Spiros Potamitis (@potamitisspiros) 's Twitter Profile
Spiros Potamitis

@potamitisspiros

Data Scientist | Global Product Marketing Manager | Guest lecturer | Loves snowboarding & cooking

ID: 1352600245574787072

linkhttp://www.linkedin.com/in/spirospotamitis calendar_today22-01-2021 12:53:30

83 Tweet

54 Takipçi

79 Takip Edilen

Spiros Potamitis (@potamitisspiros) 's Twitter Profile Photo

A2. #Fairness in #AI comes from unbiased decisions. #Trustworthiness means that the decisions produced are explainable and repeatable. #saschat #trustworthyAI #analytics

Spiros Potamitis (@potamitisspiros) 's Twitter Profile Photo

A3. Collecting data from #unbiased processes to build #AI systems is the biggest challenge for organizations I’ve seen. #trustworthyAI #analytics #saschat

Spiros Potamitis (@potamitisspiros) 's Twitter Profile Photo

A4. Organizations try to build #diverse and #inclusive teams so AI systems are as fair as they could be, by design. They also adopt tools to identify #bias in the development phase and correct #bias that is observed while monitoring ai decisions through time. #saschat

Spiros Potamitis (@potamitisspiros) 's Twitter Profile Photo

A5. Hire the right people and adopt the right tools and processes. An #AI #Ethics team is essential in every organization to ensure the above. #saschat #trustworthyAI #analytics

SAS Software Cares (@sas_cares) 's Twitter Profile Photo

This week's #SASchat explores the dynamics of analytics, productivity and innovation. Join us on Feb. 10. 2.sas.com/60143YAhk 2.sas.com/60173YAhb 2.sas.com/60103YAhe

Spiros Potamitis (@potamitisspiros) 's Twitter Profile Photo

A1. Speed and scalability in the cloud are among the top priorities. Automatic data exploration and insights along with embedded best practices and a no/low-code environment so everyone can participate and contribute also help a lot. #SASchat #analytics

Spiros Potamitis (@potamitisspiros) 's Twitter Profile Photo

A2. Time is money in the cloud. The faster you run analytics workloads, the less you spend on infrastructure. Even if you want to maintain the same spend, running more analytics workloads lead to more accurate and robust results - hence better decisions #SASchat #analytics

Spiros Potamitis (@potamitisspiros) 's Twitter Profile Photo

A3. First, time to value is accelerated as you don’t have to manually code everything. Second, everyone can get in on the fun. Data science programs that focus on professional data scientists ignore the vast majority of people and business opportunities. #SASchat #analytics

Spiros Potamitis (@potamitisspiros) 's Twitter Profile Photo

A4. Organizations are conscious of cloud costs and the pressure is on data scientists to deliver the best results in the least amount of time. Without efficient tooling in place such as #SASViya, companies choose to cut on experimentation to drive down costs. #SASchat #analytics

Spiros Potamitis (@potamitisspiros) 's Twitter Profile Photo

A5. Organizations without a cloud optimization process tend to overspend by 40%, due to unmanaged costs, unexpected usage and suboptimal design.Algorithmic optimization can help tackle those issues by providing optimal performance and scalability in the cloud. #SASchat #analytics

Franklin Manchester (@fjmanchester) 's Twitter Profile Photo

There's powerful tools out there... make sure whatever you choose to wield, it's done so responsibly. AI can and will do great things for humanity. For more, check out this asset shared by our friend Spiros Potamitis sas.com/en_gb/whitepap… #SASchat #analytics

Blake Sheldon (@blaketsheldon) 's Twitter Profile Photo

SAS Forums Don't forget the reason for analytics is to achieve value for your business. Over-rotating on the means to the end is wasteful. Focus on what makes your problem-owners become problem-solvers with the least amount of technology & people in process -- that's just friction #saschat

Spiros Potamitis (@potamitisspiros) 's Twitter Profile Photo

A1: Like any other project it needs to have a beginning, an end date and important milestones. It’s very important also to have clear expectations from the projects’ sponsors about expected benefit and flexibility for models’ refinements or methodology reviews. #saschat

Spiros Potamitis (@potamitisspiros) 's Twitter Profile Photo

A2: It’s critical to take your stakeholders on a journey with you and communicate challenges as you go along. You should also make efforts to explain in simple way the techniques that are used to gain the trust you need to put your project into production. #DataScience #saschat

Spiros Potamitis (@potamitisspiros) 's Twitter Profile Photo

A3: Because of their iterative and probabilistic nature of ds projects. You don’t know how good a model will be unless you start collecting data and try different techniques. In such a way the expected benefits may differ from initial requirements #saschat #datascience

Spiros Potamitis (@potamitisspiros) 's Twitter Profile Photo

A4: Information knowledge management should be a key course at every university that teaches #datascience. Learning the techniques alone won’t be adequate for a successful career in in the field #saschat

Spiros Potamitis (@potamitisspiros) 's Twitter Profile Photo

A5: Yes but would have to be adapted to the needs of a #datascience project. A project manager should be knowledgeable about the extra limitations that are introduced when dealing with such projects and be able to communicate that clearly to projects’ sponsors. #saschat