This message aims to introduce a new custom AI gym library: https://pypi.org/project/puzzle15Gym/. It is the library for playing the classical sliding block puzzle: https://en.wikipedia.org/wiki/15_puzzle with AI agents. Motivation Introduce a way f...
Since I have joined a team which deals with AI and LLMs, I have decided to review a paper in relation to an LLM which deals with reinforcement learning of LLM and how it turns out to be better than the zero-shot learning. It had been only 3 days in t...
Introduction Brief overview of Tic Tac Toe Tic Tac Toe — the classic game we’ve all played with pen and paper — is one of the simplest yet most iconic games in existence. Played on a 3×3 grid, two players take turns marking spaces with Xs and Os. The...
In the fast-moving world of artificial intelligence, we've been dazzled by increasingly impressive claims about "reasoning models" and their problem-solving abilities. These models, enhanced through Reinforcement Learning for Reasoning (RLfR), suppos...
It has not been a long time since DeepSeek was released. It was indeed a shock to those who are in AI industry. I was not familiar with LLM’s algorithm and the computing resource usage of the LLMs. All I was doing was to utilise the LLM APIs for deve...
This paper by David Silver and Richard Sutton, prominent figures in Reinforcement Learning (RL), proposes a transition to an "Era of Experience" where AI agents, learning autonomously from world interaction, will achieve "superhuman capabilities." Wh...
Introduction Imagine teaching a dog to sit. Every time it sits correctly, you give it a treat. If it doesn’t, you wait or correct it gently. Over time, the dog learns which actions lead to rewards. Now, imagine teaching machines in a similar way — we...
“Machine Learning is the science of getting computers to learn without being explicitly programmed.” – Arthur Samuel Machine Learning (ML) is a branch of Artificial Intelligence that enables computers to learn from data and make predictions or decisi...
Introduction Artificial Intelligence (AI) has revolutionized automation, from chatbots to self-driving cars. However, most AI agents follow predefined rules or require human intervention to improve. What if AI agents could learn from their own mistak...
Abstract The financial sector faces increasing risks due to sophisticated fraud techniques, data breaches, and regulatory challenges. Artificial intelligence (AI) solutions leveraging neural networks, reinforcement learning, and generative models are...
![[Paper Review] Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://cdn.hashnode.com/res/hashnode/image/upload/v1748160260606/de66987d-98f3-4fb6-bfdd-937f4029b4bc.png)
![[Paper Review] DeepSeek-R1 Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://cdn.hashnode.com/res/hashnode/image/upload/v1745151315420/5c785e0c-f3b5-4df1-97fa-a420779b9422.png)
