Reviewing latest Research Papers

Goal:
The goal of the Research paper titled "The Rise of Agentic AI: A Review of Definitions, Frameworks, and Challenges" is to differentiate between Agentic AI and Gen AI or other types of AI in regards to Core Definitions,Functional Components,Architectural Modals,Frameworks and tools,Application Domain,Evaluation Metrics and Challenges and Limitations.
Narrowing down,this paper simply shows shifting from simple agents to independent acting agents,requiring no human intervention.
After scanning through this paperI learnt about the Anatomy of Agentic AI systems,starting from scratch to actually deploying them.Specifically about Core Definitions,Functional Components,Architectural Modals,Frameworks and tools,Application Domain,Evaluation Metrics and Challenges and Limitations.
For instance,In class, we learned that a simple reflex agent acts only on the current percept. This paper describes how Agentic AI functions as a Goal-Based Agent on steroids. It uses an orchestration layer to break a high-level goal (like "plan a 3-day trip") into sub-tasks, choosing the right tools (APIs) to execute them.
Furthermore,in memory,standard LLMs are usually stateless.However,this paper focuses integration of Episodic and long-term memory into AI systems in order for the systems to learn from past failures and keep updating itself.

What I found Interesting:
What i found pretty interesting was the hierarchical Architecture wherein it mirrors corporate sector.Like how a boss in the office assigns task to his employees .I found out that although these systems are highly powerful but when it comes to Long horizon coordination these can be extremely vulnerable.And this is why human intervention is always needed for systems.

Goal:
The goal of the Research paper titled "Research on the A* Algorithm Based on Adaptive Weights" is to explain and highlight the various approaches ,methods by which the A* algorithm is and can be optimized alot more than what we have currently.

Optimization strategies refer to change of approach i.e four neighbor to eight neighbor search with a diagonal free five way search automatically filter out diagonal moves when an obstacle is detected,lowering Time and Space complexity,increasing efficiency,improving path safety,and less burdening memory.

Next strategy includes, calculation of distance i.e Euclidean distance to Manhattan distance.Manhattan distance lowers the g(n) function in the output of algorithm and hence this greatly effects the output ,the path chosen by the algorithm.

What I found Interesting:
I found it interesting that the mathematical approach to the path smoothing particularly using Bessel Curves turns sharp turns into navigable arcs.More practically the improved algorithm increases search nodes average by 76.4% and turning angles by 71.7%.

Find my video following this URL:https://youtu.be/MsnhFPYBaIY

@raqeeb_26