Machine Learning

Thursday, August 19, 2021

- EDT
Practical Machine Learning 👨🏽‍🔬
Join on Crowdcast
Atif Farid Mohammad PhD
Atif Farid Mohammad PhD
UNC Charlotte, Artificial Intelligence Professor

Workshop Overview:

The use of Machine Learning in the arena of Social Determinants of Health.

Intended Audience:

  • Data Scientists
  • Machine Learning Engineers

Topics Covered:

  • Python
  • Data Acquisition
  • Feature Extraction and Extrapolation
  • Machine Learning Model Design

Workshop Takeaways:

  • Comprehend the outcomes of Machine Learning Models.

To-Do Before Workshop / How to Prepare:

  • Anaconda, Python installed on participant's laptops





- EDT
Building ML Pipelines: Automating Parameter Search 🔍
Join on Crowdcast
Avinash Gopal
Avinash Gopal
Metabob, CTO

Workshop Overview:

The workshop will begin by discussing Model Tuning - a process by which developers fine tune their search with overlifting the variance. Model Tuning acts as the “dials” and “knobs” of the Machine Learning process, and is the critical first step to achieve proper Automated Training of a given machine. From here, the workshop will show attendees how to then take their newly tuned and ready to function automated training system, and implement it into the proper ML Pipeline. There are various pipelines given the needs of one’s given task, and we the workshop will go about explaining what types are preferred for certain situations over others.

Intended Audience**:

  • New and experienced data scientists and engineers.
  • Interested in learning how to implement automated training and ML pipelines
  • Note** Workshop is intended for individuals with an intermediate level proficiency with ML

Topics Covered:

  • Model Tuning / Hyperparamter Optimization 
  • Automated Training
  • Building ML pipelines

Workshop Takeaways:

  • Determining when to start model optimization
  • Which search methodology is effective for your domain
  • Configuring stop conditions and validation checks
  • Automating parameter searching in a scalable way

To-Do Before Workshop / How to Prepare:

No advanced preparation necessary

- EDT
Improving Cyber Threat Detection with Machine Learning, Visualizations and Graph Analytics
Join on Crowdcast
David Braun
David Braun
TigerGraph, Sr. Solutions Architect

Key Takeaways:

The sophistication of cyber criminals is increasing relentlessly. Accenture found that 68% of business leaders feel their cybersecurity risks are increasing. More and better technologies are required to detect attacks and prevent them, we’ll discuss: 

  •  How graph analytics, machine learning, and visualizations, can directly assist in the identification of threats in your environment. 
  • Using the same approach as many other security tools, we examine how TigerGraph can help you identify threats earlier along the kill chain of the MITRE Attack Framework.


Friday, August 20, 2021

- EDT
Cyber Situational Awareness and Resilience 💥
Join on Crowdcast
Deepinder Uppal
Deepinder Uppal
TIBCO, Chief Architect, Federal / Director, Vertical Innovations

The rapid digitization of government services has led to a dramatic increase in the number of cyber incidents. As governments require more flexibility in data sharing, federated data access and edge analytics, governments become greater targets for cyber intrusion. 

In essence, governments require cyber situational awareness and resiliency on both ends of the firewall, and in this way need to have the capabilities to sense, resist and react to disruptive cyber events -- and to recover from incidents in a timely fashion. 

One such technique is to merge cyber open source intelligence (OSI) with the ability to proactively source network anomalies thereby providing actionable intelligence to SOC analysts and threat hunters; leveraging machine learning (ML) and behavior recognition techniques empowers personnel to get ahead of potential incidents and delivers the missing context to what is happening on the network. This technology when combined with OSI allows for immediate deconfliction of known threats while alerting to those that may be unknown.

- EDT
1001 Things To Do With Your Data to Achieve Analytics Mastery 🏆
Join on Crowdcast
Kirk Borne, PhD
Kirk Borne, PhD
DataPrime Inc., Data Science Consulting

This presentation will demonstrate the vast number of analytics opportunities for businesses who have access to many different data sources. Analytics strategies will be presented that will accelerate return on data investments, including predictive, prescriptive, precursor, and sentinel analytics, plus an observability strategy for edge analytics with IoT (Internet of Things), insights-as-a-service, and the STELLAR Analytics Scorecard.

- EDT
An Introduction to Drifter-ML 🤖
Join on Crowdcast
Eric Schles
Eric Schles
Johns Hopkins University Hospital, Principal Data Scientist

Drifter-ML is a novel framework for testing machine learning models. In this talk you will learn the basic idea behind the framework from its creator.