Joshua Taylor

Artificial Intelligence, Logic, Programming, Semantic Web, Interoperability, Cybersecurity

Education

Experience

Semantic Technologies Software Engineer, 5 Sep 2023–present

CACI International, Inc.
Rome, NY

Architect, design, and develop knowledge-driven applications using semantic technologies (including RDF, OWL, SPARQL) in support of national defense, AI/ML research, and company needs.

Chief Scientist, 2022–15 Aug 2023

Siege Technologies
Rome, NY

Work alongside Siege’s Senior Leadership Team to provide technical guidance for program execution, business development, and internal research and development. Primary author of Siege’s technical blog, Cyber Under Siege. Continued development of the Cyber Quantification Framework, and sustained growth of Siege’s semantic technologies group.

Principal Software Engineer, 2020–2022

Siege Technologies
Rome, NY

Led development of Siege’s internal research and development extensions to the Cyber Quantification Framework for commercialization and a broader range of virtualization platforms. Continued building Siege’s semantic technologies group, bringing in sufficient work to support a team of engineers. Continued building customer and industry relationships.

Senior Software Engineer, 2014–2020

Siege Technologies
Rome, NY

A primary developer of Siege’s Cyber Quantification Framework (CQF) for designing and executing cyber-experiments, data collection, analysis, and AI/ML-based prediction of software behavior. Also participated in reverse engineering efforts. Initiated Siege’s ventures into semantic technologies in support of defense applications. Participated in business development, customer interaction, and collaboration with industry partners.

Research Engineer, 2011–2014

Assured Information Security
Rome, NY

Worked in AIS’s Decision Sciences and Analytics group, contributing to a variety of efforts including semantic behavioral analysis of malware and cyber-environment modeling for situational awareness. Mentored other developers in semantic technologies (RDF, OWL, rule-based reasoning). Active in business development and proposals. Executed internal research and development efforts.

Graduate Research Assistant, 2008–2011

Department of Computer Science
Department of Cognitive Science
Rensselaer Polytechnic Institute
Troy, NY, USA

Continued development of Slate in support of the intelligence community as well as for pedagogical use in Rensselaer’s Introduction to Logic and Intermediate Logic courses. Translations between logical systems enabled students to use automated reasoners in one logic as oracles within others (for instance, using an automated theorem prover for first order logic to solve problems in modal logic using Kripke frame semantics).

Translations between logics led to my PhD dissertation topic, Fluid Logics, an approach to logic-based interoperability and proof translation rooted in category theory.

Gained significant experience in proposal writing, especially for government and defense programs, as well as proper documentation procedures for the same, including adherance to technical report guidelines.

Graduate Research Assistant, 2005–2007

Department of Computer Science
Department of Cognitive Science
Rensselaer Polytechnic Institute
Troy, NY, USA

A primary developer of Slate, a system for formal and informal argument mapping, for use by professionals and students in the domains of intelligence analysis and mathematics.

Graduate research in provability based interoperability culminated in my MS thesis, Provability-Based Semantic Interoperability Between Knowledgebases and Databases via Translation Graphs. Worked with RAIR Lab colleagues and teams from IBM, Stanford University, Pacific Northwest National Laboratories, and Oculus Inc. to achieve semantic interoperability between software systems developed for intelligence analysis, including Slate.

Graduate Teaching Asssistant, Fall 2006

Department of Computer Science
Rensselaer Polytechnic Institute
Troy, NY, USA

Teaching assistant for CSCI 4150, Artificial Intelligence in Fall 2006. Graded student assignments, examinations, final projects, and term papers. Held office hours. Course covered techniques in AI including: search, game theory, logic, knowledge representation, and basic programming in Scheme and Prolog.

Co-lecturer for Introduction to Logic, Fall 2005

Department of Cognitive Science
Rensselaer Polytechnic Institute
Troy, NY, USA

Instructed students in the use of Slate in PHIL 2140, Introduction to Logic. Students learned to construct and check arguments and proofs in formal logic using Slate.

Undergraduate Research Assistant, 2003–2005

Department of Cognitive Science
Rensselaer Polytechnic Institute
Troy, NY, USA

Started development of Slate.

Undergraduate Teaching Assistant, 2002–2005

Department of Cognitive Science
Rensselaer Polytechnic Institute
Troy, NY, USA

Teaching assistant for PHIL 2140, Introduction to Logic in Fall 2002, 2003, and 2004. Graded student assignments and examinations. Held office hours. Course covered traditional propositional, predicate, and first-order logics, and introduced visual formalisms such as Peirce's existential graphs, and the RAIR Lab’s Slate.

Teaching assistant for PHIL 4963, Intermediate Logic in Spring 2005. Graded student assignments, examinations, and final projects. Held office hours. Course covered axiom systems, modal logics, soundness and completeness theorems for proof systems, and axiom independence proofs.

Undergraduate Teaching Assistant, 2002–2003

Department of Computer Science
Rensselaer Polytechnic Institute
Troy, NY, USA

Teaching assistant for CSCI 1190, Beginning C Programming for Engineers in Fall 2003. Graded student assignments and examinations. Held office hours by appointment. Present during class to aid students with in class assignments.

Teaching assistant for CSCI 2300, Data Structures and Algorithms in Spring 2002. Helped students during lab/recitation sessions with programming assignments.

Skills

Scientific Software

Cyber Quantification Framework (CQF)

Siege’s flagship quantification software for experiment execution, data collection, analysis, and artificial intelligence/machine learning (AI/ML) based prediction. Developed under government and in-house funding for over ten years.

Slate

Developed at Rensselaer Polytechnic Institute, Slate, a System for Logic and Theorem Extract, was developed to assist members of the intelligence community (IC) and students of logic, mathematics, and related disciplines. Slate was also used as courseware within RPI’s introduction and intermediate logic classes, and integrated with the Logic: A Modern Approach.

Early versions of Slate were developed in Macintosh Common Lisp (MCL) using OpenGL bindings. Later versions of Slate were developed under LispWorks, using the CAPI UI layer.

Publications

Taylor, J., Zaffarano, K., Koller, B., Bancroft, C., & Syversen, J. (2016, October). Automated effectiveness evaluation of moving target defenses: Metrics for missions and attacks. In Proceedings of the 2016 ACM Workshop on Moving Target Defense (pp. 129-134).

In this paper, we describe the results of several experiments designed to test two dynamic network moving target defenses against a propagating data exfiltration attack. We designed a collection of metrics to assess the costs to mission activities and the benefits in the face of attacks and evaluated the impacts of the moving target defenses in both areas. Experiments leveraged Siege’s Cyber Quantification Framework to automatically provision the networks used in the experiment, install the two moving target defenses, collect data, and analyze the results. We identify areas in which the costs and benefits of the two moving target defenses differ, and note some of their unique performance characteristics.

Kara Zaffarano, Joshua Taylor, Samuel Hamilton, A Quantitative Framework for Moving Target Defense Effectiveness Evaluation. In Proc. MTD’15, Second ACM Workshop on Moving Target Defense. October 2015. doi:10.1145/2808475.2808476.

Static defense has proven to be a brittle mechanism for defending against cyber attack. Despite this, proactive defensive measures have not been widely deployed. This is because flexible proactive defensive measures such as Moving Target Defense (MTD) have as much potential to interfere with a network’s ability to support the mission as they do to defend the network. In this paper we introduce an approach to defining and measuring MTD effects applied in a network environment to help guide MTD deployment decisions that successfully balance the potential security benefits of MTD deployment against the potential productivity costs.

N. S. Govindarajalulu, S. Bringsjord, J. Taylor, Proof verification and proof discovery for relativity{.title}. Synthese, April 2014.

The vision of machines autonomously carrying out substantive conjecture generation, theorem discovery, proof discovery, and proof verification in mathematics and the natural sciences has a long history that reaches back before the development of automatic systems designed for such processes. While there has been considerable progress in proof verification in the formal sciences, for instance the Mizar project’ and the four-color theorem, now machine verified, there has been scant such work carried out in the realm of the natural sciences—until recently. The delay in the case of the natural sciences can be attributed to both a lack of formal analysis of the so-called “theories” in such sciences, and the lack of sufficient progress in automated theorem proving. While the lack of analysis is probably due to an inclination toward informality and empiricism on the part of nearly all of the relevant scientists, the lack of progress is to be expected, given the computational hardness of automated theorem proving; after all, theoremhood in even first-order logic is Turing-undecidable. We give in the present short paper a compressed report on our building upon these formal theories using logic-based AI in order to achieve, in relativity, both machine proof discovery and proof verification, for theorems previously established by humans. Our report is intended to serve as a springboard to machine-produced results in the future that have not been obtained by humans.

J. Taylor, R. T. Hall. Software Analysis in the Semantic Web. In Proc. SPIE 8757, Cyber Sensing 2013. 2013. doi:10.1117/12.2016122.

Many approaches in software analysis, particularly dynamic malware analyis, benefit greatly from the use of linked data and other Semantic Web technology. In this paper, we describe AIS, Inc.’s Semantic Extractor (SemEx) component from the Malware Analysis and Attribution through Genetic Information (MAAGI) effort, funded under DARPA’s CyberGenome program. The SemEx generates OWL-based semantic models of high and low level behaviors in malware samples from system call traces generated by AIS’s introspective hypervisor, IntroVirt™. Within MAAGI, these semantic models were used by modules that cluster malware samples by functionality, and construct “genealogical” malware lineages. Herein, we describe the design, implementation, and use of the SemEx, as well as the C2DB, an OWL ontology used for representing software behavior and cyber-environments.

R. T. Hall, J. Taylor. A Framework for Network-Wide Semantic Event Correlation. In Proc. SPIE 8757, Cyber Sensing 2013. 2013. doi:10.1117/12.2016126.

An increasing need for situational awareness within network-deployed Systems Under Test has increased desire for frameworks that facilitate system-wide data correlation and analysis. Massive event streams are generated from heterogeneous sensors which require tedious manual analysis. We present a framework for sensor data integration and event correlation based on Linked Data principles, Semantic Web reasoning technology, complex event processing, and blackboard architectures. Sensor data are encoded as RDF models, then processed by complex event processing agents (which incorporate domain specific reasoners, as well as general purpose Semantic Web reasoning techniques). Agents can publish inferences on shared blackboards and generate new semantic events that are fed back into the system. We present AIS, Inc.’s Cyber Battlefield Training and Effectiveness Environment (CBTEE) to demonstrate use of the framework.

S. Bringsjord, J. Taylor. Introducing Divine-Command Robot Ethics. In P. Lin, K. Abney, & G. Bekey, eds, Robot Ethics: The Ethical and Social Implications of Robotics. MIT Press.

  1. ISBN 978-0-262-01666-7.

Perhaps it is generally agreed that robots on the battlefield, especially if they have lethal power, should be ethically regulated. But in what does such regulation consist? Presumably in the fact that all the significant actions performed by such a robot are in accordance with some ethical code. But then the question arises as to which code. One possibility, a narrow one, is that the code is a set of rules of engagement affirmed by some nation or group. Another possibility is that the code is a utilitarian one represented in computational deontic logic, as explained elsewhere by Bringsjord and colleagues. Another possibility is likewise based on computational logic, but with a logic that captures some other mainstream ethical theory (e.g., Kantian deontology, or Ross’ “right mix” direction). But there is another radically different possibility that hitherto has not arrived on the scene: viz., the controlling code could be viewed by the human as coming straight from God. There is some very rigorous work in ethics along this line, which is known as divine-command ethics. In a world in which human fighters and the general populations supporting them often see themselves as indeed championing God’s will in war, divine-command ethics is quite relevant to military robots. This chapter introduces divine-command ethics in the form of the computational logic LRT^*^, intended to eventually be suitable for regulating a real-world warfighting robot.

J. Taylor, S. Bringsjord, M. Clark. Getting Started with Slate. September 2010.

A. Shilliday, J. Taylor, M. Clark, S. Bringsjord. Provability-Based Semantic Interoperability for Information Sharing and Joint Reasoning. In L. Obrst, T. Janssen & W. Ceusters, eds, Ontologies and Semantic Technologies for Intelligence, Volume 213 of Frontiers in Artificial Intelligence and Applications, 109–128. IOS Press. 2010. ISBN 978-1-60750-580-8.

We describe provability-based semantic interoperability (PBSI), a framework transcending syntactic translation that enables robust, meaningful, knowledge exchange across diverse information systems. PBSI is achieved through translation graphs that capture complex ontological relationships, and through provability-based queries. We work through an example of automating an unmanned aerial vehicle by reasoning over information from a number of sources.

S. Bringsjord, M. Clark, J. Taylor. Sophisticated Knowledge Representation and Reasoning Requires Philosophy. In R. Hagengruber, ed., Philosophy’s Relevance in Information Science, in press. Papers from the conference Philosophy’s Relevance in Information Science.

Knowledge Representation and Reasoning (KR&R) is based on the idea that propositional content can be rigorously represented in formal languages long the province of logic, in such a way that these representations can be productively reasoned over by humans and machines; and that this reasoning can be used to produce knowledge-based systems (KBSs). As such, KR&R is a discipline conventionally regarded to range across parts of artificial intelligence (AI), computer science, and especially logic. This standard view of KR&R’s participating fields is correct — but dangerously incomplete. The view is incomplete because, as we explain herein, sophisticated KR&R must rely heavily upon philosophy. Encapsulated, the reason is actually quite straightforward: Sophisticated KR&R must include the representation of not only simple properties, but also concepts that are routine in the formal sciences (theoretical computer science, mathematics, logic, game theory, etc.), and everyday socio-cognitive concepts like mendacity, deception, betrayal, and evil. Because in KR&R the representation of such concepts must be rigorous in order to enable machine reasoning (e.g., machine-generated and machine-checked proofs that a is lying to b) over them, philosophy, devoted as it is in no small part to supplying analyses of such concepts, is a crucial partner in the overall enterprise. To put the point another way: When the knowledge to be represented is such as to require lengthy formulas in expressive formal languages for that representation, philosophy must be involved in the game. In addition, insofar as the advance of KR&R must allow formalisms and processes for representing and reasoning over visual propositional content, philosophy will be a key contributor into the future.

S. Bringsjord, J. Taylor, T. Housten, B. van Heuveln, K. Arkoudas, M. Clark, R. Wojtowicz. Piagetian roboethics via category theory: Moving beyond mere formal operations to engineer robots whose decisions are guaranteed to be ethically correct. In Proceedings of the 2009 IEEE International Conference on Robotics and Automation (ICRA2009), Workshop on Roboethics, IEEE Press. 2009.

This is an extended abstract, not a polished paper; an approach to, rather than the results of, sustained research and development in the area of roboethics is described herein. Encapsulated, the approach is to engineer ethically correct robots by giving them the capacity to reason over, rather than merely in, logical systems (where logical systems are used to formalize such things as ethical codes of conduct for warfighting robots). This is to be accomplished by taking seriously Piaget’s position that sophisticated human thinking exceeds even abstract processes carried out in a logical system, and by exploiting category theory to render in rigorous form, suitable for mechanization, structure-preserving mappings that Bringsjord, an avowed Piagetian, sees to be central in rigorous and rational human ethical decision-making.

S. Bringsjord, J. Taylor, A. Shilliday, M. Clark, and K. Arkoudas. Slate: An Argument-Centered Intelligent Assistant to Human Reasoners. In Proceedings of CMNA 08, 8th Workshop on Computational Models of Natural Argumentation, held in conjunction with ECAI 2008, 18th European Conference on Artificial Intelligence. University of Patras, Patras, Greece. July 21, 2008.

We describe Slate, a logic-based, robust interactive reasoning system that allows human “pilots” to harness an ensemble of intelligent agents in order to construct, test, and express various sorts of natural argumentation. Slate empowers students and professionals in the business of producing argumentation, e.g., mathematicians, logicians, intelligence analysts, designers and producers of standardized reasoning tests. We demonstrate Slate in several examples, describe some distinctive features of the system (e.g., reading and generating natural language, immunizing human reasoners from “logical illusions”), present Slate’s theoretical underpinnings, and note upcoming refinements.

S. Bringsjord, A. Shilliday, J. Taylor, D. Werner, M. Clark, E. Charpentier, and A. Bringsjord. Toward Logic-Based Cognitively Robust Synthetic Characters in Digital Environments. Artificial General Intelligence 2008 — Proceedings of the First AGI Conference. In the series of Frontiers in Artificial Intelligence and Applications, 171:87–98. IOS Press. 2008.

With respect to genuine cognitive faculties, present synthetic characters inhabiting online virtual worlds are, to say the least, completely impaired. Current methods aimed at the creation of “immersive” virtual worlds only avatars and NPCs the illusion of mentality and, as such, will ultimately fail. Like behaviorism, this doomed approach focuses only on the inputs and outputs of virtual characters and ignores the rich mental structures that are essential for any truly realistic social environment. While this “deceptive” tactic may be suitable so long as a human is in the driver’s seat compensating for the mental deficit, truly convincing autonomous synthetic characters must possess genuine mental states, which can only result from a formal theory of mind. We report here on our attempt to invent part of such a theory, one that will enable artificial agents to have and reason about the beliefs of others, resulting in characters that can predict and manipulate the behavior of even human players. Furthermore, we present the “embodiment” of our recent successes: Eddie, a four year old child in Second Life who can reason about his own beliefs to draw conclusions in a manner that matches human children his age.

J. Taylor, S. Bringsjord. Discovery Using Heterogeneous Combined Logics. In Papers from the 2008 AAAI Fall Symposium on Automated Scientific Discovery, AAAI Press, Menlo Park, CA, 31–32. Technical Report FS-08-03.

Research in hybrid logic systems and, later, description logics, has revealed a tradeoff between the expressivity of a logical formalism, and the complexity of reasoning within that formalism. This is why, for instance, tractable inference procedures are known for certain classes of description logics and for (some) formalisms underlying knowledge representation on the Semantic Web. … Recognizing the diversity of knowledge representation systems currently in existence, the different properties of proof calculi which may be employed over these systems, and the growing need to combine inferences made under multiple logical systems, we propose the development of formalisms to govern these interactions, and call this the study of combined logics.

A. Shilliday, J. Taylor, S. Bringsjord. Toward Automated Provability-Based Semantic Interoperability between Ontologies for the Intelligence Community. In Proceedings of the Second International Ontology for the Intelligence Community (OIC-2007), 66–72. CEUR Workshop Proceedings. November 28–29, 2007.

The need for interoperability is dire: Knowledge representation systems employ ontologies that use disparate formalisms to describe related domains; to be truly useful to the intelligence community, they must meaningfully share information. Ongoing research strives toward the holy grail of complete interoperability, but has been hindered by techniques that are specialized for particular ontologies, and that lack the expressivity needed to describe complex ontological relationships. In the sequel, we describe provability-based semantic interoperability, a means to surmount these hindrances; translation graphs, one of our key formalism for describing the complex relationships among arbitrary ontologies; and ways in which these techniques might be automated.

Taylor, J.; Shilliday, A.; Bringsjord, S.. (2007). “Provability-Based Semantic Interoperability Via Translation Graphs”. Advances in Conceptual Modeling – Foundations and Applications. Lecture Notes in Computer Science 4802. pp. 180–189. doi:10.1007/978-3-540-76292-8_21. ISBN 978-3-540-76291-1. This paper appeared in the 2007 International Workshop on Ontologies and Information Systems for the Semantic Web (ONISW).

Provability-based semantic interoperability (PBSI) is a kind of interoperability that transcends mere syntactic translation to allow for robust, meaningful information exchange across systems employing ontologies for which mappings or matchings may not exist, and which can be evaluated by provability-based (PB) queries. We introduce a system of translation graphs to formalize the relationships between diverse ontologies and knowledge representation and reasoning systems, and to automatically generate the translation axioms governing PB information exchange and inter-system reasoning. We demonstrate the use of translation graphs on a small number of simple systems to achieve interoperability.

S. Bringsjord, K. Arkoudas, D. Mukherjee, A. Shilliday, J. Taylor, M. Clark, and E. Bringsjord. The Multi-Mind Effect. In Proceedings of the 2007 International Conference on Artificial Intelligence (ICAI), 43–49. CSREA Press. June 25–28, 2007.

Courtesy of experiments carried out by such thinkers as Wason, Johnson-Laird, and Kahneman & Tversky, there is overwhelming empirical evidence that the vast majority of logically untrained humans are unable to reason in context-independent, normatively correct fashion. However, the multi-mind effect, which is predicted by our earlier success at teaching this kind of reasoning, and also by our general theory of human and machine reasoning, shows that while individual persons (with rare exceptions) are unable to solve problems that demand context-independent reasoning, groups of persons can often solve such problems.

S. Bringsjord, K. Arkoudas, M. Clark, A. Shilliday, J. Taylor, B. Schimanski, and Y. Yang. Reporting on Some Logic-Based Machine Reading Research. In Proceedings of the 2007 AAAI Spring Symposium on Machine Reading. 2007.

Much sponsored research in our lab either falls under or intersects with machine reading. In this short paper we give an encapsulated presentation of some of the research in question, leaving aside, for the most part, the considerable detailed technical information that underlies our work. Demonstrations of our technology will be provided at the symposium itself.

S. Bringsjord, A. Shilliday, J. Taylor, P. Bello, Y. Yang, and K. Arkoudas. Harnessing Intelligent Agent Technology to ‘Superteach’ Reasoning. International Journal of Technology in Teaching and Learning, 2(2):88–116. 2006.

After briefly explaining our ultimate educational goal with respect to reasoning (to “superteach” reasoning), and our theoretical foundation, we give an overview of some of our attempts to build and harness intelligent agents in order to reach this goal. We end with coverage of the Slate system, which inherits its power from lessons learned in connection with the engineering of its more primitive predecessors.

Bringsjord, S.; Kellett, O.; Shilliday, A.; Taylor, J.; Van Heuveln, B.; Yang, Y.; Baumes, J.; Ross, K. (2006). “A new Gödelian argument for hypercomputing minds based on the busy beaver problem”. Applied Mathematics and Computation 176(2):516–530. doi:10.1016/j.amc.2005.09.071.

Do human persons hypercompute? Or, as the doctrine of computationalism holds, are they information processors at or below the Turing Limit? If the former, given the essence of hypercomputation, persons must in some real way be capable of infinitary information processing. Using as a springboard Gödel's little-known assertion that the human mind has a power “converging to infinity”, and as an anchoring propblem Rado's Turing-uncomputable “busy-beaver” (or Σ) function, we present in this short paper a new argument that, in fact, human persons can hypercompute. The argument is intended to be a formidable, not conclusive: it brings Gödel's intuition to a greater level of precision, and places it within a sensible case against computationalism.

Community

Handbell Choir Director, 2017–2020

First Baptist Church
Rome, NY

Directed a mixed-age choir of handbell ringers of various musical backgrounds, aiming for three to four performances yearly during Sunday sevices. Responsible for selecting music, instructing ringers in handbell technique and general music education, scheduling and directing rehearsals.

Substitute Musician, 2012–present

First Baptist Church
Rome, NY

Substitute for church pianist/organist as needed, occasionally warming up and directing chancel choir in service music. Also provide piano accompaniment at nursing home services during First Baptist’s services (as part of the Rome Clergy Association’s rotation).

Diaconate, 2013–present

First Baptist Church
Rome, NY

Prepare and set up for various sevices as needed, typically monthly. Attend board meetings, and assist as able.

Sunday School Teacher, 2004–2005

Korean Presbyterian Church of Albany
7 Knox Drive
Schenectady, NY

Under direction of deacon responsible for junior and senior high youth group, prepared and taught weekly lessons for a group of five to fifteen students in grades 6–8.