WEBINAR: Deploying an Image Classification Model to the OpenMV Camera - 26 July 2022

 

Workshop Series Deploying an Image Classification Model to the OpenMV Camera Register Today Date: Tuesday, July 26, 2022 Time: 9:00 am PDT | 4 pm UTC | 7:00 pm EET Duration: 16 minutes In this IoT Central MicroSession with Edge Impulse, first see how the OpenMV camera is an all-in-one computer vision system that can be programmed with MicroPython. Then learn how to download a trained image classification model from Edge Impulse and deploy it to the OpenMV camera. You will use the OpenMV TensorFlow Lite library to perform inference and use the inference results to perform some action (e.g. write information on the preview screen when a particular object is seen). Instructor: Shawn Hymel, Senior Developer Relations Engineer, Edge Impulse Optional Hardware: OpenMV H7, OpenMV H7 R2, or OpenMV H7 Plus Preparation: Free sign-up at https://studio.edgeimpulse.com/signup Register Today

Improve your Model Performance with Auto-Encoders by Satyam Kumar via @TDataScience

Use Autoencoders as a Feature Extractor.

This was really useful and could potentially save a lot of time and effort if you can get the balance right.

WEBINAR: Build – and Choose – Better Models Faster: Data Scientists in a Box - 25 February 2021

 

‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌

Join us Feb. 25 How do you know which modeling methods will work best on your data? With so many predictive modeling methods and ways to compare and combine models, how can you easily identify the best models? Join us for a free live webinar Feb. 25, where we will explore case studies on coffee ratings, polymer manufacturing and biomarker identification, showing you how to easily build, assess and select the best-performing model. Kemal Oflus will demonstrate how to utilize machine learning methods without having to write code. Who should attend? Scientists, engineers and data explorers who want to take greater advantage of machine learning methods. Register now

AI Is More Than a Model: Four Steps to Complete Workflow Success by Johanna Pingel via @kdnuggets

The key element for success in practical AI implementation is uncovering any issues early on and knowing what aspects of the workflow to focus time and resources on for the best results—and it’s not always the most obvious steps.

I found this really interesting and gives food for thought as well as a great short roadmap for success.

The Frameworks that Google, DeepMind, Microsoft and Uber Use to Train Deep Learning Models at Scale by @jrdothoughts via @Medium

GPipe, Horovod, TF-Replicator and DeepSpeed combine cutting edge aspects of deep learning research and infrastructure to scale the training of deep learning models.

I found this fascinating.  I really hadn't quite connected all the dots in my mind to connect the frameworks up like this.

WEBINAR: How to Create Mathematical Optimization Models with Python - 29 April 2020

Data Science Central Webinar Series Event

How to Create Mathematical Optimization Models with Python Join us for the latest DSC Webinar on April 29th, 2020 With mathematical optimization, companies can capture the key features of their business problems in an optimization model and can generate optimal solutions (which are used as the basis to make optimal decisions). Data scientists with some basic mathematical programming skills can easily learn how to build, implement, and maintain mathematical optimization applications. The Gurobi Python API borrows ideas from modeling languages, enabling users to deploy and solve mathematical optimization models with scripts that are easy to write, read, and maintain. Such modules can even be embedded in decision support systems for production-ready applications. In this latest Data Science Central webinar, we will: Discuss the motivation for using Python in mathematical optimization applications Help you understand the importance of parameterizing a mathematical optimization model Review some of the best practices for deploying mathematical optimization models in Python Speaker: Juan Antonio Orozco, Optimization Support Engineer -- Gurobi Optimization Hosted by: Sean Welch, Host and Producer -- Data Science Central Title:  How to Create Mathematical Optimization Models with Python Date:  Wednesday, April 29th, 2020 Time:  9 AM - 10 AM PDT Space is limited so please register early: Reserve your Webinar seat now

All Machine Learning Models Explained in 6 Minutes by Terence Shin in @TDataScience

Intuitive explanations of the most popular machine learning models.

This deserves applause, a bookmark and sharing.

Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead by Adrian Colyer via @kdnuggets

The two main takeaways from this paper: firstly, a sharpening of my understanding of the difference between explainability and interpretability, and why the former may be problematic; and secondly some great pointers to techniques for creating truly interpretable models.

I enjoyed this article and his points which are very relevant.

WEBINAR - ML/AI Models: Continuous Integration & Deployment 11 December 2019

Data Science Central Webinar Series Event

ML/AI Models: Continuous Integration & Deployment Join us for this latest DSC Webinar on December 11th, 2019 Some things are best learned through real-world experience. Machine learning is no different. Getting machine learning right requires evolving your analytics platform to support moving data science from research into operations. It all begins with repeatable data wrangling processes that support building and deploying models. It also requires collaboration between data scientists, engineers and business analysts. With the help of tools like SAS® Model Manager, these teams can continuously and automatically train models at scale and ensure the best models are put into production. In this latest Data Science Central webinar we will discuss: Model validation best practices Various model deployment options including open source models Model scoring and training services Model performance monitoring Orchestrating a continuous learning platform Featured Speakers: Wayne Thompson, Chief Data Scientist -- SAS Lora Edwards, Principal Product Manager -- SAS Hosted by: Rafael Knuth, Contributing Editor -- Data Science Central Title:  ML/AI Models: Continuous Integration & Deployment Date:  Wednesday, December 11th, 2019 Time:  9:00 AM - 10:00 AM PST Space is limited so please register early: Reserve your Webinar seat now

WEBINAR: Train & Tune Your Computer Vision Models at Scale - 5 December 2019

Data Science Central Webinar Series Event

Train & Tune Your Computer Vision Models at Scale Join us for this latest DSC Webinar on December 5th, 2019 Whether you are training a self-driving car, detecting animals with drones, or identifying car damage for insurance claims, the steps needed to effectively train a computer vision model at scale remain the same. In this latest Data Science Central webinar, we’ll walk through best practices for managing a computer vision project including staffing, budgeting, and roles and responsibilities. Learn how to collect and label the data that will train and tune your machine learning algorithm, and which types of data labeling best fit your project along with the tools that will get the job done. In this webinar, you’ll learn how to: Identify key success factors when scoping a computer vision project Determine what kind of source data you need to make it successful Select tools that best fit your project Label your dataset so your algorithms can learn and perform as designed Speaker: Meeta Dash, Director of Product -- Figure Eight Hosted by: Stephanie Glen, Editorial Director -- Data Science Central Title:  Train & Tune Your Computer Vision Models at Scale Date:  Thursday, December 5th, 2019 Time:  9:00 AM - 10:00 AM PST Space is limited so please register early: Reserve your Webinar seat now

Facebook Has Been Quietly Open Sourcing Some Amazing Deep Learning Capabilities for PyTorch by @jrdothoughts via @TDataScience

The new release of PyTorch includes some impressive open-source projects for deep learning researchers and developers.

Interesting new features that definitely call for some experimenting to see what they can really do.

What a little more computing power can do for Deep Learning by Kim Martineau via @MIT

A deep learning model may need to see millions of photos before it can successfully identify a cat. The process is computationally intensive. But there may be a more efficient way - new MIT research shows that models only a fraction of the size are necessary.

An interesting viewpoint which could help to save money and time when developing this kind of model.

Google AI Blog:Predicting the Generalisation Gap in Deep Neural Networks by Yiding Jiang via @googleai

Here’s a description of a new technique that uses margin distributions to better predict a DNN’s generalization gap.

Seems a great idea to use what the Google AI team have made available in their Github is a great idea and should not be ignored. Links to a lot of sources are given thought the article.

How Etsy taught style to an algorithm by/via @FastCompany

Is it romantic or rustic? Boho or minimal? Etsy needed to offer searchers a way to find goods that matched their style aesthetics, but since descriptions aren’t uniform and don’t always describe the style, text mining the descriptions wasn’t enough. Colour and patterns don’t reliably predict style, so image recognition alone didn’t do it either. Enter a model that blends text analysis with image recognition based on 43 human-identified styles.

I love this real-life example detailing the steps they took to work out how to do this. Definitely, a methodology that could be used by other organisations to do a similar type of thing.

WEBINAR: Maximizing Data Science Applications And Model Development - 24th July 2019

Data Science Central Webinar Series Event

Maximizing Data Science Applications And Model Development Join us for the latest DSC Webinar on July 24th, 2019 In this latest Data Science Central webinar, you will learn the value of an optimized workstation for Data Scientists. We will demonstrate NVIDIA CUDA-X AI software stack and how it has enhanced data science workflows. Featured Speakers: Tim Lawrence, Founder and VP of Engineering and Operations -- BOXX Technologies Allen Bourgoyne, Senior Product Marketing Manager, Quadro -- NVIDIA Hosted by: Stephanie Glen, Editorial Director -- Data Science Central   Title:  Maximizing Data Science Applications And Model Development Date:  Wednesday, July 24th, 2019 Time:  9 AM - 10 AM PDT   Space is limited so please register early: Reserve your Webinar seat now

A Brief Introduction To GANs by/via @SarvasvKulpati

With explanations of the math and code

This is a great article with lots of links and examples so you can understand it. If you already have a Medium account please make sure you give him some applause for it and a follow.

WEBINAR: Creating Business Applications With R & Python - 12 February 2019

Creating Business Applications With R & Python Join us for the latest DSC Webinar on February 12th, 2019 Across industries, data scientists are creating powerful models and analytics to solve urgent business problems. However, in far too many cases, these analytics never reach their intended business users. The result is wasted time and effort, as well as a failure to achieve the fundamental goal of transforming data and analytics into solutions. Please join this latest Data Science Central webinar to see how data science teams can stop this trend and start putting analytics into action. With FICO® Xpress Insight, it's easy to take any advanced analytic asset (such as an R or Python script) and turn it into a fully functioning application for business users. We'll demonstrate some key features, including: An environment that fosters collaboration between data scientists and business users during model creation A robust interface for rapidly deploying validated models into business user-friendly applications Enablement tools for business users to run models, perform simulations, compare scenarios and visualize outcomes Data scientists can finally stop seeing their efforts go to waste and start empowering business users with the predictive and prescriptive analytics capable of transforming businesses–join us to learn more! Speakers: Bill Doyle, VP of Decision Management Solutions -- FICO Libin Varghese, Principal Sales Consultant, Decision Management Solutions -- FICO Hosted by: Bill Vorhies, Editorial Director -- Data Science Central   Title:  Creating Business Applications With R & Python Date:  Tuesday, February 12th, 2019 Time:  9 AM - 10 AM PST

Register here

Open sourcing wav2letter++, the fastest state-of-the-art speech system, and flashlight, an ML library going native by/via via @fbOpenSource

The Facebook AI Research (FAIR) Speech team is sharing the first fully convolutional speech recognition system. It uses convolutional neural networks (CNNs) for acoustic modeling and language modeling, and is reproducible. The team says that wav2letter++ is composed only of convolutional layers, which yields performance that’s competitive with recurrent architectures.

There are two articles linked of the landing page from the links in this post. This reads as a great achievement and looks very interesting.

Activation Regularisation for Reducing Generalisation Error in Deep Learning Neural Networks by/via @TeachTheMachine

This tutorial on activation regularisation for reducing generalisation error in deep learning neural networks will help you create better-learned representations and improve predictive models that make use of the learned features.

This is great and I recommend a bookmark to Jason's website as well as subscribing there. Everything he does and explains is very clear and easy to understand.

Comparing the performance of machine learning models and algorithms using statistical tests and nested cross-validation by/via @rasbt

Sebastian Raschka compares the performance of machine learning models and algorithms using statistical tests and nested cross-validation.

This blog is great and very much worth a bookmark.  Go and look through the entire series of articles - this is useful bot both those new to data science and those who are experienced too.