- KDnuggets Top Blogs Rewards for September 2021, by Gregory Piatetsky - Oct 15, 2021.
The September blogs that earned KDnuggets Rewards include: Do You Read Excel Files with Python? There is a 1000x Faster Way; Data Scientists Without Data Engineering Skills Will Face the Harsh Truth; Path to Full Stack Data Science; Nine Tools I Wish I Mastered Before My PhD in Machine Learning
- Learn from Northwestern Data Science experts, by Northwestern - Oct 15, 2021.
Build the essential technical, analytical, and leadership skills needed for careers in today's data-driven world in Northwestern’s Master of Science in Data Science program. Apply now.
- How our Obsession with Algorithms Broke Computer Vision: And how Synthetic Computer Vision can fix it, by Paul Pop - Oct 15, 2021.
Deep Learning radically improved Machine Learning as a whole. The Data-Centric revolution is about to do the same. In this post, we’ll take a look at the pitfalls of mainstream Computer Vision (CV) and discuss why Synthetic Computer Vision (SCV) is the future.
- New Computing Paradigm for AI: Processing-in-Memory (PIM) Architecture, by Nam Sung Kim - Oct 15, 2021.
As larger deep neural networks are trained on the latest and fastest chip technologies, an important challenge remains that bottlenecks performance -- and it is not compute power. You can try to calculate a DNN as fast as possible, but there is data -- and it has to move. Data pipelines on the chip are expensive and new solutions must be developed to advance capabilities.
- How to calculate confidence intervals for performance metrics in Machine Learning using an automatic bootstrap method, by David B Rosen (PhD) - Oct 15, 2021.
Are your model performance measurements very precise due to a “large” test set, or very uncertain due to a “small” or imbalanced test set?
- Amazon Web Services Webinar: Leverage data sets to create a customer-centric strategy and improve business outcomes, by Roidna - Oct 14, 2021.
Register now for this webinar, Oct 28, to learn how using third-party data enhances applications to better prioritize your target customer - helping you build a more customer-centric business.
- Deploying Your First Machine Learning API, by Abid Ali Awan - Oct 14, 2021.
Effortless way to develop and deploy your machine learning API using FastAPI and Deta.
- The 20 Python Packages You Need For Machine Learning and Data Science, by Sandro Luck - Oct 14, 2021.
Do you do Python? Do you do data science and machine learning? Then, you need to do these crucial Python libraries that enable nearly all you will want to do.
- What is Clustering and How Does it Work?, by Satoru Hayasaka - Oct 14, 2021.
Let us examine how clusters with different properties are produced by different clustering algorithms. In particular, we give an overview of three clustering methods: k-Means clustering, hierarchical clustering, and DBSCAN.
- How Hasura Improved Conversion Rates By 20% With PostHog, by PostHog - Oct 13, 2021.
Find out how Hasura increased conversion rates by 10-20% by using PostHog for self-hosted product analytics!
- Will Your Job be Replaced by a Machine?, by Martin Perry - Oct 13, 2021.
Yes! It will happen. However, you can pivot and thrive in this disruptive time by becoming a Citizen Developer!
- How to Ace Data Science Interview by Working on Portfolio Projects, by Abid Ali Awan - Oct 13, 2021.
Recruiters of Data Science professionals around the world focus on portfolio projects rather than resumes and LinkedIn profiles. So, learning early how to contribute and share your work on GitHub, Deepnote, and Kaggle can help you perform your best during data science interviews.
- Building Multimodal Models: Using the widedeep Pytorch package, by Rajiv Shah - Oct 13, 2021.
This article gets you started on the open-source widedeep PyTorch framework developed by Javier Rodriguez Zaurin.
- Top September Stories: Do You Read Excel Files with Python? There is a 1000x Faster Way, by KDnuggets - Oct 12, 2021.
Also: Data Scientists Without Data Engineering Skills Will Face the Harsh Truth; Nine Tools I Wish I Mastered Before My PhD in ML; A Data Science Portfolio That Will Land You The Job
- Transforming your business with SAS® Viya® on Microsoft Azure, by SAS - Oct 12, 2021.
Faster, trusted decisions are in the cloud. See how you can use the flexibility, scalability and agility of modern technologies to advance your organization’s goals. Read our blog with 3-part video demo.
- Create Synthetic Time-series with Anomaly Signatures in Python, by Tirthajyoti Sarkar - Oct 12, 2021.
A simple and intuitive way to create synthetic (artificial) time-series data with customized anomalies — particularly suited to industrial applications.
- How I Built A Perfect Model And Got Into Trouble, by Oleg Novikov - Oct 12, 2021.
Data-driven decisions, actionable insights, business impact—you've seen these buzzwords in data science jobs descriptions. But, just focusing on these terms doesn't automatically lead to the best results. Learn from this real-world scenario that followed data-driven indecisiveness, found misleading insights, and initially created a negative business impact.
- Step by Step Building a Vacancy Tracker Using Tableau, by Dotun Opasina - Oct 12, 2021.
Step-by-step explanations of vacancies valued in tens of millions of dollars in the small town of Fitchburg, Massachusetts.
- PASS Data Community Summit – Free Online Conference for Data Professionals, by PASS - Oct 11, 2021.
PASS Data Community Summit 2021 is the year’s largest gathering of Microsoft data platform professionals. This FREE online conference (taking place November 8 – 12, 2021) features 200+ world-class speakers and sessions, and gives you the opportunity to connect, share, and learn with thousands of your peers from the global data platform community.
- Top Stories, Oct 4-10: How to Build Strong Data Science Portfolio as a Beginner; 38 Free Courses on Coursera for Data Science, by KDnuggets - Oct 11, 2021.
Also: Data science SQL interview questions from top tech firms; Here’s Why You Need Python Skills as a Machine Learning Engineer; 8 Must-Have Git Commands for Data Scientists; Introduction to PyTorch Lightning
- AutoML: An Introduction Using Auto-Sklearn and Auto-PyTorch, by Kevin Vu - Oct 11, 2021.
AutoML is a broad category of techniques and tools for applying automated search to your automated search and learning to your learning. In addition to Auto-Sklearn, the Freiburg-Hannover AutoML group has also developed an Auto-PyTorch library. We’ll use both of these as our entry point into AutoML in the following simple tutorial.
- Scaling human oversight of AI systems for difficult tasks – OpenAI approach, by OpenAI - Oct 11, 2021.
The foundational idea of Artificial Intelligence is that it should demonstrate human-level intelligence. So, unless a model can perform as a human might do, its intended purpose is missed. Here, recent OpenAI research into full-length book summarization focuses on generating results that make sense to humans with state-of-the-art results that leverage scalable AI-enhanced human-in-the-loop feedback.
- Query Your Pandas DataFrames with SQL, by Matthew Mayo - Oct 11, 2021.
Learn how to query your Pandas DataFrames using the standard SQL SELECT statement, seamlessly from within your Python code.
- Choose The Right Job in Data: 5 Signs To Look For In An Engineering Culture, by Niv Sluzki - Oct 8, 2021.
Software engineers seeking jobs at data companies face a new problem: choosing the right job out of all the options. Learn the 5 signs that signal an agile and innovative engineering culture.
- Are you familiar with data labeling?, by Toloka - Oct 8, 2021.
Are you familiar with common data labeling approaches and tools? Take a simple 2-minute survey.
- 8 Must-Have Git Commands for Data Scientists, by Soner Yildirim - Oct 8, 2021.
Git is a must-have skill for data scientists. Maintaining your development work within a version control system is absolutely necessary to have a collaborative and productive working environment with your colleagues. This guide will quickly start you off in the right direction for contributing to an existing project at your organization.
- Dealing with Data Leakage, by Susan Currie Sivek, Ph.D. - Oct 8, 2021.
Target leakage and data leakage represent challenging problems in machine learning. Be prepared to recognize and avoid these potentially messy problems.
- Transforming the Shop Floor: A No-BS Look at Data Science in Manufacturing, by RapidMiner - Oct 7, 2021.
Join RapidMiner live on LinkedIn, Oct 28, to learn how you can lead a digital transformation—not by starting from scratch, but by getting more from what you already have. We’ll walk through a series of real-world examples to demonstrate how your data, when paired with machine learning, can be used to make smarter process decisions.
- The Evolution of Tokenization – Byte Pair Encoding in NLP, by Harshit Tyagi - Oct 7, 2021.
Though we have SOTA algorithms for tokenization, it's always a good practice to understand the evolution trail and learning how have we reached here. Read this introduction to Byte Pair Encoding.
- Building and Operationalizing Machine Learning Models: Three tips for success, by Jason Revelle - Oct 7, 2021.
With more enterprises implementing machine learning to improve revenue and operations, properly operationalizing the ML lifecycle in a holistic way is crucial for data teams to make their projects efficient and effective.
- How to do “Limitless” Math in Python, by Tirthajyoti Sarkar - Oct 7, 2021.
How to perform arbitrary-precision computation and much more math (and fast too) than what is possible with the built-in math library in Python.
- Here’s Why You Need Python Skills as a Machine Learning Engineer, by UCSD - Oct 6, 2021.
If you want to learn how to apply Python programming skills in the context of AI applications, the UC San Diego Extension Machine Learning Engineering Bootcamp can help. Read on to find out more about how machine learning engineers use Python, and why the language dominates today’s machine learning landscape.
- Four Different Pipes for R with magrittr, by Gregory Janesch - Oct 6, 2021.
The magrittr package supplies the pipe operator (%>%), but it turns out that the package actually contains four pipe operators in total. Let's go into them a bit.
- 38 Free Courses on Coursera for Data Science, by Aqsa Zafar - Oct 6, 2021.
There are so many online resources for learning data science, and a great deal of it can be used at no cost. This collection of free courses hosted by Coursera will help you enhance your data science and machine learning skills, no matter your current level of expertise.
- My AI Plays Piano for Me, by Kathrin Melcher - Oct 6, 2021.
Training an RNN with a Combined Loss Function.
- Eight Data Science Specializations, and Why You Should Pick One, by Pace University - Oct 5, 2021.
With so many Data Science specializations, where should you focus? The Pace University online Master of Science in Data Science features elective courses which allow you to focus on topics that suit your career path so that you can begin to develop a unique specialization.
- Will Data Analysts be Replaced by AI?, by Ngwa Bandolo Bobga Cyril - Oct 5, 2021.
It's the question so many are asking: will data analysts be replaced by AI? Read this well-reasoned and concise opinion by someone with insight into the matter.
- Data science SQL interview questions from top tech firms, by Nate Rosidi - Oct 5, 2021.
As a data scientist, there is one thing you really need to understand and know how to handle: data. With SQL being a foundational technical approach for working with data, it should not be surprising that the top tech companies will ask about your SQL skills during an interview. Here, we cover the key concepts tested so you can best prepare for your next data science interview.
- The Architecture Behind DeepMind’s Model for Near Real Time Weather Forecasts, by Jesus Rodriguez - Oct 5, 2021.
Deep Generative Model of Rain (DGMR) is the newest creation from DeepMind which can predict precipitation in short term intervals.
- Top Stories, Sep 27 – Oct 3: Path to Full Stack Data Science, by KDnuggets - Oct 4, 2021.
Also: How To Build A Database Using Python; Surpassing Trillion Parameters and GPT-3 with Switch Transformers – a path to AGI?; Nine Tools I Wish I Mastered Before My PhD in Machine Learning; 20 Machine Learning Projects That Will Get You Hired
- Parallelizing Python Code, by Borycki & Galarnyk - Oct 4, 2021.
This article reviews some common options for parallelizing Python code, including process-based parallelism, specialized libraries, ipython parallel, and Ray.
- How to Build Strong Data Science Portfolio as a Beginner, by Abid Ali Awan - Oct 4, 2021.
After learning the basics of data science, you can start to work on real-world problems. But how do you showcase your work? In this article, we are going to learn a unique way to create a data science portfolio.
- Introduction to PyTorch Lightning, by Kevin Vu - Oct 4, 2021.
PyTorch Lightning is a high-level programming layer built on top of PyTorch. It makes building and training models faster, easier, and more reliable.
- Cartoon: How Deep Is That Data Lake?, by Gregory Piatetsky - Oct 2, 2021.
New KDnuggets Cartoon looks at some of the problems data engineers may encounter when trying to measure data lakes.
- Teaching AI to Classify Time-series Patterns with Synthetic Data, by Tirthajyoti Sarkar - Oct 1, 2021.
How to build and train an AI model to identify various common anomaly patterns in time-series data.
- Surpassing Trillion Parameters and GPT-3 with Switch Transformers – a path to AGI?, by Kevin Vu - Oct 1, 2021.
Ever larger models churning on increasingly faster machines suggest a potential path toward smarter AI, such as with the massive GPT-3 language model. However, new, more lean, approaches are being conceived and explored that may rival these super-models, which could lead to a future with more efficient implementations of advanced AI-driven systems.
- How to Auto-Detect the Date/Datetime Columns and Set Their Datatype When Reading a CSV File in Pandas, by David B Rosen (PhD) - Oct 1, 2021.
When read_csv( ) reads e.g. “2021-03-04” and “2021-03-04 21:37:01.123” as mere “object” datatypes, often you can simply auto-convert them all at once to true datetime datatypes.