13 post articles, 2 pages.
Welcome to My Blog
Welcome to my blog—a space to explore ideas in math and algorithms without any pretense. Here, I'm going to share posts on interesting research papers, intriguing computational problems, and more. I’m currently considering features like monthly quizzes to keep things interactive. My goal is to create a friendly, accessible environment where we can all learn and discuss these topics together.
Feedback and participation are always welcome. In the future, I plan on adding Gitalk for comments and a Discord server for discussions.
A Taxonomy of Quantitative Trading Strategies
Quantitative trading strategies come in many shapes and sizes,
but can we classify them clearly and intuitively?
This post presents a simple, taxonomy to help you understand
and categorize quantitative strategies based on their
assumptions about current and future market data.
Additionally, what is a taxonomy you might ask?
A taxonomy rule ...
Generalized Policy Iteration (GPI) - Theory and Code Example
In this tutorial, we’ll explore Generalized Policy Iteration (GPI) through two approaches:
SARSA (model-free) – Learns directly from experience without knowing the environment’s transition dynamics.
Policy Iteration (model-based) – Uses a known transition model to compute an optimal policy.
We’ll implement both on FrozenLake-v1 (determini...
Messaging Patterns with ZeroMQ
Messaging Patterns - What are they and how to use them with ZeroMQ
Messaging systems are essential for building scalable and reliable distributed
applications.
They provide a way for different components and processes to communicate and
coordinate actions and data exchange.
Messaging systems are usually very simple and easy to use, but they can...
Interfacing core logic in C++ with Python - The Black-Litterman model
Interfacing core logic in C++ with Python - The Black-Litterman model
Our main motivation with this tutorial is to show how intensive computations can
be ported to C++ for performance and then interfaced with Python for ease of use.
Additionally, loads of parallel and distributed computing libraries are available,
such as cuBLAS, cuDNN, and cuF...
Algorithm Showcase - Stochastic Programming
Guaranteeing Minimum Wealth via Two-Stage Stochastic Programming in Finance
In many real-world financial settings—such as asset–liability management for pension funds or risk management for hedge funds—
a portfolio manager must ensure that the portfolio’s value never falls below a critical threshold, which also helps with reducing liquidity ris...
JAX - Vectorizing High-Dimensional Functions
Vectorizing High-Dimensional Functions with JAX
If you understand how neural networks work, you know that they are essentially a series of matrix multiplications and non-linear transformations.
In Python, a simple algorithm to perform matrix multiplication might look like this:
import numpy as np
def matrix_multiply(A, B):
return np.dot(A...
JAX - Bayesian Inference with Hamiltonian Monte Carlo for Financial Time Series
Bayesian Inference with Hamiltonian Monte Carlo for Financial Time Series
Introduction
Financial time series often exhibit complex dynamics, making them difficult to model using traditional statistical methods. Bayesian inference provides a probabilistic framework for estimating model parameters while incorporating prior knowledge. However, man...
JAX - Automatic Differentiation and JIT Compilation
JAX JIT Compilation and Autograd: Understanding the Basics
1. JAX JIT Compilation:
In JAX, Just-In-Time (JIT) compilation is a powerful optimization technique
that speeds up the execution of functions by compiling them into efficient machine code.
When you apply jax.jit to a function, JAX transforms it into a highly optimized,
low-level repres...