A Guide Towards Responsible Artificial Intelligence
As the AI industry continues to mature and more companies begin to adopt this powerful technology, it is important for professionals to start considering how to build their AI journey in a responsible manner.
Responsible AI covers some of the following areas: explainable AI, ethical AI, risk in AI, and more. All of these areas can have a major impact on businesses who are starting their AI journey.
Join Harib Bakhshi, Lead Data Scientist for UK and Ireland at H2O.ai, on May 14th, who will walk you through some of the best practices around Responsible AI.
As part of this webinar, Harib will cover:
- A short overview of why Responsible AI is an important aspect to take into account
- The business benefits of implementing Responsible AI
- Some practical tips to consider whilst going through your Responsible AI journey
RecordedMay 14 202054 mins
Your place is confirmed, we'll send you email reminders
In this webinar, we start off with an introduction to the basics of Machine Learning and how it is being applied to a variety of industry use cases. We will then provide overviews of H2O-3, the #1 open source machine learning platform for the enterprise and H2O Driverless AI, our award-winning automatic machine learning platform. In addition to discussing H2O-3 and Driverless AI, we will highlight Sparkling Water (machine learning pipelines on Spark).
We’ll walk you through demos of both H2O-3 and Driverless AI and highlight features and capabilities for each platform.
What you will learn:
- Introduction to machine learning
- Overviews of the features and capabilities of H2O-3, Driverless AI, and Sparkling Water
- Live demos of both H2O-3 and Driverless AI applied to real world datasets
Presenter:
Megan Kurka, Customer Data Scientist at H2O.ai
Every company can be an artificial intelligence (AI) company. Machine Learning is a specific subset of AI that has exploded the applications and adoption of AI but many times has required special skills. In this session learn about the basics of ML and how Automatic ML is making these tools accessible to a wider community of people.
Presenter: Rafael Coss, Director of Technical Marketing, H2O.ai
Featured Speaker: Brandon Purcell, Principal Analyst at Forrester Research & Ben Cox, Director of Product Marketing at H2O.ai
In tandem with the rapid adoption and excitement around Artificial Intelligence & Machine Learning, there has been a sub-sector of data science that has been growing at a rapid rate, Responsible AI. One of the early criticisms of machine learning models was that they were inherently "black-box" or too opaque for users to understand why models were coming to the conclusions that they were. The goal of Responsible AI has been to create new more transparent methods, develop post-hoc explanations for ML models, and define frameworks that will minimize potential risk or lack of understanding in AI systems.
This field has been growing rapidly and there are many different ideas from practitioners, in this webinar, Benjamin Cox of H2O.ai and featured speaker, Brandon Purcell of Forrester, will discuss their current views on the state of Responsible AI in 2020.
Presenters:
Featured speaker: Brandon Purcell, Principal Analyst - Customer Insights Professionals at Forrester Research
Ellen Friedman PhD, Technical Evangelist at H2O.ai
What makes the biggest difference to AI success? New and better algorithms? More trained data scientists? More data?
All these factors are important, but it turns out that the thing that makes the biggest difference to AI success is this: aiming AI at the right business goal. To make that happen, it’s necessary to have an effective conversation between data scientists and non-data scientists who have expert domain knowledge – and the biggest group of domain experts turn out to be business people, with the specific and valuable understanding of how their business works. But can domain experts talk to data scientists, and can data scientists communicate with them?
In this webinar you will learn:
- Why the conversation between data scientists and non-data scientists matters
- How to build the skills necessary to have that conversation
- Examples from real world use cases about putting AI into the right business context
Presenter:
Ellen Friedman PhD, Technical Evangelist at H2O.ai
Nick Patience, Co-Founder & Research Director, AI Apps & Platforms, 451 Research & Ben Cox, Dir. of Product Marketing, H2O.ai
With the rapid adoption of machine learning and artificial intelligence, there has been a subsequent desire for increased transparency and explainability of these models. But Explainable AI is not just focused on risk mitigation, it aims to drive deeper business value through enriched dynamic data about what your models are thinking. In this webinar we will talk through the current state of Explainable AI and how it can be leveraged to enable companies to make better decisions and increase understanding of their customers.
Presenters:
Nick Patience
Co-Founder & Research Director, AI Apps & Platforms
451 Research, part of S&P Global Market Intelligence
Akhil Sood, Associate Director, Allergan, Dr. Michael Proksch, Sr. Director, H2O.ai, Vijay Raghavan, Associate VP, Allergan
Together with the marketing sciences team at Allergan, H2O.ai recently compared the traditional linear marketing mix with the new machine learning based approach to experience the differences firsthand using H2O.ai’s Driverless AI - AutoML solution. Not only were they able to build a model within hours for what took weeks before, they also did not need to make assumptions as they did with a linear modeling approach. Learn more about how the latest in AI technologies reinvented media and marketing analytics at Allergan by joining us for this informative webinar.
In this webinar, you will learn:
- Traditional linear marketing mix models vs. new algorithms
- Overcoming the weaknesses of linear marketing mix models
- How Allergan is using it to optimize their marketing efforts
Presenters:
Akhil Sood, Associate Director at Marketing Sciences at Allergan
Dr. Michael Proksch, Senior Director at H2O.ai
Vijay Raghavan, Associate Vice President at Marketing Sciences at Allergan
Niki Athanasiadou MRes, PhD, Customer Data Scientist at H2O.ai
H2O.ai brings the best practices of the world’s leading data scientists in the form of comprehensive solutions to specific problems. With the digitization of primary healthcare information almost complete, many hospitals are looking to leverage the investment and maximize the returns of this gargantuan effort that took place over the past two decades. Supporting clinicians in the diagnosis of sepsis through AI-backed clinical decision support tools makes value-based care a reality in the intensive care unit.
In this webinar, you will learn about:
- The cost of hospital-acquired sepsis on patient outcomes and the burden it imposes on the healthcare system
- Building predictive models for Sepsis risk
- Leveraging AI for clinical decision support in the intensive care unit
Presenter:
Niki Athanasiadou MRes, PhD
Customer Data Scientist at H2O.ai
Ben Cox, Dir. Product Marketing, H2O.ai, Navdeep Gill, Lead Data Scientist, H2O.ai, Patrick Hall, Advisory Consultant, H2O.ai
Like other powerful technologies, AI and machine learning present significant opportunities. To reap the full benefits of ML, organizations must also mitigate the considerable risks it presents. This webinar outlines a set of actionable best practices for people, processes, and technology that can enable organizations to innovate with ML in a responsible manner. The goal is to promote human safety in ML practices so that in the near future, there will be no need to differentiate between the general practice and the responsible practice of ML.
In this webinar, you will learn:
- Why an organization’s ML culture is an important aspect of responsible ML practice
- Suggestions for changing or updating your processes to govern ML assets
- Tools that can help organizations build human trust and understanding into their ML systems
- Core considerations for companies that want to drive value from ML
Presenters:
Ben Cox, Director of Product Marketing at H2O.ai
Navdeep Gill, Lead Data Scientist at H2O.ai
Patrick Hall, Advisory Consultant at H2O.ai
Karthik Guruswamy, Sr. Principal Solutions Architect at H2O.ai
A business has a product or service that attracts customers and generates revenue streams. Everything is great for a while. Now the business sees customers are churning. Business analysts and domain experts in your company are pouring over KPIs and doing surveys to identify a few trends, yet cannot quite pinpoint the mix of causes for every customer churned. Most importantly, marketing does not have a rank ordered list of potential churners next month and the reasons why they would leave - to effectively craft relevant and personalized campaigns, offers etc., and bring them back in!
If this sounds like your business, this is where building high accurate AI/ML models with explainability can help greatly. Churn is a common issue. It can happen to a business for a variety of reasons - Pricing, Competition, Quality of product, Service, etc. This webinar is about building AI/ML Churn models and explaining the reasons for churn at a customer level.
In this webinar, you will learn:
- How to define churn in your business - Hard churn vs Soft Churn
- How to structure the problem with the historical data you have
- How to create training, test and holdout sets for building models
- How to build AI/ML models using Driverless AI in a few min
- Generate rank ordered churners with reason codes.
- Identify areas (Prescription) quantitatively in your business for improving customer retention
Presenter:
Karthik Guruswamy, Sr. Principal Solutions Architect at H2O.ai
Gregory Kanevsky, Principal Solutions Architect at H2O.ai
Time series make up a growing share of real-world use cases across many industries. In this webinar, we explain various applications of H2O Driverless AI, which enables businesses to automate development and deployment of state of the art AI/ML forecasting and signal processing models in record times and at scale.
In this webinar, you will learn:
- Types of time series data and problems
- H2O.ai's approach to automating time series problems using the Driverless AI solution extensibility and customization and other tips to solve wider variety of problems and achieve best results
- A live demo of Driverless AI showcasing problem setup, automated ML pipeline and model deployment
Presenter:
Gregory Kanevsky, Principal Solutions Architect at H2O.ai
Tom Ott, Senior Customer Solutions Engineer at H2O.ai
Congratulations on your new role as a Data Scientist! A rewarding career as a Data Scientist goes beyond just coding in Python and R. It’s not a Venn Diagram, rather it’s your first step as an analytic professional in an ever changing environment. Success as a Data Scientist is less about getting the best AUC but creatively solving business problems.
In this webinar, you will learn:
- Are Data Scientists Born or Made? Is there such a thing as a classically trained Data Scientist?
- Which is better? Coding or No Code? Weighing the speed of No Code environments vs Coding from scratch
- Open Source vs Closed Source libraries. Which is better?
- Aligning your work to a business problem. Understanding how your Data Science work affects business outcomes.
Presenter:
Tom Ott, Senior Customer Solutions Engineer at H2O.ai
Ashrith Barthur, Principal Security Scientist at H2O.ai
Electronic Fraud is prevalent in almost every walk of life these days. The directions in which society is moving forward, monetary instruments are only going to get more digital, and transactions are only going to get more electronic. In this almost-exponential growth fraudsters have a leg up. This is because legacy systems that are fighting are old, and have not accounted for newer fraudulent behaviors. While the new systems with ML models could be accurate but slow. For one to catch fraud in an acceptable time, the systems have to be fast and quickly modifiable to changing fraudulent methods.
In this talk, we speak about different methods which can make your AI systems faster, and valuable toward identifying fraud. These systems also maintain a high level of accuracy.
Some of these methods that we would discuss are:
- Different ways of implementing the models
- Variations in hyper-parameters of models
- Highly accurate features that are valuable that are modifiable
At the end of this webinar you would be able to understand how to:
- Build better features for fraud
- How to build models and model implementations to speed up the decision
Presenter:
- Ashrith Barthur, Principal Security Scientist at H2O.ai
Elena Boiarskaia, Senior Solutions Engineer at H2O.ai
H2O-3 is an open source, in-memory, distributed machine learning platform that is optimized to build machine learning models on big data and easily deploy them in an enterprise environment with a MOJO. Spark is a powerful distributed cluster-computing framework for running large-scale data processing workloads. Sparkling Water combines the best of both worlds, by seamlessly integrating the H2O-3 ML library to run on top of Spark for building fast and accurate predictive models on big data at scale.
In this webinar, you will learn about:
- Leveraging the power of H2O-3 and Spark to build scalable machine learning models
- Embedding Sparkling Water models inside SparkML pipelines
- End-to-end Sparkling Water use cases from data preparation to model deployment
Gregory Keys PhD, Senior Solutions Architect at H2O.ai
The H2O Open Source Machine Learning Platform (H2O-3) empowers data scientists to train ML models at massive scale (GBs to TBs of training data) using familiar languages and IDEs on existing or greenfield distributed compute environments. Data scientists export a standardized scoring artifact of a trained model and dev-ops teams deploy these as low latency prediction software to diverse production systems (Rest endpoint, RDBMS, Kafka, etc) all via existing SDLC processes. Numerous top companies across verticals have leveraged the power and simplicity of H2O-3 to become innovative AI companies while adhering to strict enterprise security and governance provided by the platform. Let’s learn how.
In this webinar, you will learn:
- The value of H2O Open Source from the data scientist’s perspective
- The value of H2O Open Source from the dev-ops and business owner perspective
- Technical architecture of H2O Open Source ML Platform
- Enterprise choices and best practices in implementing H2O Open Source on distributed compute for model building
- Enterprise choices and best practices in deploying trained models to diverse production software environments
- Enterprise security and governance controls of H2O Open Source
Presenter: Gregory Keys PhD, Senior Solutions Architect at H2O.ai
AutoML is helping lower the barrier to entry and accelerate the time it takes to build and deploy machine learning models. H2O's Driverless AI and Open Source library both feature AutoML capabilities that range from hyperparameter selection to advanced feature engineering and model ensembling. However, it is important to note that AutoML is not intended to replace expert data science teams but can instead be used by such teams to augment their scalability and efficiency while still allowing them to maintain control over the end to end machine learning lifecycle.
In this webinar, you will learn:
- Evolution of AutoML
- How expert data science teams can leverage AutoML
- How H2O Driverless AI allows expert Data Science teams to maintain control when using AutoML
Presenter:
Travis Couture, Solutions Engineer at H2O.ai
Vinod Iyengar, VP Customer Success and Product at H2O.ai and Rafael Coss, Community and Partner Maker at H2O.ai
AI is a central part of every CIO’s strategy and in 2020, we’re seeing every enterprise looking to use data and AI to run business. An AI-driven enterprise is not only building machine learning models but infusing and integrating AI across every application in the enterprise. This requires internal and external applications to be AI-ready and agile to the demands of the enterprise and evolving machine learning models.
In this webinar, you will learn:
- The different stages of maturity as it relates to data science and analytics in an organization and consequently their AI maturity itself.
- How an enterprise can move from just using data science to improve operational efficiency to actually creating new lines of business or revenue generation opportunities using their data assets.
- How all of this can lead to more data, better models and ultimately help you build a really defensible moat against your competitors while continuing to delight your own customers to improve your margins.
- What components are required to build this AI-aware organization and how to get there as soon as possible.
- Demo of a modern AI application and review some of the key elements that make them native and agile.
Presenters:
Vinod Iyengar, VP Customer Success and Product at H2O.ai
Rafael Coss, Community and Partner Maker at H2O.ai
Erin LeDell, Chief Machine Learning Scientist at H2O.ai
Automatic Machine Learning (AutoML) is a subfield of machine learning which aims to automate the training & tuning of machine learning models. One of the main goals of an AutoML tool is to train the “best” model possible in the least amount of computation time, with zero/minimal configuration by the user. AutoML tools reduce the expertise required for practitioners to train powerful machine learning models, which has expanded and accelerated the application of machine learning to problems in both academic research and industry. AutoML greatly speeds up the workflow and efficiency of even the most experienced data scientist.
As automation and use of machine learning increases, in particular with the proliferation of open source AutoML tools, there’s an increased risk in misuse of, or harm by, machine learning models used in real world applications. In order to reduce the risk of harmful models being deployed, machine learning tools, and especially AutoML tools, can offer easy-to-use or automated interpretability and algorithmic fairness methods that can be used to evaluate and probe machine learning models. Interpretability and fairness methods should always be applied to machine learning models before they are deployed into production where they can make or influence important decisions affecting people’s lives.
In this session, you will learn about:
- Automated Machine Learning and open source H2O AutoML
- Interpretability methods for H2O models
- Algorithmic fairness (disparate impact) for H2O models
- Demo using U.S. Home Mortgage Disclosure Act (HMDA) data
Patrick Hall, Advisory Consultant at H2O.ai and Michael Proksch, Senior Director, Customer Data Science at H2O.ai
Insurance and credit lending are highly regulated industries that have relied heavily on mathematical modeling for decades. In order to provide explainable results for their models, data scientists and statisticians in both industries relied heavily on generalized linear models (GLMs). However, new machine learning algorithms like GBMs are not only more sophisticated estimators of risk, but due to a Nobel-laureate breakthrough known as Shapley values, they are now seemingly just as interpretable as traditional GLMs. More nuanced risk estimation means less payouts and write-offs for policy and credit issuers, but it also means a broader group of customers can participate in mainstream insurance and credit markets.
In this webinar, you will learn about:
- The Advantages of Machine Learning vs. Linear Models
- Why you should think about the AI shift in perspective
- How to move to new ML Methods (e.g. GBM)
Presenters:
Patrick Hall, Advisory Consultant at H2O.ai
Michael Proksch, Senior Director, Customer Data Science at H2O.ai
Benjamin Cox, Director of Product Marketing at H2O.ai
With the latest release of H2O Driverless AI (1.9.0), we have added a litany of new features to enhance the user experience and empower companies to build models in the most responsible and transparent manner. With the addition of multiple fairness metrics such as, Disparate Impact Analysis, and leading edge explainable modeling methods such as Explainable Neural Networks (XNN) and GA2M, Driverless AI users are equipped to further explore model explainability techniques within the platform.
In this webinar, you will learn about:
- Disparate Impact Analysis and Standard Mean Difference
- Exporting Decision tree model rules as txt & kernel explainer for Shapley Values
- XNNs & GA2M
Presenter:
Benjamin Cox, Director of Product Marketing at H2O.ai
SRK, Kaggle Grandmaster/Data Scientist at H2O, Max Jeblick, Data Scientist, H2O, and Trushant Kalyanpur, Data Scientist, H2O
H2O Driverless AI brings the best practices of the world’s leading data scientists to your team to build high-quality production-ready models in hours, not weeks or months. Driverless AI users can now use state-of-the-art contextual pretrained language models for their text related datasets. Advanced or novice data scientists can build models like BERT, DistilBERT, XLNET, Roberta with the power of full Driverless AI automation.
In this webinar, you will learn about:
- NLP features in Driverless AI 1.9
- Demo of how to use BERT like models as modeling algorithms or for feature transformation
- Custom BERT recipes for domain specific problems
Presenters:
Sudalai Rajkumar (SRK), Kaggle Grandmaster and Data Scientist at H2O.ai
Maximilian Jeblick, Kaggle Master and Data Scientist at H2O.ai
Trushant Kalyanpur, Data Scientist at H2O.ai
H2O.ai is the maker of H2O, the world's best machine learning platform and Driverless AI, which automates machine learning. H2O is used by over 200,000 data scientists and more than 18,000 organizations globally. H2O Driverless AI does auto feature engineering and can achieve 40x speed-ups on GPUs.
A Guide Towards Responsible Artificial IntelligenceHarib Bakhshi, Lead Data Scientist for UK and Ireland, H2O.ai[[ webcastStartDate * 1000 | amDateFormat: 'MMM D YYYY h:mm a' ]]53 mins