ZhiZhi Gewu

AI

Why Supervised Learning Works?

In essence, "Low Training Error + Sufficient Data Relative to Model Complexity → Low Test Error". Supervised learning must work on the SAME distribution. Key Concepts * Training Error (\(\text{Train}_S(f)\)): The error of a model \(f\) evaluated on the training dataset \(S\). * Test Error (\(\text{Test}_D(f)

Matrices Explained

Linear Transformations and Matrices Linear Transformations A transformation \(T\) is linear if it satisfies the following two properties: 1. Additivity: For all vectors \(x, y\) in the domain of \(T\), we have \(T(x + y) = T(x) + T(y)\). 2. Homogeneity: For all vectors \(x\) in the domain of \(T\

AI

[Study Note] V and Q Function in RF

UC Berkeley's CS285 course: lecture 4. The goal of reinforcement learning is to find an optimal policy \(\pi^*\) that maximises the cumulative rewards over time. Define the trajectory \(\tau = (s_0,a_0,s_1,a_1,...) \), which is a sequence of states and actions. The probability under

One Month Later: Getting Back on Track with Updates

Dear readers, It's been nearly a month since my last post, and I wanted to update you on my recent activities. While my day job has kept me quite busy, I've still managed to explore some fascinating topics in my spare time. I'm pleased

My Current View on Practical Asset Allocation

Recently, I read "A Fundamental Approach to Personal Investment" by David Swensen. Swensen strongly recommends that investors focus on asset allocation, rather than market timing and security selection, to generate returns. He introduces concepts from Modern Portfolio Theory (MPT) to support this approach. Modern Portfolio Theory: A Brief

Books

[Book Summary] Core Asset Classes: A Fundamental Approach to Personal Investment by David Swensen

Core Asset Classes Defining asset classes combines art and science in an attempt to group like with like, seeking as an end result a relatively homogeneous collection of investment opportunities. The author discusses the characteristics of core asset classes: * Contribute basic, valuable, and differentiable characteristics to a portfolio * Rely on

Books

[Book Summary] A Fundamental Approach to Personal Investment by David Swensen

Overview: Sources of Return * Capital markets provide three tools for investors to generate returns: asset allocation, market timing, and security selection. * Asset Allocation: long-term decision regarding the proportion of assets that an investor chooses to place in particular asset classes. * Market Timing: deviations from the long-term asset-allocation targets. It represents

The Art and Science of Decision Making (Credit Risk)

Originally post in: https://kychanbp-credit-risk-mgmt-fintech.readthedocs.io/en/latest/fundamentals/decision_making.html TL;DR Credit risk management involves complex decision-making, often driven by experience rather than innate ability. This article explores how reinforcement learning (RL) can provide a scientific framework for these decisions, focusing on trial-and-error search and delayed

Mexico's Economic Landscape: Size, Sectors, Income Distribution, Industries and Economic Base

The Size of the Economy Mexico is the 12th largest economy in the world by GDP (PPP). Table: GPD (PPP) (Millions of current international dollars) by Country Rank Country (or territory) IMF Forecast (2024) World Bank Estimate (2023) CIA Estimate (2020) 1 China 35,291,015 34,643,706 23,

Exploring Complexity and Intelligence: A Curated Reading List

Lately, I've been delving into the concept of Complexity and its connection to learning and intelligence. I came across a website called LessWrong, and I've been engrossed in its dense and informative content. I want to take notes on some of the interesting articles I'

[Write to Learn] Minimal Description Length and Kolmogorov Complexity

Recently, the concept of compression and learning has piqued my interest, as discussed in a previous blog post. To explore this further, I've chosen to delve into Peter Grünwald's tutorial note, "A Tutorial Introduction to the Minimal Description Length Principle," with the intention of

Learning <=> Compression <=> Prediction

Introduction Occasionally, I will contemplate the meaning of learning. I learned very early that learning is not memorisation. One of the reasons is that I hate memorising facts without understanding them. Also, thanks to the Hong Kong education system, there is no need to memorise many facts as equations and