Category: automation

ML Model Degradation, and why work only just starts when you reach production

ML Model Degradation, and why work only just starts when you reach production

The assumption that a Machine Learning (ML) project is done when a trained model is put into production is quite faulty. Neverthless, according to Alexandre Gonfalonieri — artificial intelligence (AI) strategist at Philips — this assumption is among the most common mistakes of companies taking their AI products to market.

Actually, in the real world, we see pretty much the opposite of this assumption. People like Alexandre therefore strongly recommend companies keep their best data scientists and engineers on a ML project, especially after it reaches production!

Why?

If you’ve ever productionized a model and really started using it, you know that, over time, your model will start performing worse.

In order to maintain the original accuracy of a ML model which is interacting with real world customers or processes, you will need to continuously monitor and/or tweak it!

In the best case, algorithms are retrained with each new data delivery. This offers a maintenance burden that is not fully automatable. According to Alexandre, tending to machine learning models demands the close scrutiny, critical thinking, and manual effort that only highly trained data scientists can provide.

This means that there’s a higher marginal cost to operating ML products compared to traditional software. Whereas the whole reason we are implementing these products is often to decrease (the) costs (of human labor)!

What causes this?

Your models’ accuracy will often be at its best when it just leaves the training grounds.

Building a model on relevant and available data and coming up with accurate predictions is a great start. However, for how long do you expect those data — that age by the day — continue to provide accurate predictions?

Chances are that each day, the model’s latent performance will go down.

This phenomenon is called concept drift, and is heavily studied in academia but less often considered in business settings. Concept drift means that the statistical properties of the target variable, which the model is trying to predict, change over time in unforeseen ways.

In simpler terms, your model is no longer modelling the outcome that it used to model. This causes problems because the predictions become less accurate as time passes.

Particularly, models of human behavior seem to suffer from this pitfall.

The key is that, unlike a simple calculator, your ML model interacts with the real world. And the data it generates and that reaches it is going to change over time. A key part of any ML project should be predicting how your data is going to change over time.

Read more about concept drift here.

Via

How do we know when our models fail?

You need to create a monitoring strategy before reaching production!

According to Alexandre, as soon as you feel confident with your project after the proof-of-concept stage, you should start planning a strategy for keeping your models up to date.

How often will you check in?

On the whole model, or just some features?

What features?

In general, sensible model surveillance combined with a well thought out schedule of model checks is crucial to keeping a production model accurate. Prioritizing checks on the key variables and setting up warnings for when a change has taken place will ensure that you are never caught by a surprise by a change to the environment that robs your model of its efficacy.

Alexandre via

Your strategy will strongly differ based on your model and your business context.

Moreover, there are many different types of concept drift that can affect your models, so it should be a key element to think of the right strategy for you specific case!

Image result for concept drift
Different types of model drift (via)

Let’s solve it!

Once you observe degraded model performance, you will need to redesign your model (pipeline).

One solution is referred to as manual learning. Here, we provide the newly gathered data to our model and re-train and re-deploy it just like the first time we build the model. If you think this sounds time-consuming, you are right. Moreover, the tricky part is not refreshing and retraining a model, but rather thinking of new features that might deal with the concept drift.

A second solution could be to weight your data. Some algorithms allow for this very easily. For others you will need to custom build it in yourself. One recommended weighting schema is to use the inversely proportional age of the data. This way, more attention will be paid to the most recent data (higher weight) and less attention to the oldest of data (smaller weight) in your training set. In this sense, if there is drift, your model will pick it up and correct accordingly.

According to Alexandre and many others, the third and best solution is to build your productionized system in such a way that you continuously evaluate and retrain your models. The benefit of such a continuous learning system is that it can be automated to a large extent, thus reducing (the human labor) maintance costs.

Although Alexandre doesn’t expand on how to do these, he does formulate the three steps below:

Via the original blog

In my personal experience, if you have your model retrained (automatically) every now and then, using a smart weighting schema, and keep monitoring the changes in the parameters and for several “unit-test” cases, you will come a long way.

If you’re feeling more adventureous, you could improve on matters by having your model perform some exploration (at random or rule-wise) of potential new relationships in your data (see for instance multi-armed bandits). This will definitely take you a long way!

Solving concept drift (via)
AutoML-Zero: Evolving Machine Learning Algorithms From Scratch

AutoML-Zero: Evolving Machine Learning Algorithms From Scratch

Google Brain researchers published this amazing paper, with accompanying GIF where they show the true power of AutoML.

AutoML stands for automated machine learning, and basically refers to an algorithm autonomously building the best machine learning model for a given problem.

This task of selecting the best ML model is difficult as it is. There are many different ML algorithms to choose from, and each of these has many different settings ([hyper]parameters) you can change to optimalize the model’s predictions.

For instance, let’s look at one specific ML algorithm: the neural network. Not only can we try out millions of different neural network architectures (ways in which the nodes and lyers of a network are connected), but each of these we can test with different loss functions, learning rates, dropout rates, et cetera. And this is only one algorithm!

In their new paper, the Google Brain scholars display how they managed to automatically discover complete machine learning algorithms just using basic mathematical operations as building blocks. Using evolutionary principles, they have developed an AutoML framework that tailors its own algorithms and architectures to best fit the data and problem at hand.

This is AI research at its finest, and the results are truly remarkable!

GIF for the interpretation of the best evolved algorithm

You can read the full paper open access here: https://arxiv.org/abs/2003.03384 (quick download link)

The original code is posted here on github: github.com/google-research/google-research/tree/master/automl_zero#automl-zero

GIF for the experiment progress
Building a realistic Reddit AI that get upvoted in Python

Building a realistic Reddit AI that get upvoted in Python

Sometimes I find these AI / programming hobby projects that I just wished I had thought of…

Will Stedden combined OpenAI’s GPT-2 deep learning text generation model with another deep-learning language model by Google called BERT (Bidirectional Encoder Representations from Transformers) and created an elaborate architecture that had one purpose: posting the best replies on Reddit.

The architecture is shown at the end of this post — copied from Will’s original blog here. Moreover, you can read this post for details regarding the construction of the system. But let me see whether I can explain you what it does in simple language.

The below is what a Reddit comment and reply thread looks like. We have str8cokane making a comment to an original post (not in the picture), and then tupperware-party making a reply to that comment, followed by another reply by str8cokane. Basically, Will wanted to create an AI/bot that could write replies like tupperware-party that real people like str8cokane would not be able to distinguish from “real-people” replies.

Note that with 4 points, str8cokane‘s original comments was “liked” more than tupperware-party‘s reply and str8cokane‘s next reply, which were only upvoted 2 and 1 times respectively.

gpt2-bert on China
Example reddit comment and replies (via bonkerfield.org/)

So here’s what the final architecture looks like, and my attempt to explain it to you.

  1. Basically, we start in the upper left corner, where Will uses a database (i.e. corpus) of Reddit comments and replies to fine-tune a standard, pretrained GPT-2 model to get it to be good at generating (red: “fake”) realistic Reddit replies.
  2. Next, in the upper middle section, these fake replies are piped into a standard, pretrained BERT model, along with the original, real Reddit comments and replies. This way the BERT model sees both real and fake comments and replies. Now, our goal is to make replies that are undistinguishable from real replies. Hence, this is the task the BERT model gets. And we keep fine-tuning the original GPT-2 generator until the BERT discriminator that follows is no longer able to distinguish fake from real replies. Then the generator is “fooling” the discriminator, and we know we are generating fake replies that look like real ones!
    You can find more information about such generative adversarial networks here.
  3. Next, in the top right corner, we fine-tune another BERT model. This time we give it the original Reddit comments and replies along with the amount of times they were upvoted (i.e. sort of like likes on facebook/twitter). Basically, we train a BERT model to predict for a given reply, how much likes it is going to get.
  4. Finally, we can go to production in the lower lane. We give a real-life comment to the GPT-2 generator we trained in the upper left corner, which produces several fake replies for us. These candidates we run through the BERT discriminator we trained in the upper middle section, which determined which of the fake replies we generated look most real. Those fake but realistic replies are then input into our trained BERT model of the top right corner, which predicts for every fake but realistic reply the amount of likes/upvotes it is going to get. Finally, we pick and reply with the fake but realistic reply that is predicted to get the most upvotes!
What Will’s final architecture, combining GPT-2 and BERT, looked like (via bonkerfield.org)

The results are astonishing! Will’s bot sounds like a real youngster internet troll! Do have a look at the original blog, but here are some examples. Note that tupperware-party — the Reddit user from the above example — is actually Will’s AI.

COMMENT: 'Dune’s fandom is old and intense, and a rich thread in the cultural fabric of the internet generation' BOT_REPLY:'Dune’s fandom is overgrown, underfunded, and in many ways, a poor fit for the new, faster internet generation.'
bot responds to specific numerical bullet point in source comment

Will ends his blog with a link to the tutorial if you want to build such a bot yourself. Have a try!

Moreover, he also notes the ethical concerns:

I know there are definitely some ethical considerations when creating something like this. The reason I’m presenting it is because I actually think it is better for more people to know about and be able to grapple with this kind of technology. If just a few people know about the capacity of these machines, then it is more likely that those small groups of people can abuse their advantage.

I also think that this technology is going to change the way we think about what’s important about being human. After all, if a computer can effectively automate the paper-pushing jobs we’ve constructed and all the bullshit we create on the internet to distract us, then maybe it’ll be time for us to move on to something more meaningful.

If you think what I’ve done is a problem feel free to email me , or publically shame me on Twitter.

Will Stedden via bonkerfield.org/2020/02/combining-gpt-2-and-bert/

Building a $86 million car theft AI in 57 lines of JavaScript

Building a $86 million car theft AI in 57 lines of JavaScript

Tait Brown was annoyed at the Victoria Police who had spent $86 million Australian dollars on developing the BlueNet system which basically consists of an license-plate OCR which crosschecks against a car theft database.

Tait was so disgruntled as he thought he could easily replicate this system without spending millions and millions of tax dollars. And so he did. In only 57 lines of JavaScript, though, to be honest, there are many more lines of code hidden away in abstraction and APIs…

Anyway, he built a system that can identify license plates, read them, and should be able to cross check them with a criminal database.

Via Medium

I really liked reading about this project, so please do so if you’re curious via the links below:

Part 1: How I replicated an $86 million project in 57 lines of code

Part 2: Remember the $86 million license plate scanner I replicated?

Part X: the code on Github

Cover image via Medium via Freepik

The 12 Truths of Machine Learning – by Delip Rao

The 12 Truths of Machine Learning – by Delip Rao

In this original blog, with equally original title, Delip Rao poses twelve (+1) harsh truths about the real world practice of machine learning. I found it quite enlightning to read a non-hyped article about ML for once. Particularly because Delip’s experiences seem to overlap quite nicely with the principles of software design and Agile working.

Delip’s 12 truths I’ve copied in headers below. If they spark your interest, read more here:

  1. It has to work
  2. No matter how hard you push and no matter what the priority, you can’t increase the speed of light
  3. With sufficient thrust, pigs fly just fine. However, this is not necessarily a good idea
  4. Some things in life can never be fully appreciated nor understood unless experienced firsthand
  5. It is always possible to agglutinate multiple separate problems into a single complex interdependent solution. In most cases, this is a bad idea
  6. It is easier to ignore or move a problem around than it is to solve it
  7. You always have to tradeoff something
  8. Everything is more complicated than you think
  9. You will always under-provision resources
  10. One size never fits all. Your model will make embarrassing errors all the time despite your best intentions
  11. Every old idea will be proposed again with a different name and a different presentation, regardless of whether it works
  12. Perfection has been reached not when there is nothing left to add, but when there is nothing left to take away

Delip added in a +1, with his zero-indexed truth: You are Not a Scientist.

Yes, that’s all of you building stuff with machine learning with a “scientist” in the title, including all of you with PhDs, has-been-academics, and academics with one foot in the industry. Machine learning (and other AI application areas, like NLP, Vision, Speech, …) is an engineering research discipline (as opposed to science research).

Delip Rao via deliprao.com/archives/227

Delip [bio] is the VP of Research at AI Foundation where he leads speech, language, and vision research efforts for generating and detecting artificial content. You can find his personal webblog here.

Cover image via the-vital-edge.com/lie-detector

Python Web Scraping: WordPress Visitor Statistics

Python Web Scraping: WordPress Visitor Statistics

I’ve had this WordPress domain for several years now, and in the beginning it was very convenient.

WordPress enabled me to set up a fully functional blog in a matter of hours. Everything from HTML markup, external content embedding, databases, and simple analytics was already conveniently set up.

However, after a while, I wanted to do some more advanced stuff. Here, the disadvantages of WordPress hosting became evident fast. Anything beyond the most simple capabilities is locked firmly behind paywalls. Arguably rightfully so. If you want to use WordPress’ add-ins, I feel you should pay for them. That’s their business model after all.

However, what greatly annoys me is that WordPress actively hinders you from arranging matters yourself. Want to incorporate some JavaScript in your page? Upgrade to a paid account. Want to use Google Analytics? Upgrade and buy an add-in. Want to customize your HTML / CSS code? Upgrade or be damned. Even the simplest of tasks — just downloading visitor counts — WordPress made harder than it should be.

You can download visitor statistics manually — day by day, week by week, or year by year. However, there is no way to download your visitor history in batches. If you want to have your daily visiting history, you will manually have to download and store every day’s statistics.

For me, getting historic daily data would entail 1100 times entering a date, scrolling down, clicking a button, specifying a filename, and clicking to save. I did this once, for 36 monthly data snapshots, and the insights were barely worth the hassle, I assure you.

Fortunately, today, after nearly three years of hosting on WordPress, I finally managed to circumvent past this annoyance! Using the Python script detailed below, my computer now automonously logs in to WordPress and downloads the historic daily visitor statistics for all my blogs and pages!

Let me walk you through the program and code.

Modules & Setup

Before we jump into Python, you need to install Chromedriver. Just download the zip and unpack the execution file somewhere you can find it, and make sure to copy the path into Python. You will need it later. Chromedriver allows Python’s selenium webdriver to open up and steer a chrome browser.

We need another module for browsing: webdriver_manager. The other modules and their functions are for more common purposes: os for directory management, re for regular expression, datetime for working with dates, and time for letting the computer sleep in between operations.

from selenium import webdriver
from webdriver_manager.chrome import ChromeDriverManager
from time import sleep
from datetime import datetime, timedelta
import os
import re

Helper Functions

I try to write my code in functions, so let’s dive into the functions that allow us to download visitor statistics.

To begin, we need to set up a driver (i.e., automated browser) and this is what get_driver does. Two things are important here. Firstly, the function takes an argument dir_download. You need to give it a path so it knows where to put any downloaded files. This path is stored under preferences in the driver options. Secondly, you need to specify the path_chromedriver argument. This needs to be the exact location you unpacked the chromedriver.exe. All these paths you can change later in the main program, so don’t worry about them for now. The get_driver function returns a ready-to-go driver object.

def get_driver(dir_download, path_chromedriver):
    chrome_options = webdriver.ChromeOptions()
    prefs = {'download.default_directory': dir_download}
    chrome_options.add_experimental_option('prefs', prefs)
    driver = webdriver.Chrome(executable_path=path_chromedriver, options=chrome_options)
    return driver

Next, our driver will need to know where to browse to. So the function below, compile_traffic_url, uses an f-string to generate the url for the visitor statistics overview of a specific domain and date. Important here is that you will need to change the domain default from paulvanderlaken.com to your own WordPress adress. Take a look at the statistics overview in your regular browser to see how you may tailor your urls.

Now, in the rest of the program, I work dates formatted and stored as datetime.datetime.date(). By default, the compile_traffic_url function also uses a datetime date argument for today’s date. However, WordPress expects simple string dates in the urls. Hence, I need a way to convert these complex datetime dates into simpler strings. That’s what the strftimefunction below does. It formats a datetime date to a date_string, in the format YYYY-MM-DD.

def compile_traffic_url(domain='paulvanderlaken.com', date=datetime.today().date()):
    date_string = date.strftime('%Y-%m-%d')
    return f'https://wordpress.com/stats/day/posts/{domain}?startDate={date_string}'

So we know how to generate the urls for the pages we want to scrape. We compile them using this handy function.

If we would let the driver browse directly to one of these compiled traffic urls, you will find yourself redirected to the WordPress login page, like below. That’s a bummer!

Hence, whenever we start our program, we will first need to log in once using our password. That’s what the signing_in function below is for. This function takes in a driver, a username, and a password. It uses the compile_traffic_url function to generate a traffic url (by default of today’s traffic [see above]). Then the driver loads the website using its get method. This will redirect us to the WordPress login page. In order for the webpages to load before our driver starts clicking away, we let our computer sleep a bit, using time.sleep.

def signing_in(driver, username, password):
    print('Sign in routine')

    url = compile_traffic_url()

    driver.get(url)
    sleep(1)

    field_email = driver.find_element_by_css_selector('#usernameOrEmail')
    field_email.send_keys(username)

    button_submit = driver.find_element_by_class_name('button')
    button_submit.click()

    sleep(1)

    field_password = driver.find_element_by_css_selector('#password')
    field_password.send_keys(password)

    button_submit = driver.find_element_by_class_name('button')
    button_submit.click()

    sleep(2)

Now, our automated driver is looking at the WordPress login page. We need to help it find where to input the username and password. If you press CTRL+SHIFT+C while on any webpage, the HTML behind it will show. Now you can just browse over the webpage elements, like the login input fields, and see what their CSS selectors, names, and classes are.

If you press CTRL+SHIFT+C on a webpage, the html behind it will show.

So, next, I order the driver to find the HTML element of the username-input field and input my username keys into it. We ask the driver to find the Continue-button and click it. Time for the driver to sleep again, while the page loads the password input field. Afterwards, we ask the driver to find the password input field, input our password, and click the Continue-button a second time. While our automatic login completes, we let the computer sleep some more.

Once we have logged in once, we will remain logged in until the Python program ends, which closes the driver.

Okay, so now that we have a function that logs us in, let’s start downloading our visitor statistics!

The download_traffic function takes in a driver, a date, and a list of dates_downloaded (an empty list by default). First, it checks whether the date to download occurs in dates_downloaded. If so, we do not want to waste time downloading statistics we already have. Otherwise, it puts the driver to work downloading the traffic for the specified date following these steps:

  1. Compile url for the specified date
  2. Driver browses to the webpage of that url
  3. Computer sleeps while the webpage loads
  4. Driver executes script, letting it scroll down to the bottom of the webpage
  5. Driver is asked to find the button to download the visitor statistics in csv
  6. Driver clicks said button
  7. Computer sleeps while the csv is downloaded

If anything goes wrong during these steps, an error message is printed and no document is downloaded. With no document downloaded, our program can try again for that link the next time.

def download_traffic(driver, date, dates_downloaded=[]):
    if date in dates_downloaded:
        print(f'Already downloaded {date} traffic')
    else:
        try:
            print(f'Downloading {date} traffic')
            url = compile_traffic_url(date=date)
            driver.get(url)
            sleep(1)
            driver.execute_script("window.scrollTo(0, document.body.scrollHeight);")
            button = driver.find_element_by_class_name('stats-download-csv')
            button.click()
            sleep(1)
        except:
            print(f'Error during downloading of {date}')

We need one more function to generate the dates_downloaded list of download_traffic. The date_from_filename function below takes in a filename (e.g., paulvanderlaken.com_posts_day_12_28_2019_12_28_2019) and searches for a regular expression date format. The found match is turned into a datetime date using strptime and returned. This allows us to walk through a directory on our computer and see for which dates we have already downloaded visitor statistics. You will see how this works in the main program below.

def date_from_filename(filename):
    match = re.search(r'\d{2}_\d{2}_\d{4}', filename)
    date = datetime.strptime(match.group(), '%m_%d_%Y').date()
    return date

Main program

In the end, we combine all these above functions in our main program. Here you will need to change five things to make it work on your computer:

  • path_data – enter a folder path where you want to store the retrieved visitor statistics csv’s
  • path_chromedriver – enter the path to the chromedriver.exe you unpacked
  • first_date – enter the date from which you want to start scraping (by default up to today)
  • username – enter your WordPress username or email address
  • password – enter your WordPress password
if __name__ == '__main__':
    path_data = 'C:\\Users\\paulv\\stack\\projects\\2019_paulvanderlaken.com-anniversary\\traffic-day\\'
    path_chromedriver = 'C:\\Users\\paulv\\chromedriver.exe'

    first_date = datetime(2017, 1, 18).date()
    last_date = datetime.today().date()

    username = "insert_username"
    password = "insert_password"

    driver = get_driver(dir_download=path_data, path_chromedriver=path_chromedriver)

    days_delta = last_date - first_date
    days = [first_date + timedelta(days) for days in range(days_delta.days + 1)]
    dates_downloaded = [date_from_filename(file) for _, _, f in os.walk(path_data) for file in f]

    signing_in(driver, username=username, password=password)

    for d in days:
        download_traffic(driver, d, dates_downloaded)
    driver.close()

If you have downloaded Chromedriver, have copied all the code blocks from this blog into a Python script, and have added in your personal paths, usernames, and passwords, this Python program should work like a charm on your computer as well. By default, the program will scrape statistics from all days from the first_date up to the day you run the program, but this you can change obviously.

Results

For me, the program took about 10 seconds to download one csv consisting of statistics for one day. So three years of WordPress blogging, or 1095 daily datasets of statistics, were extracted in about 3 hours. I did some nice cooking and wrote this blog in the meantime : )

The result after 3 hours of scraping

Compare that to the horror of having to surf, scroll, and click that godforsaken Download data as CSV button ~1100 times!!

The horror button (in Dutch)

Final notes

The main goal of this blog was to share the basic inner workings of this scraper with you, and to give you the same tool to scrape your own visitor statistics.

Now, this project can still be improved tremendously and in many ways. For instance, with very little effort you could add some command line arguments (with argparse) so you can run this program directly or schedule it daily. My next step is to set it up to run daily on my Raspberry Pi.

An additional potential improvement: when the current script encounters no statistics do download for a specific day, no csv is saved. This makes the program try again a next time it is run, as the dates_downloaded list will not include that date. Probably this some minor smart tweaks will solve this issue.

Moreover, there are many more statistics you could scrape of your WordPress account, like external clicks, the visitors home countries, search terms, et cetera.

The above are improvement points you can further develop yourself, and if you do please share them with the greater public so we can all benefit!

For now, I am happy with these data, and will start on building some basic dashboards and visualizations to derive some insights from my visitor patterns. If you have any ideas or experiences please let me know!

I hope this walkthrough and code may have help you in getting in control of your WordPress website as well. Or that you learned a thing or two about basic web scraping with Python. I am still in the midst of starting with Python myself, so if you have any tips, tricks, feedback, or general remarks, please do let me know! I am always happy to talk code and love to start pet projects to improve my programming skills, so do reach out if you have any ideas!

Join 332 other followers