Artificial Intelligence is evolving rapidly, especially in the fields of autonomous agents and intelligent search systems. As part of my AI coursework, I explored two recent research papers that connect strongly with the concepts we studied in class such as intelligent agents, search algorithms, optimization, and decision making.
The two papers I selected were:
- “The Rise of Agentic AI: A Review of Definitions, Frameworks, and Challenges” (2025)
- “Research on the A* Algorithm Based on Adaptive Weights” (2025)
To better understand these papers, I used Google NotebookLM to summarize difficult sections, explain technical terminology, and compare the papers with classroom concepts. This experience helped me realize how theoretical AI concepts are directly connected to real-world intelligent systems.
Paper 1 — The Rise of Agentic AI
The first paper focuses on Agentic AI, which refers to AI systems capable of acting autonomously with minimal human supervision. Unlike traditional AI systems that simply respond to commands, agentic AI systems can plan, reason, remember information, and make independent decisions.
The paper explains how modern AI agents combine several important components:
- Memory systems
- Planning modules
- Decision-making strategies
- Tool usage
- Multi-agent collaboration
This directly relates to the concept of intelligent agents discussed in our AI course. In class, we studied different types of agents such as reflex agents, goal-based agents, utility-based agents, and learning agents. Agentic AI can be seen as an advanced form of learning and utility-based agents because these systems continuously adapt while optimizing goals.
One interesting idea discussed in the paper is the rise of multi-agent collaboration. Instead of a single AI system performing all tasks, multiple specialized agents cooperate with each other. This reminded me of disaster rescue robots where different robots communicate and coordinate rescue operations.
The paper also discusses challenges in agentic AI, including:
- Safety concerns
- Hallucinations
- Memory limitations
- Long-term planning issues
- Ethical risks
I found this section particularly interesting because it shows that modern AI is not only about intelligence but also about reliability and responsibility.
Paper 2 — Adaptive A* Search Algorithm
The second paper focuses on improving the famous A* search algorithm using adaptive weights.
The traditional A* algorithm uses the formula:
f(n) = g(n) + h(n)
Where:
- g(n) is the actual path cost
- h(n) is the heuristic estimate to the goal
A* is widely used in robotics, autonomous vehicles, games, and navigation systems because it efficiently finds optimal paths.
The research paper proposes adaptive weights to improve search efficiency in dynamic environments. Instead of using a fixed heuristic strategy, the algorithm adjusts weights depending on environmental conditions.
This connects directly to the search algorithms studied in class, especially:
- Breadth-First Search (BFS)
- Uniform Cost Search (UCS)
- Best-First Search
- A* Search
What makes adaptive A* important is its ability to perform better in changing environments. For example, in a flood rescue scenario, road conditions and obstacles may continuously change. A static heuristic may not work efficiently, but adaptive weighting helps the algorithm respond intelligently to dynamic situations.
The paper also highlights improvements in:
- Search speed
- Path optimization
- Reduced node expansion
- Computational efficiency
This showed me that even classical AI algorithms are still evolving and being improved for modern applications.
My Learning Experience with NotebookLM
Reading research papers manually was challenging because many sections used highly technical language and mathematical explanations. Google NotebookLM helped simplify difficult concepts into easier summaries.
One major benefit of NotebookLM was that it explained complex ideas conversationally. For example, instead of simply defining adaptive heuristics mathematically, it explained their practical purpose in navigation and robotics.
NotebookLM also helped me:
- Summarize long sections quickly
- Understand technical terminology
- Compare concepts between papers
- Relate theory to practical applications
However, manual reading was still important because it helped me understand the actual methodology and structure of the research papers.
Personal Insights
After studying these papers, I realized that modern AI systems combine multiple areas of artificial intelligence together. Intelligent agents require search algorithms, optimization, memory, planning, and learning mechanisms simultaneously.
I also learned that AI research is becoming increasingly focused on autonomy. Systems are moving from simple programmed behavior toward intelligent decision-making systems capable of adapting independently.
The most interesting realization for me was that concepts we study in introductory AI courses, such as A* search and utility-based agents, are still extremely relevant in modern research and real-world applications.
Conclusion
These two research papers helped me better understand the relationship between theoretical AI concepts and real-world intelligent systems. Agentic AI demonstrates the future of autonomous intelligent systems, while adaptive A* shows how classical search algorithms are still improving for modern environments.
Using NotebookLM alongside manual reading made the learning process more interactive and effective. Overall, this assignment improved both my research understanding and technical communication skills.
Special thanks to @raqeeb_26 for motivating students to explore AI research and technical blogging within the FAST AI community.