Into the Unknown
Defaultable Bonds A defaultable bond has these characters: We have to specify the probability of default by working directly with risk-neutral probability Q. We will model the term structure of default using 1-step default probability h(t). As with other fixed income securities, we are going to calibrate h(t) to market prices. We will need to…
Fixed Income Derivatives Fixed income markets are enormous and in fact bigger than equity markets. Fixed income derivative markets are also enormous, including: Fixed income models are inherently more complex than security models, because it needs to model evolution of entire term-structure of interest rates. One of the classic ways to get around this problem…
Non-personalized recommenders systems are remarkably effective tools and still useful in various situations. For example: recommendation in print for restaurants. Sometimes we know a little about users: age, gender, zip code. We can do a little week personalization. Also product associations allow recommendations based on current page, item, or context. Summary Statistics This is a…
Agency Agency is nothing more than the idea that you give somebody else authority to act on your behave. There is a principal (the person in charge) and an agent (who takes on some responsibility from the principal). There are 4 types of common agency relationship: Agent’s Duties Agent’s Rights Principal’s Duties Principle’s Rights Tort…
Information Retrieval and Filtering Information retrieval evolved in response to the need to be able to ask questions about a large collection of documents. We have a static content base, and there is dynamic information need (a query). So we spend our time and invest in indexing the content base. The common approach used is…
Exploration is needed to find unknown actions which lead to very large rewards. Most of the reinforcement learning algorithms share one problem: they learn by trying different actions and seeing which works better. We can use a few made-up heuristics (e.g. epsilon-greedy exploration) to mitigate the problem and speed up the learning process. Multi-armed bandits…
What is fashion Fashion can be many things: creativity, business, trend that people share, or a way to express personality. There is no consensus about what fashion is. Fashion is about impulse and desire. We live not according to reason, but according to fashion. Seneca, mid 1st century. Fashion is a balancing act between the…
Third party beneficiaries Third party beneficiaries are people who benefit under a contract, but are not part to the contract. What rights do third party beneficiaries have? The answer depends on what type of beneficiary each of these parties is: ONLY intended beneficiaries can enforce agreements as a third party, who must have been vested….
The problems of value-based methods The idea behind value-based reinforcement learning (say, Q-learning) is to find an optimal action, in a state, based on how much discounted reward you will get, by following a policy. The first problem here is value-based methods do not explicit learn “what to do”, instead it learns “what kind of…
TensorFlow is an open-source, high-performance library for any numerical computation (not just for machine learning). For example, you could use TensorFlow to solve partial differential equations. In order to make TensorFlow work, you need to create directed graph, which represent the computation you want to do. In a graph, nodes represent mathematical operations, edges represent…
What is SEO? SEO stands for Search Engine Optimization, which is the practice of improving visibility of a website in search. SEO fits into a larger digital marketing strategy, it is more focused on free organic traffic. Other roles include: SEOs want their page to show on the first page of search result. The goal…
UCC stands for Uniform Commercial Code, UCC Article 2 pertains sales of goods, which provides additional rules when it comes to contracts for the sales of tangible, movable goods. Sometimes a contract calls for both goods and services, in which case use “Predominant factor test’ to distinguish between them. The question is “What is the…
AI First Think about the complex pipeline, to be successful at Machine Learning, besides creating models, you need to think about serving out machine learning predictions using notebooks, dashboards, applications, reports, etc. Operationalizing a machine model is hard. Batch data and stream data should be treated the same way. To be good at Machine Learning,…
Multi-period binomial model Multi-period binomial model is really just a series of one-period model spliced together. When pricing an European option, you can calculate it backward 1 period at a time. But you may also do the same thing just as one calculation. It appears the true probabilities p (price going up) and 1-p (price…
Deep Q-Network (DQN) is the first successful application of learning, both directly from raw visual inputs as humans do and in a wide variety of environments. It contains deep convolutional network without hand-designed features. DQN is actually no more than standard Q-learning bundled with stability and epsilon-greedy exploration. In any application, the first thing you…