A journey of self-growth

Image for post
Image for post

Where shall I start?

I guess I’ll start with the most recent.

I am currently a research data scientist at Accenture AI Labs, we do ML research and we prototype them. We actually do a bunch of cool stuff, from automatically generating labels for training data, causal inference analysis on a distributed system, to remove bias in data and models. I read papers, I do math, and I code.

Before that, I was at Stanford doing a master’s degree in statistics, where I got a full scholarship to attend. Even before that, I was a journalist for three years, the…


Image for post
Image for post
After sunset, Top of the World, CA

“All the world’s a stage, And all the men and women merely players; They have their exits and their entrances, And one man in his time plays many parts, His acts being seven ages. At first, the infant, Mewling and puking in the nurse’s arms. Then the whining schoolboy, with his satchel And shining morning face, creeping like snail Unwillingly to school. And then the lover, Sighing like furnace, with a woeful ballad Made to his mistress’ eyebrow. Then a soldier, Full of strange oaths and bearded like the pard, Jealous in honor, sudden and quick in quarrel, Seeking the…

To my best friend’s daughter, may you live with ease, may you know how to love, may you stay feisty, fierce and brave

Image for post
Image for post

Dear Jane,

I’m auntie Yao. I hope by the time you know how to read this letter, I can still see you as often as I do now. I very much enjoy taking a walk with your mom, your dad, and your grandma on the Los Gatos Creek Trail near the first house your parents bought. We would take turns carrying you or put you in the stroller when our arms get too sore. Unlike a lot of babies who typically sit back in the stroller, you always use your little fingers to grab the bar firmly to hold your…

Here and there

Image for post
Image for post
Sausalito, CA, April, 2020

Subconsciousness. Using how we think to think about how we think. Inception.

Attachment. I’ve come to the realization that being connected doesn’t mean being attached and vice versa, and you can have both!

Albert Camus. I’ve been liking him a lot recently, gives me a lot of courage facing absurdity. Read this quote from him ‘In the midst of winter, I found there was, within me, an invincible summer. And that makes me happy. For it says that no matter how hard the world pushes against me, within me, there’s something stronger — something better, pushing right back.’

Assertiveness. I…

Demystifying T-learner, S-learner and X-learner in causal ML

Image for post
Image for post

Recently I saw Uber published causalML python library on github. It includes tree-based algorithms and meta-algorithms for estimating treatment effect in causal inference. I’m really excited about this effort — combining state-of-art machine learning techniques with causal analysis.

The repository contains a good example on Jupyter Notebook of how to use all these algorithms. For people who haven’t heard about meta-learners, it can be a little confusing though. This post I’ll try to explain some of the meta-learners mentioned in the code, all these T-Learner, S-Learner, X-learner, etc…

To begin with, I assume that you already know what is the…

When we look at politics, we tend to think that a country makes policy decisions as one entity, especially for an authoritarian country like China. However, the reality is more complicated than that and the policy outcome is always a result of games different players play behind it.

Image for post
Image for post
picture source: http://knowledge.ckgsb.edu.cn/2017/11/01/chinese-economy/state-owned-enterprise-reform-china/

Part I: Background

China started SOE reform since the mid-1990s under a dual track of “grasping the large and letting go the small”. This means: (a) privatize small and medium SOEs at the local level; (b) restructure (including merge, group/conglomerate, corporatize, and initial public offerings) large SOEs controlled by the central government. An important distinction to be made here is the large and the small/medium SOEs. Large SOEs are under the direct supervision of the central government. They are mostly located in strategic or natural monopoly industries (e.g. iron and metal, aerospace, telecommunication, airlines, banks, etc); while small/medium SOEs are under the supervision…

Love, courage, and dedication in the time of madness

Image for post
Image for post

I got a call from my mom during the Memorial Day long weekend, she told me that grandfather was hospitalized. He rode the exercise bicycle in the morning, then started coughing in the afternoon, which quickly developed into pneumonia, followed by high fever — things all of a sudden went out of control.

“I’m afraid he can’t escape this time,” my mom sobbed.

Buddhism calls sufferings ‘Kalpa’, meaning a long period of time (by human calculation) in Hindu and Buddhist cosmology. It is the period of time between the creation and recreation of a world or universe. When you escape…

A counterfactual method for causal inference

Image for post
Image for post

G-computation algorithm was first introduced by Robins in 1986 [1] to estimate the causal effect of a time-varying exposure in the presence of time-varying confounders that are affected by exposure, a scenario where traditional regression-based methods would fail.

G-computation or G-formula belongs to the G-method family [2] which also includes inverse probability weighted marginal structural models and g estimation of a structural nested model. They provide consistent estimates of contrasts (e.g. differences, ratios) of average potential outcomes under a less restrictive set of identification conditions than standard regression methods.

In this post, I’ll explain more in details of how G-computation…

A double-robust automated way for causal inference

Image for post
Image for post

What is causal inference?

Imagine you have two drugs, A and B, for treating cancer. You test the drugs on different people, now you want to measure which is more effective, how are you going to test it?

You will probably first do randomization of the patients, and give half of them drug A, and half drug B, test key metrics over a period of time to compare which drug lowers more of the key metrics. Is that enough?

No. A common misunderstanding people have is that as long as randomization is fulfilled, we can do infer that different treatments…

Yao Yang

The combination of words, numbers, and stories makes life

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store