Intelligent Agents

Intelligent Agents - Unit 2: Introducing First Order Logic

Overview

As per the course website, "During this week we have the first seminar for the module and will focus on discussing the fundamental elements of first order or predicate logic. It will involve students engaging with this underpinning mathematical notation in order to provide a greater understanding of some of the key elements of intelligent agents and wider topics."

My Reflection

Overall Reflection

This unit covered First Order Logic (FOL), which we previously covered in the Knowledge Representation and Reasoning (KRR) module, especially starting from Unit 3, so it was a good chance to connect the two modules as we move forward.

This time, the topic was tackled in the context of intelligent agents, the topic of this module. The topic was covered by a introductory seminar, in addition to reading Chapter 8 from Artificial Intelligence: A Modern Approach (4th. ed.) by Russel and Norvig (2021).

Additionally, we were expected to respond to two peers as part of the first collaborative discussion of the module, started in Unit 1.

Regarding the preparation for the group project, I have not received a reply for the group mates, so I sent another email as a kind reminder.

Readings Reflection

Like the entirety of the book, Chapter 8 from Artificial Intelligence: A Modern Approach (4th. ed.) by Russel and Norvig (2021) provided a detailed explanation of what FOL is, why it was developed, how it is constructed, in addition to how it relates to propositional logic, and more, interestingly, natural language.

Below are some quotes that I specially liked or found though-provoking:

"There is a long tradition in linguistics and the philosophy of language that views natural language as a declarative knowledge representation language."
"Clearly, one could not store a sentence such as “Look!” in a knowledge base and expect to recover its meaning without also storing a representation of the context—which raises the question of how the context itself can be represented."
"The famous Sapir–Whorf hypothesis (Whorf, 1956) claims that our understanding of the world is strongly influenced by the language we speak."
"More interesting is the case in which a concept is completely absent in a language. Speakers of the Australian aboriginal language Guugu Yimithirr have no words for relative (or ego centric) directions, such as front, back, right, or left. Instead they use absolute directions, saying, for example, the equivalent of “I have a pain in my north arm.” This difference in language makes a difference in behavior: Guugu Yimithirr speakers are better at navigating in open terrain, while English speakers are better at placing the fork to the right of the plate."
"The language of first-order logic, whose syntax and semantics we define in the next section, is built around objects and relations. It has been important to mathematics, philosophy, and artificial intelligence precisely because those fields—and indeed, much of everyday human existence—can be usefully thought of as dealing with objects and the relations among them."
"The primary difference between propositional and first-order logic lies in the ontological commitment made by each language—that is, what it assumes about the nature of reality."
"Knowledge representation languages should be declarative, compositional, expressive, context independent, and unambiguous."
"Both propositional logic and first-order logic share a difficulty in representing vague propositions. This difficulty limits their applicability in domains that require personal judgments, like politics or cuisine."

Reference

Russell, S. and Norvig, P. (2021) ‘First Order Logic’, in Artificial Intelligence: A Modern Approach - 4th. ed.. Harlow, UK: Pearson, pp. 269–397.

Artefacts

Collaborative Learning Discussion 1

The collaborative discussion we started in Unit 1 is about what has led to the rise of agent-based systems and what benefits organisations are gaining out of it. This week, it was the time to repond to at least two posts from the module peers.

Peer Response 1

The first post I responded to was from Matthew, who mainly attributed the rise of agents to the distributed nature of business and static nature of traditional software, and acknowledged the main benefit of agents for business is their ability to find solutions "more quickly" and lower cost. Below is my repsonse.

My Response:

Hi Matthew,

I appreciate how your post is clear and straight to the point, framing the topic at hand in two concise questions.

For the first topic - as you framed it - I would add internal technological factors of related technologies advancement (Priyadarshi, 2025), whether on the side of infrastructure, like the overall increase computational capabilities, or the advent of natural language processing (NLP) and large language models (LLMs). I would argue that interaction with the machine through natural language due to the advancement in latter areas made agents and agent-based systems (ABS) more accessible and, therefore, more adoptable, and gave them one their main traits that Wooldridge (2009) asserted in the first place, which them being 'socially able'. Furthermore, I would add that ABS has special glamour around its capability of scaling, connecting and communicating with other agents and systems. This area still yields great promises as it is still a wide area for live research and continuous development (Joshi, 2025).

For the second topic, I agree with you and emphasise that the future of ABS, after all, would still be hinged on whether they would be able to increase profits and lower costs, through various ways, including rapid problem-detection and decision making, reducing the 'bottleneck' for the need of human intervention like the traditional software architecture, as well as reducing the need for wide labour with diversified set of skills (Joshi, 2025). However, I would prefer to be suspicious on how would business frame the limits between reducing the need for human intervention, and how much delegation of decision making to machines would actually prove to be useful and sensible.

As much it is exciting to witness and discuss the rise and proliferation of ABS, as much it raises a lot of suspicions and second thoughts, probably like the rise of any new technology.

Reference List

Joshi, S. (2025) ‘Review of Autonomous and Collaborative Agentic AI and Multi-Agent Systems for Enterprise Applications’, International Journal of Innovative Research in Engineering and Management (IJIREM), 12(3), pp. 65–76. Available at: https://doi.org/10.55524/ijirem.2025.12.3.9.

Priyadarshi, M. (2025) ‘Autonomous AI Agents Transforming Enterprise Operations: From Static Automation to Intelligent Decision-Making Systems’, Sarcouncil Journal of Multidisciplinary, 5(7), pp. 863–870. Available at: https://doi.org/10.5281/zenodo.16408261.

Wooldridge, M. (2009) ‘Intelligent Angents’, in An Introduction to MultiAgent Systems - Second Edition. West Sussex, United Kingdom: John Wiley & Sons, pp. 21–48.

Peer Response 2

The second post that I reponded to was from Ioanna, who pointed out the factors of complexity and agents adaptability and scalability. However, what was more interesting for me is that she pointed out unpredictability and instability as limitations, espcially when agent systems are scaled to multi-agent systems. This made me think about what I found out in the KRR module on how ontologies can act as 'guardrails' for large language models (LLMs). Hence, this was my response.

My Response:

Hi Ioanna,

Thanks for your well-structured post. I agree with you that agent-based systems (ABS) offer a significant shift when compared to traditional software architecture, especially regarding ABS reactiveness, proactiveness, social ability and, most importantly in the context of this response, autonomy, as stressed by Wooldridge (2009).

What I found most interesting about your post is the mention of ABS limitations in your last paragraph. So far, I have not thought about the unpredictability and instability of ABS, especially multi-agent systems (MAS) and autonomous systems in general. Inspired by our last module on Knowledge Representation and Reasoning (KRR), and also the second unit of this module on First Order Logic (FOL), I thought to look for resources on whether logic formalisation and ontologies can help as 'guardrails' for MAS, as they can be for large language models (LLMs) (Magana and Monti, 2025).

The result of my research is that I found accumulative body of research on the exact topic of adopting logic formalisation and ontologies for 'safer' and reliable MAS. For example, Zaki et al. (2021) propose a modelling paradigm for self-certification of autonomous systems, based on ontologies enriched with "expressive semantic relationships and modality constraints to relate finite state automatons together". They propose a system that is not just designed to not derail, but also of which derailment can be diagnosed and prognosed.

Felicíssimo et al. (2005) also have a similar proposal that uses ontologies to define regulations over roles in open MAS, with five main concepts that are Role, Norm, Penalty, Action and Place. They claim that their proposed ontology's structure "provides a semantic support for agents to base their behaviour according to norms and to reason about action selection." Applying on a specific domain of knowledge and a specific case scenario, Sperotto, Belchior and de Aguiar (2019), developed an ontology to support MAS designed to operate in the legal domain, governing those MAS with and constraints, placing ontologies as "a fundamental piece" in the middleware of the system.

Thank you for shedding light on such a critical aspect. I hope that the example research papers that I mentioned here answer part of your question, and I hope we would be able to cover this further in detail later in the module.

Reference List

Felicíssimo, C. et al. (2005) ‘Normative Ontologies to Define Regulations over Roles in Open Multi- Agent Systems’, in AAAI Fall Symposium Series. - AAAI 2005 Fall Symposium on Roles, an Interdisciplinary Perspective, Menlo Park, California, USA: AAAI Press. Available at: https://cdn.aaai.org/Symposia/Fall/2005/FS-05-08/FS05-08-011.pdf (Accessed: 6 February 2026).

Magana, I. and Monti, M. (2025) Enhancing Large Language Models through Neuro-Symbolic Integration and Ontological Reasoning, arXiv.org. Available at: https://arxiv.org/abs/2504.07640v1 (Accessed: 23 December 2025).

Sperotto, F.A., Belchior, M. and de Aguiar, M.S. (2019) ‘Ontology-Based Legal System in Multi-agents Systems’, Lecture Notes in Computer Science, pp. 507–521. Available at: https://doi.org/10.1007/978-3-030-33749-0_41.

Wooldridge, M. (2009) ‘Intelligent Angents’, in An Introduction to MultiAgent Systems - Second Edition. West Sussex, United Kingdom: John Wiley & Sons, pp. 21–48.

Zaki, O. et al. (2021) ‘Reliability and Safety of Autonomous Systems Based on Semantic Modelling for Self-Certification’, Robotics, 10(1), p. 10. Available at: https://doi.org/10.3390/robotics10010010.