitting-type complex structures.

We determine the 6-dimensional solvmanifolds admitting an invariant complex structure with holomorphically trivial canonical bundle. Such complex structures are classified up to isomorphism, and the existence of strong K\"ahler with torsion (SKT), generalized Gauduchon, balanced and strongly Gauduchon metrics is studied. As an application we construct a holomorphic family $(M,J_a)$ of compact complex manifolds such that $(M,J_a)$ satisfies the $\partial\bar\partial$-Lemma and admits a balanced metric for any $a\not=0$, but the central limit neither satisfies the $\partial\bar\partial$-Lemma nor admits balanced metrics., Comment: 32 pages; to appear in IMRN

We classify invariant complex structures on 6-dimensional nilmanifolds up to equivalence. As an application, the behaviour of the associated Fr\"olicher sequence is studied as well as its relation to the existence of strongly Gauduchon metrics. We also show that the strongly Gauduchon property and the balanced property are not closed under holomorphic deformation., Comment: 30 pages; the title has changed and the presentation has been improved, to appear in J. Geom. Anal

We classify solvable Lie groups admitting left invariant symplectic half-flat structure. When the Lie group has a compact quotient by a lattice, we show that these structures provide solutions of supersymmetric equations of type IIA., Comment: 16 pages; presentation improved, one reference added and typos corrected; to appear in Ann. Global Anal. Geom

R. Baer has proved that if the factor-group G/{\zeta}_{n}(G) of a group G by the member {\zeta}_{n}(G) of its upper central series is finite (here n is a positive integer) then the member {\gamma}_{n+1}(G) of the lower central series of G is also finite. In particular, in this case, the nilpotent residual of G is finite. This theorem admits the following simple generalization that has been published very recently by M. de Falco, F. de Giovanni, C. Musella and Ya. P. Sysak: "If the factor-group G/Z of a group G modulo its upper hypercenter Z is finite then G has a finite normal subgroup L such that G/L is hypercentral". In the current article we offer a new simpler very short proof of this theorem and specify it substantially. In fact, we prove that if |G/Z|=t then |L|\leqt^{k}, where k=(1/2)(log_{p}t+1), and p is the least prime divisor of t.

We show that for any hyperbolic surface of genus g, the eigenvalue $\lambda _{2g-2}$ of the Laplace operator is > 1/4., Comment: The final version of this article will be published in the Duke Mathematical Journal