Jul 12, 2025 · 2 min read · What is Reinforcement Learning? Reinforcement Learning is a trial-and-error learning process where an agent learns to make decisions by interacting with an environment. Key Components • Agent: The learner/decision-maker • Environment: The world agent...
Join discussion
Jun 25, 2025 · 5 min read · If you're reading this, you're probably curious about how machine learning can help in healthcare — or how we can teach models to work with very little data. In this blog, I’ll walk you through my recent project where I built a chest X-ray classifier...
Join discussionApr 14, 2025 · 3 min read · In 2017, I published a short reflection on the 10,000-hour rule and its relevance to programming. Years later, I find myself returning to that idea—not to reaffirm it, but to reframe it. This time, through the lens of antifragility. The 10,000 Hour ...
Join discussionMar 30, 2025 · 3 min read · 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...
Join discussionFeb 27, 2025 · 14 min read · Introduction One of the most debated aspects of artificial intelligence (AI) is creativity—the ability to generate novel, original, and meaningful solutions beyond memorization and pattern recognition. While AI models like Latent Program Networks (LP...
Join discussionFeb 27, 2025 · 17 min read · Introduction As artificial intelligence continues to advance, models like Latent Program Networks (LPNs), Large Language Models (LLMs), and hybrid AI architectures are pushing the boundaries of program synthesis, reasoning, and creativity. However, s...
Join discussionFeb 20, 2025 · 18 min read · The Abstraction and Reasoning Corpus (ARC) benchmark presents one of the most significant challenges in artificial intelligence today: assessing a model’s ability to generalize to truly novel tasks. Unlike traditional machine learning benchmarks, whi...
Join discussionFeb 20, 2025 · 17 min read · Introduction Despite the remarkable success of deep learning models, particularly large neural networks and pre-trained language models (LLMs), significant challenges arise when applying these architectures to program synthesis and abstract reasoning...
Join discussionDec 1, 2024 · 12 min read · Deep NeuroEvolution will leverage Evolution Strategies to find the best weights of the controller's linear matrix, i.e. the weights that will maximize the total reward on a full game episode. This technique is at the heart of the Policy Gradient bran...
Join discussion
