RIP Johann Cruijff

Zijn jij een van de beste Oranje spelers; jij had voor altijd de spelen en de historie van het Nederlandse voetbal veranderden. Hartelijk dank.

Thoughts on AlphaGo vs Lee Sedol - What does Artificial Intelligence mean to our future?

I followed the recent matches of AlphaGo (by Google DeepMind) vs Lee Sedol, and even though I am not a Go player nor a computer scientist, I was genuinely intrigued by the complexity of the game, the intensity of the battle, and the capability of machine learning. AlphaGo is a prototype program that was created 2 years ago, it learned to play Go by playing against top players and itself over and over in order to learn from its mistakes. Since last year, it had defeated Fan Hui, a European Champion Go player, with five straight wins. Before that, computer programs had little success in defeating human players because of the large number of possibilities in the game and the finite playing time. Thus this is a well-anticipated battle to show the improvement and advancement in the latest development of machine learning. Lee, a Korean 9-dan (highest ranking - Divinity) Go master, was often regarded as one of the best players in the world today, played 5 matches against AlphaGo. He had faced tremendous pressure during each game and did not feel like he had any lead at all at any point. By the end of the game 3, he was losing to AlphaGo 0-3. The result caused a lot of emotional responses from the public, which some believed that it is about time that AI is more intelligent than us, and that it will eventually rule the world. Most people started to root for Lee and hoped that he would defend us human beings against machine, and when he won game 4, the world cheered. This got me thinking, is this really a battle between human and machine? Are we trying to draw to a black-and-white conclusion here that the winner (machine) reigns supreme?

After giving it some thoughts, and understanding how powerful machine learning is and what it can do, it is still probably too soon to be worried that AI will take over our world. However, its capability is a wonderful tool for us as it can empower us in an innovative way, and I think this is rather the meaning behind these games.

First of all, the battle is merely in the game of Go. It is only meant to demonstrate the capability of machine learning, and how it is capable to improve itself by learning from others and itself. Then in combination of its uncanny speed in calculating probabilities in each possibility, would be able to solve complex problems and complete very difficult tasks - to the point that it can be better than humans.

However, in real life, there are so much more complex tasks that would still require human inputs in order to be carried out and accomplished. For example, think of our roles in creating arts, performing arts, negotiations, relationship buildings etc. Even though AI can have very high intelligence quotient in solving problems, it lacks emotion quotient (EQ), and this is what makes human input/interaction meaningful and crucial. A lot of times connections and emotions are the keys to getting the jobs done. For example, doing business requires solid relationships/partnerships - people like to do business with those they feel comfortable with and are able to feel connected to; performing the Chopin concertos require deep emotional understanding of the composer and the meaning of the masterpiece etc. All of which would not be achieved without the human part. Our civilization is much broader than just 'intelligence': there are emotions, values, virtues, manner, morale, spirits, religions etc, those are essential part of the civilization which AI does not possess (at least not today).

Furthermore, with the discovery and advancement of machine learning, we should utilize it as a way to assist us to look into things that we might have missed. As the developers in AlphaGo stated, machine learning might even be able to assist the healthcare providers in some day, by identifying problems that we have neglected to observe from individual incidents and cases, and by enough repetitions that translate to learning, coming up with information and ideas on what we should pay further attention to in order to reduce errors and improve the quality of healthcare.

So after all, although there was a clear 'winner' in the battle, it is not a competition between human and AI, but a demonstration of the possibilities and capabilities in machine learning. What can we do instead, is to take this opportunity to reflect on ourselves as human being, on how we can make the best out of AI. We need to have a solid understanding of what we are doing, what we are capable of, and then determine what are the things that we can delegate to AI to make our life easier. We can optimize the strength of AI as a way to make us better human beings by allowing us to achieve more than we could do before with better awareness or resources; then maximize our uniqueness (the human aspect) to innovate, to change, and to accomplish more great things for the world.

Another takeaway from the success of AlphaGo is its ability to "learn and master" the game by playing itself over millions of times through trial and error. It is a good example and a reminder for us that the old saying "Practice Makes Perfect" is still the key to success. Although we might not be able to do the same things a million times, with failure and willingness to succeed, we will gain wisdom and our possibilities are still endless.

Youtube recordings of the 5 matches played by AlphaGo vs Lee Sedol

Avoid making these mistakes in facial masks!

I love caring for my skin, and I started at very young age.  I always loved facial masks as well, it's sort of a skin care boost to detox, remove impurities, and nourish the skin with 'nutrition'. During weekends I can easily do 3-5 a day! I have used over 50 different masks from various brands, in cream or cloth form, from pharmacy to high-end department store brands.  In retrospect, I have purchased some based on marketed benefits, promotions, or positive reviews, without thinking too much about the ingredients, but I choose them more carefully. (For product ingredients and the benefits of skincare ingredients, feel free to visit ewg.org or paulaschoices.com for individual produce reviews)

In my opinion, facial masks are not absolutely necessary but it is a great way to address specific skin concerns.  However, using facial masks incorrectly will result in wasting the products and your money, or worse, making your skin worse! So whether you're looking to do a deep cleaning or adding some hydration to your face, be sure to avoid these mistakes:

1. Not looking at the product labels - read the labels carefully to make sure that you are not allergic to the products' ingredients

2. Buying the 'natural' marketed labels - Don't get me wrong. I love natural products and try to avoid certain chemical ingredients as much as possible.  But poison ivy is natural but it isn't good for the skin, and ingredients like Retinol (Vitamin A), Hyaluronic Acid (or Sodium Hyaluronate), or Glycerin might sound foreign to you, can be good or even essential for the skin

3. Not cleaning your face thoroughly - Clogged pores and dirt prevent you from getting the most out of the facial masks. Even if you're using a cleansing mask, it's still a good habit to remove the dirt from the surface and let the mask absorbs the impurities from your pores


4. Opting for cream mask because it seems to be cheaper and lasts longer - Cloth mask also does wonders.  When you leave it on your face, it seals the ingredients onto the skin.  For cream masks, apply generously onto your skin, you do need a good amount (more than the daily moisturizer amount) in order for your skin to absorb the essence.

Not following instructions
5.  Leaving the mask on for too long (except for overnight/sleeping masks) - This is a huge mistake.  Cleaning mask with clay and charcoal can dry up your skin quickly.  When your skin is too dry, it can become flaky and cause peeling, or worse, form fine lines!  Even if you are doing a hydrating mask, some masks have intensive formulas that you shouldn't leave it on for more than 5-10.  After the recommended time the ingredients could irritate your skin. So the longer isn't always better.

5b. Using your facial/eye masks as a substitute to your regular skincare products. Facial masks ingredients are formulated for special skin concerns and provide a boost to the skin, but they don't provide the same protection as the regular skincare products (SPF, seed oils for antioxidants)

6. Not rinsing/wiping/rubbing properly - If it says "wipe off excess", use a facial tissue to gently wipe off any residues on your face; for rinsing, use lukewarm water to rinse off completely.  If it says massage onto your skin, use your fingertips to rub the essence onto your face in circular motions.  Be sure to follow the regular skincare routine as instructed.

7.  Talking/Laughing/Eating - Try to relax as much as you can.  Your skin will relax too and absorb the ingredients better.  In addition, if the ingredients are drying, intensive motion on your face will pull your skin and can create fine lines.

8.  Using only one facial masks - I would switch with a few products, or at least try something else and then go back to your favorite.  Once your skin is used to the ingredients, the products will become less effective.

9. Picking one that is colorful and smells like its marketed ingredients - If you care about the ingredients, try to stay away from the colorful ones or the ones that really smell like the ingredients (like 'rose' or 'cucumber').  A lot of times they contain synthetics colors and scents to make the products more 'believable' and don't add any benefits to your skin.  I prefer not having Green-3, Yellow-3, CI-xxxxx (artificial colors) and fragrance/parfum (could be made of synthetics) added to my skincare boost.


So that's it for now! Don't forget to  Happy Masking!

Gender Wage Gap - Discussion on Descriptive Statistics

Bernie Sanders is probably celebrating how close the gap was between him and Clinton in the Iowa Caucus last night. The senator, who aims to fight inequalities in minorities (people of color, women, LGBT, disabilities) and the welfare of the youth and the elderly, has been gaining support steadily. Anyone can visit his official campaign website to read more about where he stands on issues. There's one issue, particularly the 'facts' and its descriptive statistics, even as female (and somewhat of a feminist) I cannot bring myself to totally agree with it.

This issue is the endless 'gender wage gap" that politicians have been debating over for decades.
Does it exist? I believe so
Are the statistics correct? Most likely.
Are politicians fighting the anti-pay gap making a case based on statistics? Yes
So, they must be right? Not necessarily.

First, take a look at where Sanders stands on Gender Wage Gap from his issue on "Fighting for Women's Right" (Reference: https://berniesanders.com/issues/fighting-f...womens-rights/)

AS PRESIDENT, SENATOR BERNIE SANDERS WILL:

1. FIGHT FOR PAY EQUITY FOR WOMEN.
It is a national disgrace that women only earn 79 cents for every dollar a man earns.1 The gender pay gap is even worse for women of color. Today, African American women earn just 64 cents for every dollar a white male earns, while the figure for Hispanic women is just 54 cents. As president, Sen. Sanders will sign the Paycheck Fairness Act into law to end wage discrimination based on gender.

5. INCREASE THE MINIMUM WAGE TO $15 AN HOUR BY 2020.
According to the most recent statistics, women make up two-thirds of all minimum wage workers. Increasing the minimum wage to $15 an hour would significantly boost the wages of more than 15 million women and help close the gender wage gap.

6. RAISE THE TIPPED MINIMUM WAGE TO $15 AN HOUR BY 2023.
The federal tipped minimum wage of just $2.13 an hour hasn’t been raised since 1991. More than two-thirds of tipped workers are women. Increasing the tipped minimum wage to $15 an hour by 2023 would lift millions of women out of poverty and significantly reduce the gender pay gap.

Skipped #2-#4 as they were irrelevant to wage gap (they were about reproductive rights and childcare)

See any problem there with the descriptions? Most might say "Well it's just the median, so the extreme outliers won't skew the median". Well, true...but here are the considerations:

- Are the female minimum wage worker and the tipped minimum wage worker groups large enough so that the median salary includes their pay rates?
- Are the male minimum wage worker and the tipped minimum wage worker groups small enough so that the median salary does NOT include their pay rates?
If both answers to the questions above are yes, then sure, increasing the minimum wage would help. However, the fact is that it is inaccurate. Based on government data (Bureau of Labor Statistics), 'Hourly rate' (not limited to minimum wage) workers represent 58.7% of all wage and salary workers, and only 4% of those hourly rate workers are making minimum wage or below. So we can conclude that raising minimum wage is good for certain workers, but will NOT close the gender wage gap.

Now, one can argue this way: if the minimum wage increases, wouldn’t the wage of other non-minimum wage jobs increase as well? Yes, the cost of doing business would increase, however in order to generate positive returns or to merely break-even, businesses would mark up the prices of goods and services, which means the cost is transferred and absorbed by the customers. When price level goes up, it becomes inflation, and non-minimum wage workers would demand high wages in return. In the long run, everything would be adjusted but will also return to equilibrium. So is increasing minimum wage the answer to solve gender wage gap? Or is it more of a campaign tactic that Sanders is using in order to gain supports from the many lower wage workers? I would like to think it’s the latter.

Descriptive statistics, if used properly, can properly identify issues/problem/inefficiencies/inequality, and provide insights on what is or isn’t working. However, the misuse of descriptive statistics is misleading and deceptive, and can even cause the ill-informed to think/act based on incomplete information and biased messages. Unfortunately, politicians and the media often unintentionally or deliberately interpret statistics, make inferences, and draw conclusions only to their advantages.

Without turning this into a political debate, let’s focus on the main issue here. If gender wage gap exists, how should we properly measure this based on the data that we have? The answer is not as easy as we think.

What are the challenges?
The 79% (Refer to pg. 7 of the link provided below)
This is a very quick but rough comparison. It is inclusive of all workers, regardless of job function, location, skills (hard and soft) competency, tenure at the job, years of relevant work experience, age, education, licenses, and certification etc. The distribution of these attributes could vary based on gender but the comparison does not distinguish any of them. Also, the research report did not disclose the profile of the median earnings of female and male.

Comparison by state (pg. 9)
This is better, but it only helps a little by taking into consideration the diverse cost of living among states.

Comparison by age group, education (pg. 14, 15)
These comparisons are bit more convincing, at least they show that we are seeing gaps in every age group/education level. The difficult part would be to validate some of the soft facts, such as whether women are taking jobs that are as aligned to their education as their male counterpart; whether women are leaving the workforce at some point to start a family and returning years later; or have their male counterpart been switching jobs more often and negotiating salary more aggressively. These factors could be an explanation for some of the differences, rather than concluding it as a discrimination that women are hitting so-called glass ceiling that limits their earning potential.

Comparison by specific job title (pg. 17)
This seems to be the better comparison, but again the challenge is still there – as the other attributes/factors could still play a significant role in driving the differences in the measured result.

It is extremely difficult and complicated to properly show the accurate earnings gap between genders, if we have to take into account all factors that can be attributed to earnings. Even if all factors are being accounted for, ceteris paribus, still shows the gaps, our analysis would provide accuracy and granularity, but lose credibility.

-----------------------------------------------------------------------------------------------------
1 all figures are median earning based on AAUW's published result (http://www.aauw.org/files/2015/09/The-Simple-Truth-Fall-2015.pdf) the information published in this report is often used as the gender wage gap debate. It’s worthwhile to take a deep look for anyone who is interested. The report has better analysis and argument to explain the differences in the numbers than how the main-stream media with political biases choose to interpret them.

Oil Price: an example of Game Theory

Drivers have been enjoying low gasoline price thanks to the cheapest oil price in over 7 years. While some of the reasons of low price stem from a weaker global economy which results in lower demand for oil, and the recent conflict between Saudi Arabia and Iran, the true reason behind the steep decline in oil price is fueled by the price competition between the largest oil producing countries, largely between the OPEC (Organization of the Petroleum Exporting Countries) and US Shale producers. The control of supply, the discrepancy in break-even cost of production amongst individual oil producers, and the political strategy in this battle demonstrated a great example of Game Theory.

The OPEC (led by Saudi Arabia, other major producing countries including UAE, Iran, Iraq etc) was once the largest oil producer. Since the past two decades, with the increasing demand which resulted in rising oil price, and the drive to cut back on the dependence on oil imports, Russia, the US, as well as other countries resumed researches and investments for alternative/renewable energy and different methods of extracting oil. This fueled the rise of the oil shale industry, which involves mining and processing of oil from shale rocks. However, at higher cost in producing each barrel, the only incentive for shale producers to generate supply is when both the demand and the price of oil are high.

The shale producers run a lucrative business when the price is at the highest and supply as much as possible, but their counterparts in OPEC extracting oil from offshore have the same interest as well. However, they could not both achieve optimization by adopting the same strategy and enjoy the maximum profit. Thus the leaders in OPEC planned a way to regain the market by adopting the game theory model: price war - to compete at a cost of its own profits in order to drive down oil price and to undermine competing producers. This was a favorable strategy for OPEC (mainly Saudi Arabia) as it only costs $25-$30 USD per barrel to extract conventional oil from the Persian Gulf, as opposed to about $55-60 USD per barrel from shale oil production. Hence, as long as the oil price remains below mid 50s, shale producers would find themselves unprofitable and eventually be forced out of the market (note: Many shale producers are partnership funded by investors, and investors focus on dividends and returns from private business periodically)

In basic economic fundamentals, when supply exceeds demand, price goes down. Controlling the supply was exactly how the OPEC drove down the price of oil. Since the middle of 2014, OPEC decided to remain or even to increase production despite weaker demands due to slowing global economic growth. This caused a sharp decline in oil price, from $115 in July 2014 to $59 in December 2014. In 2015 OPEC continued to over-produce, which caused many shale producers to halt dividends and discontinue investments and projects. The downward momentum in price persisted into late 2015 when the price finally tumbled below $40, leaving not only shale producers to exit the business, but the revenues of many major oil producing countries also drastically reduced. With the cost of oil production, countries like Russia, Nigeria, Mexico, Canada, and Brazil are not be able to break-even and any production is operating under negative profit margin.

A simple table below explains the price war in Game Theory:

Profit/bbl: OPEC, Shale Producers

		OPEC	OPEC
		Maximize Production	Control Production
Shale Producers	Maximize Production	Price: $35 $5, $-25	Price: $75 $45, $15
Shale Producers	Control Production	Price: $65 $35, $5	Price: $115 $85, $55

For countries with oil export as a large source of revenue, the price war can turn the economies into deficits or even on verges of collapses. Even for Saudi Arabia, as the country had been spending heavily for the war in Yemen as well as addressing its own issues in the country, the tumbling oil price will only hurt the economy further. In the end, did Saudi Arabia win in the price war and eliminate competitions? Yes; but the bigger question is: when 80% of the country’s revenue comes from oil sales, is this the best strategy to adopt?

--------------------------------------------------------------
Update Jan 21 2016

Some updates on the continuation of the oil price turmoil:

Saudi Arabia
‧ Reported second quarters of trade deficits at -$8bil USD, albeit favorable comparing to the previous quarter of -$11bil USD
‧ Plans an IPO of Aramco, the state-owned oil giant, in order to find capitals for its exploration and production assets
‧ Plans to issue debt for the first time to fund budget deficit
‧ In talks with the two largest Chinese petroleum companies, China National Petroleum Corp (PetroChina) and China Petroleum & Chemical Corp (Sinopec) to invest in projects to build refineries in China.

Iran
‧ Sanctions have been lifted and Iran can return to the world oil market; plans to ramp up production to 500K barrels a day

Venezuela
‧ Requested an emergency OPEC meeting to discuss plans to curb productions in order to drive prices back upward
‧ As petroleum represents more than 95% of the country’s exports and 38% of total GDP, at current rate the export will generate $27bil in this year, down from $75bil two years before
‧ The country owes over $10 bil in debt payments and is on the brink of going to default
‧ With high inflations and shrinkage, the government can’t pay to import basic food items

United States
‧ As of Jan 15, inventories rose to a record level of 485.2 million
‧ 40 companies have filed for bankruptcy
‧ Oil boomtowns like Pearsall, TX., and Fargo, ND. have turn into ghost towns
‧ Output is expected to reduce drastically

Europe
‧ Oil giant Royal Dutch Shell saw profits tumbled 40%, has reduced 7,500 jobs last year and plans for additional reductions.
‧ Competitor BP plans to eliminate 4,000 positions exploration and production jobs, in additions to 4,000 job cuts from previous year

As price drops below the psychological significant level of $30, producers are getting gloomier at the potential recovery and shutting down production as current contracts are fulfilled. The turmoil also sends the currencies of the petrostates (petroleum as a large % of state revenue) to record low against the USD. IMO, Saudi's plan to continue draining reserves, producing at an 80% discount in order to dominate market share to the point of borrowing to fund budget deficit, is an imbecile move. I suspect that they will have trouble getting great rates: given the country's budget dependency (or lifeline) on a commodity that they are willing to drain and destroy the price at anytime. Think about it this way, who wants to buy Apple's stock if the company willingly discounts its iphone price by 80% anytime just to start a price war with other phone makers?

My Favorite Yogurt

Since I could remember, there were only few brands of yogurt on the market. We weren't as crazy as today with Yogurts, but today we have more frozen yogurt chains than the selections we had back then. Just like everything else, as we continue to strive for better quality product and make buying decision according to our needs/preferences, market responds by providing us a variety of choices. We have over hundreds of brands to choose from and each time I go to the grocery store, the variety is just dazzling. Here are some of the ones I've tried. What are your favorites?

Chobani - Probably the most popular Greek Yogurt brand today. They even have a Froyo-style store in NYC.  Flavors are expanded and different versions are sold to difference audiences - Chobani Champions for kids; Chobani bite for portion control.  Favorite Flavor: Strawberry Banana

Fage - Greek Yogurt.  I like the separate fruit cup where the yogurt stays fresh.

Favorite flavor: Strawberry & Goji

Dannon's Oikos

Stonyfield's Oikos - Always love Stonyfield yogurts.  Favorite flavor: Caramel

Yoplait - They've been around for a long time.  To stay competitive, they've made "Light", "Plus" (for digestion), "Greek", Smoothies and Frozen Yogurt to serve broader customer base. 

Siggi's Icelandic Skyr  - If you're looking for the 'regular' yogurt, then this is not the one you will like.  Although thick and sour, it's lean (skim milk) and low on sugar. Favorite flavor: Orange & Ginger

Amande - Almond cultured milk yogurt.  Vegan.  To me it's a bit runny. Favorite flavor: Peach

Green Valley - Lactose Free. It's is surprising good. Favorite flavor: Honey

SO Delicious.  Coconut Milk yogurt. Vegan.  A bit runny.

Brown Cow - who doesn't like the creamy, buttery yogurt? In the company fridge, there are so many Brown Cow yogurts that people had to put their names on the foil lids. Favorite flavor: Chocolate Underground

Kalona Supernatural - Very smooth and rich cream :). Organic.  Favorite flavor: Vanilla

Happy Valentine's Day

Hope everyone gets to enjoy this day and share you affection with your loved ones.  

Thank you DD BB!!