Applied AI Live Breakout Sessions
Upcoming Session
Past Session
Past Session
Past Session
WORKSHOP, 90 Minutes
WORKSHOP, 90 Minutes
WORKSHOP, 90 Minutes
TALK, 45 Minutes
AI Crash Course – Adding ML to Your Projects in an Hour
Introduction to MLOps: The Concepts and Strategies that Get Models Reliably into Production
Deep Learning in Ten Minutes or Less with AutoML
The Art of Storytelling for AI and Machine Learning
Current Speakesr: Zan Markan and Peter Klipfel
Current Speaker: Seph Mard
Current Speakesr: Emily Webber and Ben Taylor
Current Speaker: Jen Underwood
Missed Applied AI?
See Our Next Lineup of VIrtual Events
Workshop, 90 minutes, Watch now!
During this session we will be learning how to architect, instrument, deploy, and manage AI models of various types in production environments. Example code will be provided. It is recommended that attendees have a background in or familiarity with DevOps concepts…more details
Workshop, 90 Minutes, Watch now!
In this introductory session, we will be covering the basics of making AI-powered applications, from designing a dataset to making predictions in-app. This session will be primarily focused on using REST APIs to build, deploy, and consume AI output. All samples and source code will be available on GitHub so you can follow along or experiment afterward. You will be encouraged to ask questions during and before the event!…more details
Talk, 45 Minutes, Watch now!
In today’s era of artificial intelligence (AI) and machine-assisted analytics, business analysts are crucial for bridging the growing data literacy gap. Successful analytical communicators don’t wait until the end of their analysis to communicate insights. Accurately defining projects, understanding what to data to use, preventing bias, and interpreting and effectively communicating findings are all important skills for helping stakeholders understand results and get the most actionable value from automated machine learning projects.In this session, Jen Underwood will walk through the best way to communicate the value of automated machine learning results with visualizations throughout the entire analytical process, from use case definition to insight implementation.You’ll learn:How to define a business use case for machine learning and AI storytellingThe process of planning, designing, and visualizing AI storiesHow to effectively translate quantitative insights and tell a compelling story throughout the complete analytical project lifecycle…more details
Workshop, 90 Minutes, Watch now!
AutoML is coming to deep learning. Traditional deep learning models would take data scientists weeks to code and tune. Learn how to take multimodal datasets, mixing of tabular and unstructured data (images, audio, video), and create accurate deep learning models in under 10 minutes with DataRobot. AutoML lets users have access to the latest frameworks like Keras, but with a push of a button be able to access transparent interpretability tools like feature impact, partial dependence, and prediction explanations. In this session, we will reveal some recent breakthroughs in deep learning and walk through some detailed examples from data to deployment…more details
Talk, 45 Minutes, Watch now!
When considering Intellectual property rights typically you think of patents, trademarks and the like. But what about the IP architected into your algorithms and AI decision pipelines? Beau Walker, a Data Scientist with a degree in Intellectual Property Law, has explored this concept his entire career. Together with Ben Taylor, PHD, they will summarize the basics and deep dive into the key factors you need to know to protect the IP inherent in your AI…more details
Talk, 45 Minutes, Watch now!
You know Kubernetes is a great platform for the applications you’re running today. Most of the applications you’ll be excited about tomorrow are intelligent applications, which collect data and rely on machine learning to support essential functionality. These capabilities often seem like magic to users, but building applications and services that leverage artificial intelligence is more accessible than you might think. This workshop will show how Kubernetes, the most popular open source container orchestration platform, can increase collaboration and decrease time to value for machine learning workflows. Attendees will be able to create workflows with ease, deploy models as micro-services and monitor performance to understand when retraining is needed. We’ll focus on the open source infrastructure, tools and processes that will help you to get meaningful results from application intelligence and show why Kubernetes is the best place for data science workloads. You’ll leave having solved a real business problem interactively with powerful machine learning techniques and Kubernetes…more details
Talk, 45 Minutes, Watch now!
Bias in machine learning is a significant concern as technology gets increasingly ubiquitous across many industries. Some types of bias can be attributed to limits in design and tooling; however, the bias in the training data itself is a general phenomenon. Skewed training data propagates into discriminatory AI models that amplify human prejudices.
Building a data labeling framework that uses a diverse set of crowd workers to collect and label the data can help reduce bias. Additionally, when you tap into a global crowd workforce you need to optimize the quality and speed of the labeling tasks, while at the same time follow ethical pricing practices so the crowd workforce is paid fair wages. This is a tough nut to crack.
In this talk, we present some of the frameworks and approaches to minimize bias and maintain a thriving community of highly engaged crowd workers. We will talk about:
- A bias minimizer framework that routes data labeling tasks to the right crowd worker and maintains a healthy worker distribution for a given task.
- An approach to ensure a fair wage for the crowd with location, skillsets, and task complexity considerations.
- Ways to increase crowd performance and engagement with smart targeting of labeling tasks to the crowd workers who are best suited for the job.
Workshop, 90 Minutes, Watch now!
In this workshop you will learn how to expand your Python data analysis skills to geospatial data.The workshop is aimed at software engineers, data scientists, and others who are interested in data science and data analysis.
During the workshop we will analyse UK Crime Data with Pandas and GeoPandas in a Jupyter notebook. We first will look at the properties of geospatial data and explore the different commands. After you have learned the basics we will go through some exercises analysing the UK Crime Data to explore patterns and trends and create a few maps of crime rates in London… more details
Talk, 45 Minutes, Watch now!
Much of the success of your models will depend on customer adoption. Setting the right expectations, communicating the results, introducing the idea of risk, and integrating your models in your customers day-to-day lives are just the tip of the iceberg. In this presentation, we will be discussing 5 practical and effective principles you can use to make your models as impactful and meaningful to your customers as possible. With these principles, your customers will keep coming for more! more details
Talk, 45 Minutes, Watch now!
Dask is a parallel computing library for Python people. This talk will be a gentle introduction to Dask, showing how you can improve the speed of data science code on your laptop with a simple “pip install”. Then we will use the same code to process big data on a cluster of machines. We will be going through an end-to-end data science pipeline, from ETL and exploratory analysis to machine learning model training and scoring… more details
Talk, 45 Minutes, Watch now!
Prior knowledge is believed to be informative to assist the understanding of natural language and the integration of prior knowledge with machine learning models has been found useful in various NLP tasks. The prior knowledge can be categorised into two categories. Structured knowledge explicitly defined by knowledge graph and more, while unstructured knowledge implicitly contained in large text corpus. Our research focuses on the effectiveness of integrating these two kinds of prior knowledge with machine learning models on text classification and summarisation…more details
Workshop, 90 Minutes, Watch now!
During this session we will be learning how to architect, instrument, deploy, and manage AI models of various types in production environments. Example code will be provided. It is recommended that attendees have a background in or familiarity with DevOps concepts.
Workshop, 90 Minutes, Watch now!
In this session, you will learn how to architect complex AI-powered applications and train and utilize many ML models to get working. This session will be primarily focused on using REST APIs to orchestrate the business logic of the application. All samples and source code will be available on GitHub so you can follow along or experiment afterward. You will be encouraged to ask questions during and before the event!..more details
Workshop, 90 Minutes, Watch now!
The stories of bias in AI are everywhere: Amazon’s recruiting tool, Apple’s credit card limits, Google’s facial recognition, and dozens more. The quick solution is just to blame the algorithm and its designers. However, as data scientists, its incumbent on us to understand the true source of the bias and improve the underlying process.
AI does not create bias alone; it exposes the latent bias present in the system it was designed to imitate. We need to reframe the conversation around bias in AI to instead identify it as the first step in building a more ethical system.
In this talk, we show how machine learning can make the implicit bias of a human institution explicit. Bias becomes diagnosable, correctable, and ultimately preventable in a way that cannot be replicated in human decision-making, which is opaque and difficult to change. Bias is not new, but AI represents a new toolset to measure and change it.
The goal is not only to provide you a theoretical understanding of bias, but a practical plan that you can start to implement right away. After all, it’s not whether or not you have bias in your institution, but how you plan to handle it…more details
Talk, 45 Minutes, Watch now!
The age-old adage “time is money” is perhaps more applicable today than during anytime in history. In this presentation, iMerit deep learning engineer Hrishikesh Hippalgaonkar talks about how a transformational, solutions-oriented approach to addressing data labeling problems will decrease Time-to-Value and increase ROI in the Artificial Intelligence and Machine Learning ecosystems. He will speak specifically of solutions mapping to custom workflows, real-time functionalities and predictive AI and how these solutions factor into the data labeling pipeline for AI and Machine Learning. He discusses the need for taking a holistic, solutions-based approach to AI deployments, and why partnering with organizations with vision and the ability to execute at scale are critical to AI success.
Tutorial, 45 Minutes, Watch now!
A Data Scientist sends you a model and it’s your job to use that model in production to power a business application. How do you approach this? Would you use the same approach for scoring 10 million records as you would for 10 thousand? How do you minimize cost when routinely scoring large volumes of data? These questions and more will be answered along with a few deep dive demos of real scoring techniques in this 45 minute power tutorial. more details
Talk, 30 Minutes, Watch now!
Audio is everywhere, but are you able to quickly and efficiently leverage this valued source of information? In this presentation, you’ll hear from Kensho, the S&P Global company powering its AI and machine learning capabilities, on how you can turn difficult-to-use audio into text using Kensho Scribe. Built with sophisticated deep learning models, Kensho Scribe is a leading transcription service for business and financial audio. You’ll learn how Scribe was built and see it live in action ― understanding how it can parse jargon in context and transcribe numbers, currencies, product names (e.g., pharmaceutical drugs).
Additionally, you’ll hear detailed use cases highlighting the efficiencies gained with Scribe – expanding earnings call coverage by 1,500 companies, transcribing 10,000+ voicemails per year for sales teams and automating compliance processes to meet regulations. Finally, you’ll gain an understanding of how you can pair Scribe with other natural language processing capabilities to build a pipeline that unlocks insights from your audio. more details
Talk, 45 Minutes, Watch now!
Learn about the development of a DIAGNOSIS TOOL, designed to help healthcare professionals in the fight against the COVID-19. We’ll cover everything from sourcing and cleaning the input dataset to the special techniques taken to combine a wide range of data types for VISUAL AI modeling. The modeling’s results are exported and loaded into Tableau and are used to build, along with recommendations from an (also implemented) Fuzzy-Ranking Content-Based Recommendation Engine, full-interactive dashboards that would provide actionable insight to support decision-making, gain quick knowledge in new cases, deeper understanding of anomalous cases, as well as support for resources management, and.. more details
.
ODSC West
October 27th – 30th, 2020
Data Science Virtual Training
Conference & Expo
