Watch this Ask the Expert session to learn how to elevate your ModelOps strategy by integrating data more agilely and effectively.    Watch the Webinar   You will: Learn how to accelerate ModelOps strategies using data. Understand the influence of new data trends and methodologies on scalable ModelOps. Explore how adopting innovative data strategies can transform your organization, leading to improved ModelOps through better data quality, reliability and accessibility, ultimately driving significant outcomes.   The questions from the Q&A segment held at the end of the webinar are listed below and the slides from the webinar are attached.   Q&A What percentage of Data Scientists have direct access to production data for model development? That's an interesting question. Assume none. I see a lot of people on the Internet all the time saying like, hey, for me to do my job, I need access to production data. Cool. And I agree, you do need access to representative data. You need to know what your data looks like so you can model it and derive value from it. But, a lot of times, in our customer spaces, we see personally identifiable information and other sensitive information. A good AI platform provides flexibility and lets you closely mimic production conditions within testing environments. Enabling data scientists to leverage parity between their modeling experiences and the production ecosystems where their models will be deployed. That's really what it's all about. In most organizations, it's hard to provide that type of risk, i.e., letting people model against highly sensitive data and personal information.   Is there a reliable protocol to validate the trustworthiness of the data models generated from AI? Jean: It is the application of surrogate models’ proxies to play a role in identifying whether the response is aligned to what is deemed correct. But I don't know if there's a direct way of doing that at the moment. Luis: It's a moving target. Now, more than ever, things are moving really fast and different organizations and sectors are more concerned with trustworthiness. From what I understand, GDPR had been further ahead of the US regulation of AI. I know there's a risk management framework that's been put out by the National Institute of Standards and Technology (NIST). There's some basic guidance there, although legislation is currently en route. SAS has responded.   We have a trustworthy AI workflow at the model level. Data Scientists in SAS Viya can create trustworthy AI workflows with checks and measures, where you notify stakeholders, when guardrails aren’t met. Keeping AI in check in a human centric way to try to enforce trustworthiness. Keeping AI in check in a human centric way to that doesn’t compromise innovation. Also, we have the model cards. We can quantify bias and trustworthiness of the individual model. We are crafting compelling capabilities that keep up with regulation as it comes online. When people choose a technology vendor, they'll go with someone who’s values are aligned with theirs.   Is Machine Learning the end goal of all data strategies? Jean: It's definitely a goal. I don't know if it's the end goal. I think the end goal should always be the decision and how it gets applied and used. Also very important, the feedback loop to measure how you perform against what you set out as a target. So, it's definitely an outcome, but it shouldn't be the sole focus. I think there should be value. Otherwise, you are still in experimentation phase. Luis: One of the things I wanted to amplify is that all users matter. You strike a balance between feeding predictive capabilities or emerging technologies as well as providing access to your users, so they could leverage data in a way that adds value to the business. Is that always a predictive model? Is that always something that leverages the latest language model? No. Sometimes, like we articulated in a little workflow, we have 32200 users that are thirsty for data, and they leverage data their way. From a data engineering standpoint, why not serve them as part of a plan that could be comprehensive? So, machine learning is a goal, but probably not the only goal, and shouldn't be the only goal depending on where you're coming from.   Can you explain the role of data quality in MLOps? It’s very important – the data that feeds the models, that models get trained on, as well as being scored on. Models are only as good as the data that goes into it.   Are there any programs for university students who want to master data analytics? I am a senior majoring in Statistics. To learn more about SAS, as a student look at the resources here for students: https://www.sas.com/en_us/learn/academic-programs/students.html   Below is a sampling of classes to take that can give you a flavor of data management capabilities and start exploring data analytics within SAS Viya.   Data Management Flavored Courses DATA MANAGEMENT with SAS VIYA Course Writing a custom task for SAS Studio Free Using SAS Studio Flows and Custom Steps in SAS Viya Using SAS Data Preparation in SAS Viya   Analytics Flavored Courses Leading with Analytics Explore and Visualize: Getting Started   And Rising Energy Costs? While I do not see any courses for rising energy costs, below is the SAS Statement and corporate social responsibility resource link as it pertains to sustainability.   “A sustainable future requires developing solutions grounded in science and data to address climate change mitigation and adaptation. Reducing environmental impacts and ensuring continued availability of natural resources is a shared responsibility that starts with intentional and ambitious goals and actions.” https://www.sas.com/content/dam/SAS/documents/corporate-collateral/brochures/en-csr-environmental-program-110853.pdf   Will this ever be a subject in schools? E.g., High School, etc. Applied statistics, Applied Mathematics, computer science, data science, information technology, engineering and a myriad of other stem programs would be beneficial towards starting a career in AI, ML , and other software related job fields. It depends on the academic institution but many of them do teach and mirror software capabilities and concepts as part of their curriculum.   How do you measure models behaving as the and its results vs. changing/ regulatory and other macroeconomic needs? At SAS, we engineer our software to protect your data and your business. The SAS® Software Security Framework incorporates industry best practices and defines the guiding principles for our secure product development life cycle. From engineering all the way through vulnerability remediation, we are committed to ensuring that our products continually meet the business and security needs of our customers.  https://www.sas.com/en_us/company-information/security.html   Back half of the question, SAS is at the forefront of keeping pace with changing regulatory requirements across a variety of industries from public sector to health care, and financial services companies. Additionally, SAS provides AI governance advisory offerings from designed to help you institute AI governance in your organization and reap the benefits of responsible innovation.   SAS has been at the forefront of responsible AI and is proactively collaborating at the highest levels of governments across the globe to advise and steer guardrails to ensure privacy, security, and regulatory standards are developed.   From a modeling perspective, SAS Model manager enables Trustworthy AI Life Cycle workflow that reflects some of the standards and best practices set by the AI Risk Management Framework defined by the National Institute of Standards and Technology (NIST).  This is an initial step towards driving our products alignment to the blossoming AI regulations we see across the globe.      Recommended Resources ModelOps ebook Please see additional resources in the attached slide deck.   Want more tips? Be sure to subscribe to the Ask the Expert board to receive follow up Q&A, slides and recordings from other SAS Ask the Expert webinars. ... View more