CIS 4394 Agentic AI · Fall 2026 · Dr. Xinyu Fu
CIS 4394 Agentic AI · Fall 2026 · Dr. Xinyu Fu
Just enough about language models to see why they can’t act alone — then the loop that turns a text predictor into something that gets work done.
Input text goes in; the model outputs a probability distribution over possible next tokens; one gets sampled; repeat. Everything impressive an LLM does is built from that single operation at scale.
| Strong | Weak |
|---|---|
| Language: summarize, draft, translate, explain | Knowing what it doesn’t know — can be confidently wrong |
| Pattern recognition across huge context | Fresh or private information past its training data |
| Code and structured text | Math it can’t verify; precise multi-step state |
| Following instructions in context | Doing anything — it can only emit text |
The limits in the right column are exactly what tools, memory, and the loop exist to fix.
Model + Tools + Loop + Memory (Wiesinger, Marlow & Vuskovic, 2024). The research literature decomposes the same idea as profile, memory, planning, and action (Wang et al., 2023, arXiv:2308.11432).
The reasoning engine. Interprets observations, weighs options, proposes the next action.
The hands. Structured functions — search, calendar, database, code — that turn intent into effect.
The heartbeat. Reason → act → observe → repeat, until a stop condition is met.
The thread. State that persists across turns so the agent doesn’t start over each step.
One concrete episode: “Schedule a 30-minute meeting with Maya this week.” Step through it. Watch which node is active, what the tools return, how state accumulates — and where the loop stops.
The loop closes — until the stop condition is met or a human gate interrupts. An agent without a stop condition is a bug with a budget.
The academic survey literature describes agent architecture as four modules — the same idea in formal dress.
Who the agent is: role, goals, constraints, persona.
Short-term context + long-term store the agent reads and writes.
Decomposing goals, choosing next steps, revising on feedback.
Tool calls and outputs that change the environment.
Wang et al. (2024), A Survey on LLM-based Autonomous Agents, Frontiers of Computer Science 18(6), arXiv:2308.11432 · Yao et al. (2023), ReAct, ICLR, arXiv:2210.03629 — the observe→reason→act formulation.