Paulo Shakarian

• Ph.D.
• Associate Professor at Arizona State University

Machine learning models assume that training and test samples are drawn from the same distribution. As such, significant differences between training and test distributions often lead to degradations in performance. We introduce Multiple Distribution Shift -- Aerial (MDS-A) -- a collection of inter-related datasets of the same aerial domain that ar...

In this paper, we study the problem of visual question answering (VQA) where the image and query are represented by ASP programs that lack domain data. We provide an approach that is orthogonal and complementary to existing knowledge augmentation techniques where we abduce domain relationships of image constructs from past examples. After framing t...

Metacognition is the concept of reasoning about an agent's own internal processes, and it has recently received renewed attention with respect to artificial intelligence (AI) and, more specifically, machine learning systems. This paper reviews a hybrid-AI approach known as "error detecting and correcting rules" (EDCR) that allows for the learning o...

Bad actors in the maritime industry engage in illegal behaviors after disabling their vessel's automatic identification system (AIS) - which makes finding such vessels difficult for analysts. Machine learning approaches only succeed in identifying the locations of these ``dark vessels'' in the immediate future. This work leverages ideas from the li...

Predicting price spikes in critical metals such as Cobalt, Copper, Magnesium, and Nickel is crucial for mitigating economic risks associated with global trends like the energy transition and reshoring of manufacturing. While traditional models have focused on regression-based approaches, our work introduces a neurosymbolic ensemble framework that i...

In this paper, we study the problem of visual question answering (VQA) where the image and query are represented by ASP programs that lack domain data. We provide an approach that is orthogonal and complementary to existing knowledge augmentation techniques where we abduce domain relationships of image constructs from past examples. After framing t...

Recent advances in Hierarchical Multi-label Classification (HMC), particularly neurosymbolic-based approaches, have demonstrated improved consistency and accuracy by enforcing constraints on a neural model during training. However, such work assumes the existence of such constraints a-priori. In this paper, we relax this strong assumption and prese...

The ability to generate artificial human movement patterns while meeting location and time constraints is an important problem in the security community, particularly as it enables the study of the analog problem of detecting such patterns while maintaining privacy. We frame this problem as an instance of abduction guided by a novel parsimony funct...

Metacognition is the concept of reasoning about an agent's own internal processes and was originally introduced in the field of developmental psychology. In this position paper, we examine the concept of applying metacognition to artificial intelligence. We introduce a framework for understanding metacognitive artificial intelligence (AI) that we c...

Recent advances in experimental methods have enabled researchers to collect data on thousands of analytes simultaneously. This has led to correlational studies that associated molecular measurements with diseases such as Alzheimer's, Liver, and Gastric Cancer. However, the use of thousands of biomarkers selected from the analytes is not practical f...

Capturing the uncertain aspects in cyber threat analyses is an important part of a wide range of efforts, including diagnostics, threat evaluation, and preventing attacks. However, there has been insufficient research and development of modeling approaches that are able to correctly capture and handle such uncertainty. In this work, we present an a...

Classification of movement trajectories has many applications in transportation. Supervised neural models represent the current state-of-the-art. Recent security applications require this task to be rapidly employed in environments that may differ from the data used to train such models for which there is little training data. We provide a neuro-sy...

Error prediction in large language models often relies on domain-specific information. In this paper, we present measures for quantification of error in the response of a large language model based on the diversity of responses to a given prompt - hence independent of the underlying application. We describe how three such measures - based on entrop...

Neuro symbolic reasoning and learning is a topic that combines ideas from deep neural networks with symbolic reasoning and learning to overcome several significant technical hurdles such as explainability, modularity, verification, and the enforcement of constraints. While neuro symbolic ideas date back to the early 2000’s, there have been signific...

Deep learning based approaches have been used to address several problems of a sequential nature, whether using supervised learning to learn a model of a time-series system or using reinforcement learning to train an agent. However, this work has raised questions of generalizability, explainability, and verification. In this chapter, we review work...

In this chapter, we describe how a logic program can be learned from data in a neuro symbolic framework. Our focus is on the gradient-based method known as differentiable inductive logic programming (ILP), which combines concepts from ILP with a neural architecture to support gradient-based learning. Additionally, we also cover several other paradi...

In this chapter, we provide an overview of Logic Tensor Networks (LTNs, for short), a formalism that makes use of tensor embeddings—n-dimensional vector representations—of elements tied to a logical syntax, which has seen traction in NSR literature in the past few years. After briefly recalling Real Logic, the underlying language of LTNs, we discus...

Many recent neuro symbolic approaches rely on an underlying logical language. In this chapter, we provide a brief introduction to the basic concepts behind propositional logic and predicate calculus (first order logic). We cover syntax, semantics, satisfiability relationships, consistency and entailment, quantifiers and grounding, as well as how lo...

In this chapter, we explore a neuro symbolic approach to incorporate explicit knowledge about a domain while learning a model. Specifically, we use a differentiable extension of the declarative problem solving language ASP (Answer Set Programming) that is inspired by Logic Programming. In this chapter, we review basic concepts in ASP and describe a...

Various neuro symbolic approaches such as Logical Neural Networks, Logical Tensor Networks, differentiable ILP, and others rely on the use of several forms of real-valued logic and fuzzy operators. In this chapter, we review generalized annotated logic that encompasses the various logics used in neuro symbolic frameworks as well as the fuzzy operat...

Although these days neural networks and deep learning get equated with AI, there are many systems that combine neural reasoning and learning with symbolic reasoning and learning modules. In the previous chapters, we discussed some of these approaches with more focus on formalisms. In this chapter we point to some of the notable applications, elabor...

Logical Neural Networks (LNN) is a framework that assumes knowledge of a logic program a-priori and uses gradient descent to fit the logic program to training data via parameterized logical operators, resulting in fuzzy logic semantics. The framework has several desirable properties, namely the ability to support open world reasoning, omnidirection...

The SATNet framework is a neural architecture designed to learn instances of combinatorial problems by learning the set of logical constraints associated with an instance of the maximum satisfiability problem. This turns out to be quite powerful as SATNet is able to learn instances of a wide variety of combinatorial problems (including certain NP-h...

The growing popularity of neuro symbolic reasoning has led to the adoption of various forms of differentiable (i.e., fuzzy) first order logic. We introduce PyReason, a software framework based on generalized annotated logic that both captures the current cohort of differentiable logics and temporal extensions to support inference over finite period...

While deep neural networks have led to major advances in image recognition, language translation, data mining, and game playing, there are well-known limits to the paradigm such as lack of explainability, difficulty of incorporating prior knowledge, and modularity. Neuro symbolic hybrid systems have recently emerged as a straightforward way to exte...

We study the performance of a commercially available large language model (LLM) known as ChatGPT on math word problems (MWPs) from the dataset DRAW-1K. To our knowledge, this is the first independent evaluation of ChatGPT. We found that ChatGPT's performance changes dramatically based on the requirement to show its work, failing 20% of the time whe...

Reasoning about complex networks has in recent years become an important topic of study due to its many applications: the adoption of commercial products, spread of disease, the diffusion of an idea, etc. In this paper, we present the MANCaLog language, a formalism based on logic programming that satisfies a set of desiderata proposed in previous w...

Decision support tools are key components of intelligent sociotechnical systems, and their successful implementation faces a variety of challenges, including the multiplicity of information sources, heterogeneous format, and constant changes. Handling such challenges requires the ability to analyze and process inconsistent and incomplete informatio...

This article describes a scalar field topology (SFT)‐based methodology for the interactive characterization and analysis of hotspots for density fields defined on a regular grid. In contrast to the common approach of simply identifying hotspots as areas that exceed a chosen density threshold, SFT provides various data abstractions—such as the merge...

Malicious hackers utilize the World Wide Web to share knowledge. Analyzing the online communication of these threat actors can help reduce the risk of attacks. This book shifts attention from the defender environment to the attacker environment, offering a new security paradigm of 'proactive cyber threat intelligence' that allows defenders of compu...

In this book, we presented results of the efforts to detect “Pathogenic Social Media (PSM)” accounts who are responsible for manipulating public opinion and political events. There are many challenges in the area of PSM accounts detection. In Chaps. 3 and 4, standard and time-decay probabilistic causal metrics were proposed to distinguish PSM from...

The lack of sufficient labeled examples for devising and training sophisticated approaches to combat PSM accounts is still one of the foremost challenges facing social media firms. In contrast, unlabeled data is abundant and cheap to obtain thanks to the massive user-generated data produced on a daily basis. This chapter proposes a semi-supervised...

This chapter introduces a time-decay causality metric and incorporates it into a causal community detection-based algorithm to identify PSMs within a short time frame around their activity. The proposed algorithm is applied to groups of accounts sharing similar causality features and is followed by a classification algorithm to classify accounts as...

This chapter introduces an unsupervised causality-based framework built upon the causal inference presented in Chap. 2 using label propagation. The merit of this approach is that it identifies PSM users without using network structure, cascade path information, content and user’s information which are usually hard to obtain. Results on the ISIS-A d...

In this chapter, we present a feature-driven approach to detect PSM accounts in social media. Inspired by the literature, we set out to assess PSMs from three broad perspectives: (1) user-related information (e.g., user activity, profile characteristics), (2) source-related information (i.e., information linked via URLs shared by users) and (3) con...

Over the past years, political events and public opinion on the Web have been allegedly manipulated by “Pathogenic Social Media (PSM)” accounts dedicated to spreading disinformation and performing malicious activities. These accounts are often controlled by terrorist supporters, water armies, or fake news writers and hence can pose threats to socia...

In this chapter, we adopt the causal inference framework described previously along with graph-based metrics to distinguish PSMs from normal users within a short time of their activities. We propose both supervised and semi-supervised approaches without taking the network information and content into account. Results on the ISIS-A dataset demonstra...

Recently, there has been strong interest in measuring influence in online social networks. Different measures have been proposed to predict when individuals will adopt a new behavior, given the influence produced by their friends. In this article, we show that one can achieve significant improvement over these measures, extending them to consider a...

Malware reverse-engineering, specifically, identifying the tasks a given piece of malware was designed to perform (e.g., logging keystrokes, recording video, establishing remote access) is a largely human-driven process that is a difficult and time-consuming operation. In this chapter, we present an automated method to identify malware tasks using...

It is widely believed that one’s peers influence product adoption behaviors. This relationship has been linked to the number of signals a decision-maker receives in a social network. But it is unclear if these same principles hold when the “pattern” by which it receives these signals vary and when peer influence is directed towards choices which ar...

It is widely believed that one's peers influence product adoption behaviors. This relationship has been linked to the number of signals a decision-maker receives in a social network, but it is unclear if these same principles hold when the "pattern" by it receives these signals vary and when potential decisions have different utilities. To investig...

Although cybersecurity research has demonstrated that many of the recent cyberattacks targeting real-world organizations could have been avoided, proactively identifying and systematically understanding when and why those events are likely to occur is still challenging. It has earlier been shown that monitoring malicious hacker discussions about so...

Over the past few years, we have observed different media outlets' attempts to shift public opinion by framing information to support a narrative that facilitate their goals. Malicious users referred to as "pathogenic social media" (PSM) accounts are more likely to amplify this phenomena by spreading misinformation to viral proportions. Understandi...

With the rise in security breaches over the past few years, there has been an increasing need to mine insights from social media platforms to raise alerts of possible attacks in an attempt to defend conflict during competition. In this study, we attempt to build a framework that utilizes unconventional signals from the darkweb forums by leveraging...

With rise in security breaches over the past few years, there has been an increasing need to mine insights from social media platforms to raise alerts of possible attacks in an attempt to defend conflict during competition. In this study, we attempt to build a framework that utilizes unconventional signals from the darkweb forums by leveraging the...

Social media has enabled users and organizations to obtain information about technology usage like software usage and even security feature usage. However, on the dark side it has also allowed an adversary to potentially exploit the users in a manner to either obtain information from them or influence them towards decisions that might have maliciou...

It is widely believed that the adoption behavior of a decision-maker in a social network is related to the number of signals it receives from its peers in the social network. It is unclear if these same principles hold when the "pattern" by which they receive these signals vary and when potential decisions have different utilities. To investigate t...

Information diffusion mechanisms based on social influence models are mainly studied using likelihood of adoption when active neighbors expose a user to a message. The problem arises primarily from the fact that for the most part, this explicit information of who-exposed-whom among a group of active neighbors in a social network, before a susceptib...

Recent years have witnessed a surge of manipulation of public opinion and political events by malicious social media actors. These users are referred to as “Pathogenic Social Media (PSM)” accounts. PSMs are key users in spreading misinformation in social media to viral proportions. These accounts can be either controlled by real users or automated...

Pathogenic Social Media (PSM) accounts such as terrorist supporter accounts and fake news writers have the capability of spreading disinformation to viral proportions. Early detection of PSM accounts is crucial as they are likely to be key users to make malicious information "viral". In this paper, we adopt the causal inference framework along with...

The spread of harmful mis-information in social media is a pressing problem. We refer accounts that have the capability of spreading such information to viral proportions as "Pathogenic Social Media" accounts. These accounts include terrorist supporters accounts, water armies, and fake news writers. We introduce an unsupervised causality-based fram...

Network motifs are patterns of over-represented node interactions in a network which have been previously used as building blocks to understand various aspects of the social networks. In this paper, we use motif patterns to characterize the information diffusion process in social networks. We study the lifecycle of information cascades to understan...

A growing set of applications consider the process of network formation by using subgraphs as a tool for generating the network topology. One of the pressing research challenges is thus to be able to use these subgraphs to understand the network topology of information cascades which ultimately paves the way to theorize about how information spread...

Recent years have witnessed a surge of manipulation of public opinion and political events by malicious social media actors. These users are referred to as "Pathogenic Social Media (PSM)" accounts. PSMs are key users in spreading misinformation in social media to viral proportions. These accounts can be either controlled by real users or automated...

Network motifs are patterns of over-represented node interactions in a network which have been previously used as building blocks to understand various aspects of the social networks. In this paper, we use motif patterns to characterize the information diffusion process in social networks. We study the lifecycle of information cascades to understan...

Information diffusion mechanisms based on social influence models are mainly studied using likelihood of adoption when active neighbors expose a user to a message. The problem arises primarily from the fact that for the most part, this explicit information of who-exposed-whom among a group of active neighbors in a social network, before a susceptib...

The ease of use of the Internet has enabled violent extremists such as the Islamic State of Iraq and Syria (ISIS) to easily reach large audience, build personal relationships and increase recruitment. Social media are primarily based on the reports they receive from their own users to mitigate the problem. Despite efforts of social media in suspend...

Over the past years, political events and public opinion on the Web have been allegedly manipulated by accounts dedicated to spreading disinformation and performing malicious activities on social media. These accounts hereafter referred to as "Pathogenic Social Media (PSM)" accounts, are often controlled by terrorist supporters, water armies or fak...

The number of software vulnerabilities discovered and publicly disclosed is increasing every year; however, only a small fraction of these vulnerabilities are exploited in real-world attacks. With limitations on time and skilled resources, organizations often look at ways to identify threatened vulnerabilities for patch prioritization. In this chap...

Scientific work that leverages information about communities on the deep and dark web has opened up new angles in the field of security informatics. [...]

With rise in security breaches over the past few years, there has been an increasing need to mine insights from social media platforms to raise alerts of possible attacks in an attempt to defend conflict during competition. We use information from the darkweb forums by leveraging the reply network structure of user interactions with the goal of pre...

Recent incidents of data breaches call for organizations to proactively identify cyber attacks on their systems. Darkweb/Deepweb (D2web) forums and marketplaces provide environments where hackers anonymously discuss existing vulnerabilities and commercialize malicious software to exploit those vulnerabilities. These platforms offer security practit...

With the recent prevalence of darkweb/deepweb (D2web) sites specializing in the trade of exploit kits and malware, malicious actors have easy-access to a wide-range of tools that can empower their offensive capability. In this study, we apply concepts from causal reasoning, itemset mining, and logic programming on historical cryptocurrency-related...

Pathogenic Social Media (PSM) accounts such as terrorist supporters exploit large communities of supporters for conducting attacks on social media. Early detection of these accounts is crucial as they are high likely to be key users in making a harmful message "viral". In this paper, we make the first attempt on utilizing causal inference to identi...

Most social network sites allow users to reshare a piece of information posted by a user. As time progresses, the cascade of reshares grows, eventually saturating after a certain time period. While previous studies have focused heavily on one aspect of the cascade phenomenon, specifically predicting when the cascade would go viral, in this paper, w...

Cyber adversaries employ a variety of malware and exploit to attack computer systems. Despite the prevalence of markets for malware and exploit kits, existing paradigms that model such cyber-adversarial behaviour do not account for sequential application or “chaining” of attacks, that take advantage of the complex and interdependent nature of explo...

Software vulnerabilities can expose computer systems to attacks by malicious actors. With the number of vulnerabilities discovered in the recent years surging, creating timely patches for every vulnerability is not always feasible. At the same time, not every vulnerability will be exploited by attackers; hence, prioritizing vulnerabilities by asses...

Cyber attribution is a difficult problem, and conducting attribution research is made even more difficult by a lack of data with ground truth. In this chapter, we describe a game-based framework (Capture-the-Flag) to produce cyber attribution data with deception. We discuss the motivation and the design of the contest and the framework to record da...

In cyber attribution, knowledge bases consisting of all the available information for a specific domain, along with the current state of affairs, will typically contain contradictory data coming from different sources, as well as data with varying degrees of uncertainty attached. In this chapter, we propose a probabilistic structured argumentation...