anthony_doan

10 ms·

61.
by anthony_doan 7y ago
It's a nice approach to concurrency. It doesn't guarantee parallelization. Concurrency is a precursor and a requirement for parallelization but it doesn't guarantee it.
62.
by anthony_doan 7y ago
I learn about Actor model via Elixir/Erlang. A good place to start is perhaps learning Elixir and eventually move to OTP and it's actor model. Actor are just processes in Erlang/Elixir (underlying implementation is a thread t
63.
by anthony_doan 7y ago
Yeah AI world have a loose definition of explanability and interpretablity. I also see this very dogmatic mindset that Deep Learning will do prediction and interpretability. What is stopping you from building two models? a regression stati
64.
by anthony_doan 7y ago
Jobless part suck but I think it's smart of them to try. It's like they're playing the law of large number. Keep on trying until they get a unicorn.
65.
by anthony_doan 7y ago
I didn't even realize they launched it. I thought it was vapor ware. I went to their party where they talk about it and it was underwhelming. They did not give any number comparison to Nvidia and such.
66.
by anthony_doan 7y ago
It is. Health sector uses mostly statistic. We use randomized experiment and with survival analysis and longitudinal analysis. Then we're also on observation studies with casuality since the 1980s between statistician and econometricia
67.
by anthony_doan 7y ago
Correct. It makes it so for(true) in other languages doesn't hijack the cpu and strive for the low latency goals of Erlang. To be fair you can go about it by just filling up a process mail box with tons of messages.
68.
by anthony_doan 7y ago
I disagree. FDA have set guidelines for clinical trials and moral standards to it. For an example you cannot experiment on a person without their consent and you have to state out what your trials is doing. It all lay out. It's because
69.
by anthony_doan 7y ago
You're advocating between drugs and the alternative you listed. The problem is the alternative you've listed are usually for prevention not as a cure. At least all the papers I've seen are observational and retrospective. Dru
70.
by anthony_doan 7y ago
If he/she chose to use dataframe then the data is probably multiple type. Matrix data structure can only have one type.
71.
by anthony_doan 7y ago
You use one of the family of `apply` function.
72.
by anthony_doan 7y ago
> nterpretability is fundamentally not a math or statistics challenge No, I keep on seeing AI/ML/DS people keep on downplaying statistic. Statistic interpret things. The majority of the models out there have a one to one, predi
73.
by anthony_doan 7y ago
In God We Trust, All Others Must Have Data - W. Edwards Deming (Statistician) I dislike it when people make statement without data which leads to politic. You have provided an article which contain data that support climate change may caus
74.
by anthony_doan 7y ago
I agree with you. I don't think country is going to give up their control of currency and all those monetary tools that they know for something new. These monetary policies have been evolving base on past mistakes. I can only see decen
75.
by anthony_doan 7y ago
> It feels like someone could write a full "Bad Blood" style piece about all the twists and turns and it'd be really interesting. You're not wrong. He was hailed as the savior of Nissan and there were a manga of him a
76.
by anthony_doan 7y ago
> Is there much of that going on? Most of that are statistical methods in statistic and not ML. We're talking about survival analysis, longitudinal analysis, clinical trial, nonparametric statistic, etc... From my experiences ML is
77.
by anthony_doan 7y ago
There is something to be said to build a programing language to solve a certain task in mind. Being vector aware and having a dataframe support in R is much more elegant for me than Python's add on library. It's like Scala buildin
78.
by anthony_doan 7y ago
Small data, Kalman filter, belief... They're using Bayesian statistic. Probably added some expert belief to their prior distribution (informative prior). The whole article made it sound spicier than it has to be. The majority of the ML
79.
by anthony_doan 7y ago
We had an American Statistical Association workshop for the local Southern California chapter recently. The speaker for the workshop mentioned he had to often change Logistic regression wikipedia page back to statistic technique from Machin
80.
by anthony_doan 7y ago
I'm going to buck the trend and state that I love using R for modeling and statistic. The R packages for these domains are one of the best I've seen. As for R in production, I would wrap it using https://www.rplumber.io
81.
by anthony_doan 7y ago
Referrals. They work with other companies, and these companies give them a unique code to apply for discount. When users use these code they get a commission. It's like CJ (commission junctions www.cj.com) that many other websites such
82.
by anthony_doan 7y ago
https://en.wikipedia.org/wiki/List_of_countries_by_total_hea... It's actually linear if you compare it by percentage of GDP of country. I'm not entirely sure what metric you are using but with GDP it's l
83.
by anthony_doan 7y ago
Natural gas is ban so now they have to use coal. That's a weird logic statement. Coal is an option out of many options out there. Why are you so sure those new constructions in Berkeley will be using coal? Do you have any source of tha
84.
by anthony_doan 7y ago
I have a few projects I jump around when I'm in the mood. 1. web app in elixir/phoenix 2. trade system for stock & time series analysis 3. finance (stocks, bonds, options) I like having multiple projects because my passion end
85.
by anthony_doan 7y ago
To be fair they filed for bankruptcy and arranged a loan to help keep the pension in place. I thought it was a nice gestures. Most story like these I've seen the tax payers end up footing the bill for the pension. A particular coal com
86.
by anthony_doan 7y ago
> A lot of those parts are from third party suppliers (not that it excuses Tesla completely of course). What makes you think that other automaker aren't using third party suppliers? Borgwarner, Denso, etc... are all third party supp
87.
by anthony_doan 7y ago
> I don't think it's untill you get to the NN based models that they start treating time as a first-class component in the model. They're bad at prediction for the past several m1-m4 time series tournament for univariate.
88.
by anthony_doan 7y ago
> Looks like they are trying to use AI for automation and it's screwed up. Again. And it will and must happen again and again to get better. Excuse me? Why can't they get better data? It seems here they used email spam data to
89.
by anthony_doan 7y ago
R have <- operator and a quick google search Python currently isn't using <-. So being pragmatic I would choose <- over ←.
90.
by anthony_doan 7y ago
You can't get faster engine without having slower one first. How do they even know what to improve upon without building it and experiment with it? Also simulation is not a replacement for real world testing. They should launch stuff i

More ›