-
Deep Content Understanding Toward Entity and Aspect Target Sentiment Analysis on Foundation Models
Authors:
Vorakit Vorakitphan,
Milos Basic,
Guilhaume Leroy Meline
Abstract:
Introducing Entity-Aspect Sentiment Triplet Extraction (EASTE), a novel Aspect-Based Sentiment Analysis (ABSA) task which extends Target-Aspect-Sentiment Detection (TASD) by separating aspect categories (e.g., food#quality) into pre-defined entities (e.g., meal, drink) and aspects (e.g., taste, freshness) which add a fine-gainer level of complexity, yet help exposing true sentiment of chained aspe…
▽ More
Introducing Entity-Aspect Sentiment Triplet Extraction (EASTE), a novel Aspect-Based Sentiment Analysis (ABSA) task which extends Target-Aspect-Sentiment Detection (TASD) by separating aspect categories (e.g., food#quality) into pre-defined entities (e.g., meal, drink) and aspects (e.g., taste, freshness) which add a fine-gainer level of complexity, yet help exposing true sentiment of chained aspect to its entity. We explore the task of EASTE solving capabilities of language models based on transformers architecture from our proposed unified-loss approach via token classification task using BERT architecture to text generative models such as Flan-T5, Flan-Ul2 to Llama2, Llama3 and Mixtral employing different alignment techniques such as zero/few-shot learning, Parameter Efficient Fine Tuning (PEFT) such as Low-Rank Adaptation (LoRA). The model performances are evaluated on the SamEval-2016 benchmark dataset representing the fair comparison to existing works. Our research not only aims to achieve high performance on the EASTE task but also investigates the impact of model size, type, and adaptation techniques on task performance. Ultimately, we provide detailed insights and achieving state-of-the-art results in complex sentiment analysis.
△ Less
Submitted 4 July, 2024;
originally announced July 2024.
-
Graphs with three and four distinct eigenvalues based on circulants
Authors:
Milan Bašić
Abstract:
In this paper, we aim to address the open questions raised in various recent papers regarding characterization of circulant graphs with three or four distinct eigenvalues in their spectra. Our focus is on providing characterizations and constructing classes of graphs falling under this specific category. We present a characterization of circulant graphs with prime number order and unitary Cayley g…
▽ More
In this paper, we aim to address the open questions raised in various recent papers regarding characterization of circulant graphs with three or four distinct eigenvalues in their spectra. Our focus is on providing characterizations and constructing classes of graphs falling under this specific category. We present a characterization of circulant graphs with prime number order and unitary Cayley graphs with arbitrary order, both of which possess spectra displaying three or four distinct eigenvalues. Various constructions of circulant graphs with composite orders are provided whose spectra consist of four distinct eigenvalues. These constructions primarily utilize specific subgraphs of circulant graphs that already possess two or three eigenvalues in their spectra, employing graph operations like the tensor product, the union, and the complement. Finally, we characterize the iterated line graphs of unitary Cayley graphs whose spectra contain three or four distinct eigenvalues, and we show their non-circulant nature.
△ Less
Submitted 9 October, 2023;
originally announced October 2023.
-
Maximal diameter of integral circulant graphs
Authors:
Milan Bašić,
Aleksandar Ilić,
Aleksandar Stamenković
Abstract:
Integral circulant graphs are proposed as models for quantum spin networks that permit a quantum phenomenon called perfect state transfer. Specifically, it is important to know how far information can potentially be transferred between nodes of the quantum networks modelled by integral circulant graphs and this task is related to calculating the maximal diameter of a graph. The integral circulant…
▽ More
Integral circulant graphs are proposed as models for quantum spin networks that permit a quantum phenomenon called perfect state transfer. Specifically, it is important to know how far information can potentially be transferred between nodes of the quantum networks modelled by integral circulant graphs and this task is related to calculating the maximal diameter of a graph. The integral circulant graph $ICG_n (D)$ has the vertex set $Z_n = \{0, 1, 2, \ldots, n - 1\}$ and vertices $a$ and $b$ are adjacent if $\gcd(a-b,n)\in D$, where $D \subseteq \{d : d \mid n,\ 1\leq d<n\}$. Motivated by the result on the upper bound of the diameter of $ICG_n(D)$ given in [N. Saxena, S. Severini, I. Shparlinski, \textit{Parameters of integral circulant graphs and periodic quantum dynamics}, International Journal of Quantum Information 5 (2007), 417--430], according to which $2|D|+1$ represents one such bound, in this paper we prove that the maximal value of the diameter of the integral circulant graph $ICG_n(D)$ of a given order $n$ with its prime factorization $p_1^{α_1}\cdots p_k^{α_k}$, is equal to $r(n)$ or $r(n)+1$, where $r(n)=k + |\{ i \ | α_i> 1,\ 1\leq i\leq k \}|$, depending on whether $n\not\in 4N+2$ or not, respectively. Furthermore, we show that, for a given order $n$, a divisor set $D$ with $|D|\leq k$ can always be found such that this bound is attained. Finally, we calculate the maximal diameter in the class of integral circulant graphs of a given order $n$ and cardinality of the divisor set $t\leq k$ and characterize all extremal graphs. We actually show that the maximal diameter can have the values $2t$, $2t+1$, $r(n)$ and $r(n)+1$ depending on the values of $t$ and $n$. This way we further improve the upper bound of Saxena, Severini and Shparlinski and we also characterize all graphs whose diameters are equal to $2|D|+1$, thus generalizing a result in that paper.
△ Less
Submitted 18 July, 2023;
originally announced July 2023.
-
Maximum values of the Sombor-index-like graph invariants of trees and connected graphs
Authors:
Milan Bašić
Abstract:
A set of novel vertex-degree-based invariants was introduced by Gutman, denoted by \newline $SO_1, SO_2, \ldots,SO_6$. These invariants were constructed through geometric reasoning based on a new graph invariant framework. Motivated by proposed open problems in [Z. Tang, Q. Li, H. Deng, \textit{Trees with Extremal Values of the Sombor-Index-Like Graph Invariants}, MATCH Commun. Math. Comput. Chem.…
▽ More
A set of novel vertex-degree-based invariants was introduced by Gutman, denoted by \newline $SO_1, SO_2, \ldots,SO_6$. These invariants were constructed through geometric reasoning based on a new graph invariant framework. Motivated by proposed open problems in [Z. Tang, Q. Li, H. Deng, \textit{Trees with Extremal Values of the Sombor-Index-Like Graph Invariants}, MATCH Commun. Math. Comput. Chem. \textbf{90} (2023) 203-222], we have found the maximum values of $SO_5$ and $SO_6$ in the set of molecular trees with a given number of vertices, respectively, and we have found the maximum value of $SO_5$ in a class of connected graphs.
△ Less
Submitted 7 April, 2024; v1 submitted 4 May, 2023;
originally announced May 2023.
-
Another estimation of Laplacian spectrum of the Kronecker product of graphs
Authors:
Milan Bašić,
Branko Arsić,
Zoran Obradović
Abstract:
The relationships between eigenvalues and eigenvectors of a product graph and those of its factor graphs have been known for the standard products, while characterization of Laplacian eigenvalues and eigenvectors of the Kronecker product of graphs using the Laplacian spectra and eigenvectors of the factors turned out to be quite challenging and has remained an open problem to date. Several approac…
▽ More
The relationships between eigenvalues and eigenvectors of a product graph and those of its factor graphs have been known for the standard products, while characterization of Laplacian eigenvalues and eigenvectors of the Kronecker product of graphs using the Laplacian spectra and eigenvectors of the factors turned out to be quite challenging and has remained an open problem to date. Several approaches for the estimation of Laplacian spectrum of the Kronecker product of graphs have been proposed in recent years. However, it turns out that not all the methods are practical to apply in network science models, particularly in the context of multilayer networks. Here we develop a practical and computationally efficient method to estimate Laplacian spectra of this graph product from spectral properties of their factor graphs which is more stable than the alternatives proposed in the literature. We emphasize that a median of the percentage errors of our estimated Laplacian spectrum almost coincides with the $x$-axis, unlike the alternatives which have sudden jumps at the beginning followed by a gradual decrease for the percentage errors. The percentage errors confined (confidence of the estimations) up to $\pm$10% for all considered approximations, depending on a graph density. Moreover, we theoretically prove that the percentage errors becomes smaller when the network grows or the edge density level increases. Additionally, some novel theoretical results considering the exact formulas and lower bounds related to the certain correlation coefficients corresponding to the estimated eigenvectors are presented.
△ Less
Submitted 4 February, 2021;
originally announced February 2021.
-
Further results on structured regression for multi-scale networks
Authors:
Milan Bašić,
Branko Arsić,
Zoran Obradović
Abstract:
Gaussian Conditional Random Fields (GCRF), as a structured regression model, is designed to achieve higher regression accuracy than unstructured predictors at the expense of execution time, taking into account the objects similarities and the outputs of unstructured predictors simultaneously. As most structural models, the GCRF model does not scale well with large networks. One of the approaches c…
▽ More
Gaussian Conditional Random Fields (GCRF), as a structured regression model, is designed to achieve higher regression accuracy than unstructured predictors at the expense of execution time, taking into account the objects similarities and the outputs of unstructured predictors simultaneously. As most structural models, the GCRF model does not scale well with large networks. One of the approaches consists of performing calculations on factor graphs (if it is possible) rather than on the full graph, which is more computationally efficient. The Kronecker product of the graphs appears to be a natural choice for a graph decomposition. However, this idea is not straightforwardly applicable for GCRF, since characterizing a Laplacian spectrum of the Kronecker product of graphs, which GCRF is based on, from spectra of its factor graphs has remained an open problem. In this paper we apply new estimations for the Laplacian eigenvalues and eigenvectors, and achieve high prediction accuracy of the proposed models, while the computational complexity of the models, compared to the original GCRF model, is improved from $O(n_{1}^{3}n_{2}^{3})$ to $O(n_{1}^{3} + n_{2}^{3})$. Furthermore, we study the GCRF model with a non-Kronecker graph, where the model consists of finding the nearest Kronecker product of graph for an initial graph. Although the proposed models are more complex, they achieve high prediction accuracy too, while the execution time is still much better compare to the original GCRF model. The effectiveness of the proposed models is characterized on three types of random networks where the proposed models were consistently away more accurate than the previously presented GCRF model for multiscale networks [Jesse Glass and Zoran Obradovic. Structured regression on multiscale networks. IEEE Intelligent Systems, 32(2):23-30, 2017.].
△ Less
Submitted 2 September, 2019;
originally announced September 2019.
-
Path matrix and path energy of graphs
Authors:
Aleksandar Ilic,
Milan Basic
Abstract:
Given a graph $G$, we associate a path matrix $P$ whose $(i, j)$ entry represents the maximum number of vertex disjoint paths between the vertices $i$ and $j$, with zeros on the main diagonal. In this note, we resolve four conjectures from [M. M. Shikare, P. P. Malavadkar, S. C. Patekar, I. Gutman, \emph{On Path Eigenvalues and Path Energy of Graphs}, MATCH Commun. Math. Comput. Chem. {\bf 79} (20…
▽ More
Given a graph $G$, we associate a path matrix $P$ whose $(i, j)$ entry represents the maximum number of vertex disjoint paths between the vertices $i$ and $j$, with zeros on the main diagonal. In this note, we resolve four conjectures from [M. M. Shikare, P. P. Malavadkar, S. C. Patekar, I. Gutman, \emph{On Path Eigenvalues and Path Energy of Graphs}, MATCH Commun. Math. Comput. Chem. {\bf 79} (2018), 387--398.] on the path energy of graphs and finally present efficient $O(|E| |V|^3)$ algorithm for computing the path matrix used for verifying computational results.
△ Less
Submitted 18 February, 2019; v1 submitted 11 October, 2018;
originally announced October 2018.
-
Characterization of circulant graphs having perfect state transfer
Authors:
Milan Bašić
Abstract:
In this paper we answer the question of when circulant quantum spin networks with nearest-neighbor couplings can give perfect state transfer. The network is described by a circulant graph $G$, which is characterized by its circulant adjacency matrix $A$. Formally, we say that there exists a {\it perfect state transfer} (PST) between vertices $a,b\in V(G)$ if $|F(τ)_{ab}|=1$, for some positive real…
▽ More
In this paper we answer the question of when circulant quantum spin networks with nearest-neighbor couplings can give perfect state transfer. The network is described by a circulant graph $G$, which is characterized by its circulant adjacency matrix $A$. Formally, we say that there exists a {\it perfect state transfer} (PST) between vertices $a,b\in V(G)$ if $|F(τ)_{ab}|=1$, for some positive real number $τ$, where $F(t)=\exp(ıAt)$. Saxena, Severini and Shparlinski ({\it International Journal of Quantum Information} 5 (2007), 417--430) proved that $|F(τ)_{aa}|=1$ for some $a\in V(G)$ and $τ\in \R^+$ if and only if all eigenvalues of $G$ are integer (that is, the graph is integral). The integral circulant graph $\ICG_n (D)$ has the vertex set $Z_n = \{0, 1, 2, ..., n - 1\}$ and vertices $a$ and $b$ are adjacent if $\gcd(a-b,n)\in D$, where $D \subseteq \{d : d \mid n,\ 1\leq d<n\}$. These graphs are highly symmetric and have important applications in chemical graph theory. We show that $\ICG_n (D)$ has PST if and only if $n\in 4\N$ and $D=\widetilde{D_3}\cup D_2\cup 2D_2\cup 4D_2\cup \{n/2^a\}$, where $\widetilde{D_3}=\{d\in D\ |\ n/d\in 8\N\}$, $D_2= \{d\in D\ |\ n/d\in 8\N+4\}\setminus \{n/4\}$ and $a\in\{1,2\}$. We have thus answered the question of complete characterization of perfect state transfer in integral circulant graphs raised in {\it Quantum Information and Computation}, Vol. 10, No. 3&4 (2010) 0325--0342 by Angeles-Canul {\it et al.} Furthermore, we also calculate perfect quantum communication distance (distance between vertices where PST occurs) and describe the spectra of integral circulant graphs having PST. We conclude by giving a closed form expression calculating the number of integral circulant graphs of a given order having PST.
△ Less
Submitted 10 April, 2011;
originally announced April 2011.
-
Which weighted circulant networks have perfect state transfer?
Authors:
Milan Bašić
Abstract:
The question of perfect state transfer existence in quantum spin networks based on weighted graphs has been recently presented by many authors. We give a simple condition for characterizing weighted circulant graphs allowing perfect state transfer in terms of their eigenvalues. This is done by extending the results about quantum periodicity existence in the networks obtained by Saxena, Severini an…
▽ More
The question of perfect state transfer existence in quantum spin networks based on weighted graphs has been recently presented by many authors. We give a simple condition for characterizing weighted circulant graphs allowing perfect state transfer in terms of their eigenvalues. This is done by extending the results about quantum periodicity existence in the networks obtained by Saxena, Severini and Shparlinski and characterizing integral graphs among weighted circulant graphs. Finally, classes of weighted circulant graphs supporting perfect state transfer are found. These classes completely cover the class of circulant graphs having perfect state transfer in the unweighted case. In fact, we show that there exists an weighted integral circulant graph with $n$ vertices having perfect state transfer if and only if $n$ is even. Moreover we prove the non-existence of perfect state transfer for several other classes of weighted integral circulant graphs of even order.
△ Less
Submitted 10 April, 2011;
originally announced April 2011.
-
Bisimulations for fuzzy automata
Authors:
Miroslav Ćirić,
Jelena Ignjatović,
Nada Damljanović,
Milan Bašić
Abstract:
Bisimulations have been widely used in many areas of computer science to model equivalence between various systems, and to reduce the number of states of these systems, whereas uniform fuzzy relations have recently been introduced as a means to model the fuzzy equivalence between elements of two possible different sets. Here we use the conjunction of these two concepts as a powerful tool in the st…
▽ More
Bisimulations have been widely used in many areas of computer science to model equivalence between various systems, and to reduce the number of states of these systems, whereas uniform fuzzy relations have recently been introduced as a means to model the fuzzy equivalence between elements of two possible different sets. Here we use the conjunction of these two concepts as a powerful tool in the study of equivalence between fuzzy automata. We prove that a uniform fuzzy relation between fuzzy automata $\cal A$ and $\cal B$ is a forward bisimulation if and only if its kernel and co-kernel are forward bisimulation fuzzy equivalences on $\cal A$ and $\cal B$ and there is a special isomorphism between factor fuzzy automata with respect to these fuzzy equivalences. As a consequence we get that fuzzy automata $\cal A$ and $\cal B$ are UFB-equivalent, i.e., there is a uniform forward bisimulation between them, if and only if there is a special isomorphism between the factor fuzzy automata of $\cal A$ and $\cal B$ with respect to their greatest forward bisimulation fuzzy equivalences. This result reduces the problem of testing UFB-equivalence to the problem of testing isomorphism of fuzzy automata, which is closely related to the well-known graph isomorphism problem. We prove some similar results for backward-forward bisimulations, and we point to fundamental differences. Because of the duality with the studied concepts, backward and forward-backward bisimulations are not considered separately. Finally, we give a comprehensive overview of various concepts on deterministic, nondeterministic, fuzzy, and weighted automata, which are related to bisimulations.
△ Less
Submitted 5 May, 2011; v1 submitted 26 February, 2011;
originally announced February 2011.
-
Nondeterministic automata: equivalence, bisimulations, and uniform relations
Authors:
Miroslav Ćirić,
Jelena Ignjatović,
Milan Bašić,
Ivana Jančić
Abstract:
In this paper we study the equivalence of nondeterministic automata pairing the concept of a bisimulation with the recently introduced concept of a uniform relation. In this symbiosis, uniform relations serve as equivalence relations which relate states of two possibly different nondeterministic automata, and bisimulations ensure compatibility with the transitions, initial and terminal states of t…
▽ More
In this paper we study the equivalence of nondeterministic automata pairing the concept of a bisimulation with the recently introduced concept of a uniform relation. In this symbiosis, uniform relations serve as equivalence relations which relate states of two possibly different nondeterministic automata, and bisimulations ensure compatibility with the transitions, initial and terminal states of these automata. We define six types of bisimulations, but due to the duality we discuss three of them: forward, backward-forward, and weak forward bisimulations. For each od these three types of bisimulations we provide a procedure which decides whether there is a bisimulation of this type between two automata, and when it exists, the same procedure computes the greatest one. We also show that there is a uniform forward bisimulation between two automata if and only if the factor automata with respect to the greatest forward bisimulation equivalences on these automata are isomorphic. We prove a similar theorem for weak forward bisimulations, using the concept of a weak forward isomorphism instead of an isomorphism. We also give examples that explain the relationships between the considered types of bisimulations.
△ Less
Submitted 26 February, 2011;
originally announced February 2011.
