Author: admin

ICYDK: Machine Learning is guiding Artificial Intelligence capabilities.

Image Classification, Recommendation Systems, and AI in Gaming, are popular uses of Machine Learning capabilities in our everyday lives. If we breakdown machine learning further, we find that these 3 Machine Learning examples are powered by different types of machine learning:

* Image classification comes from Supervised Learning.
* Recommendation systems comes from Unsupervised Learning.
* Gaming AI comes from Reinforcement Learning.

How can we better understand Supervised, Unsupervised, and Reinforcement Learning?

Let’s start with Supervised Learning, which makes up most of the uses for Machine Learning today. In Supervised Learning, the machine already knows the output of the algorithm before it starts working on it. The algorithm is taught through a training data set that guides the machine, and the machine works out the steps from input to output. Supervised learning is used for image classification or identity fraud detection, and for weather forecasting. But how is Unsupervised Learning different?

Well first off, with Unsupervised Learning, the system does not have any concrete data sets, and the outcomes are also mostly unknown. Unsupervised Learning has the ability to interpret and find solutions to a limitless amount of data. Now when you log onto Hulu or Netflix, you have personalized recommendations because of Unsupervised Learning.

Lastly, there is Reinforcement Learning. Reinforcement Learning is different, because it gives a high degree of control to software agents and machines, which are determining what the behavior within a context should be. People are helping the machine to grow by maximizing performance, providing feedback to the machine, helping it to learn its behavior.

Reinforcement Learning requires the use of tons of different algorithms, giving control to the agent as they decide the best action based on the current results. When you are gaming on PC, Xbox, Playstation, or Nintendo, and you witness AI in Gaming, this is because of Reinforcement Learning. https://goo.gl/urCz5L #DataScience #Cloud

ICYDK: Sat across from the interviewer for your dream job, you may start to feel the pressure. A sure-fire way to quash the interview jitters is to prepare as much as possible. Typically, you can segment the types of questions you’ll get asked in a data science interview; things such as statistics, programming and technical ability, business acumen, and culture fit assessment. Studying up on these will help you prepare as best you can. 

Here are some examples of what you could expect when interviewing for a data science role. Tailor these in accordance with what the job description asks for, read it thoroughly and get clued up on the desired points!

“What ML techniques do you work with? / Are these research level or production level techniques?”

What techniques and knowledge are required for the role? Your experience should match up with what is being asked of in the description if you’re at interview level so make sure you go in with examples of your experience with these.

Try memorizing 3 different examples of where you have used particular techniques and the effect that they have had. For example, if the role requires convolutional neural network experience, prepare 3 examples of projects where you have worked with CNN and the impact they had on the business or research you’ve contributed to.

“Tell me about an in-depth example of projects you have worked on from inception to completion. What was the project, how did you approach the problem, what was the end result etc.”?

Be prepared to explain your experience and impact in granular detail!

-Why the project existed.

-Your part vs other people’s role.

– Provide a step by step walk-through of what you did, what tools and techniques you used.

-End product and what it meant to the business.

Know your own cv inside out, don’t be caught off guard by questions on experience or a project that you cannot dive into and explain thoroughly!

“What’s your favourite algorithm?”

This is a tough one and which algorithms and tools you use will be totally dependent on the job you’re working on. The best approach to a question like this, is to have an answer ready before going in, that is fitting to the role you’re going for rather than trying to think of a ‘favourite’, think of the most relevant and be able to talk about it – show that you’re able to make a decision (this is also what they could be trying to figure out!), and communicate your reasons for your choice, all the while framing it to what they will desire in a candidate.

“What level of experience do you have with [programming language]?  What do you do daily with [programming language] and what was your hardest challenges with this?”

This is a great way for interviewers to measure you up alongside other candidates in terms of technical ability. The programming language they will more than likely ask you about will have been named as a requirement in the job description so make sure you go in with your answer on this ready to go. Have an example up your sleeve and be able to frame your use of the programming language in terms of how you could use it similarly in this role. If you’re not well versed in what they’re asking for, be honest and show your willingness to learn.

“What is the largest data set that you have processed? How did you approach this, and what was the end result?”

Again, with questions like this, interviewers will be looking for a deep dive into your successes with processing large data sets, your understanding of the approach and techniques used, and how the results have benefited the company. Can you quantify your results in terms of costs, revenue and time saved? If you can, make sure these are front and centre in describing the impact you had.

 

There is, of course, no one size fits all when it comes to data science interviews, questions, and tasks but hopefully, this guide can go some way in helping you know what to expect broadly speaking.  https://goo.gl/EUeE9C #DataScience #Cloud

Despite growing four-fold since 2000, SAP emits less carbon dioxide today than it did 18 years ago – and still has advanced ambitions.

Released this week, the latest SAP Integrated Report sets it all out in black and white: In 2017, SAP cut its CO2 emissions by 55,000 tons year over year. That’s a whopping 14%. But even more important: At 325 kilotons, the company’s carbon footprint is once again below year-2000 levels. Back in 2009, SAP had set itself the goal of reducing its global greenhouse gas emissions to the year 2000 level by 2020 – despite strong company growth in the interim. A closer look at the per capita values reveals just how impressive that reduction is: In 2000, SAP recorded 13.9 tons CO2 emissions per employee. Today it’s only around 3.8 tons.

But this improvement doesn’t come by chance. It is the result of a dedicated climate policy that SAP adopted in 2009 and has been following ever since, and it underscores SAP’s commitment to the United Nations Sustainable Development Goal (SDG) 13: “Climate Action.” The policy is part of a holistic management approach that is aimed at harmonizing the economic, social, and environmental performance of a company instead of focusing exclusively on optimizing profit.

The SAP Integrated Report, the company’s “online first” financial report, showcases just how SAP is making that happen. Among other things, the report provides key figures on the financial and non-financial value add of the company and explains the key connections between SAP’s economic, social, and environmental performance.

Where to Invest Efforts?

“SAP’s key lever for a sustainable future is our product portfolio, with which we enable customers to create positive economic, environmental, and social impact. Leading by example and being a trustworthy role model has always been very important to SAP. Our transparent results in the report show that we continue to make great progress in all areas and are turning SAP’s vision and purpose into reality. I’d like to thank all employees for helping the world run better and improving peoples’ lives,” says SAP Chief Sustainability Officer Daniel Schmid.

While the focus is on innovations that help customers become more sustainable, SAP is just as committed to reducing its own greenhouse gas emissions, and has implemented a number of measures in order to achieve this. Currently, SAP’s sustainability strategy is based on the three pillars of avoid, reduce, and compensate:

* Avoid: Emissions are avoided wherever and whenever possible. Substituting business flights with video conferencing is just one example.
* Reduce: SAP has programs in place to increase efficiency in data centers, for example, that scale significantly and not least benefit our customers in the cloud.
* Compensate: SAP compensates for remaining emissions by investing in certified climate projects with high quality standards.

Cornerstone No. 1: The Green Cloud

Switching to electricity from renewable energy sources was a key milestone. Between 2010 and the end of 2013, SAP converted all of its data centers and office buildings to green energy, and at the same time increased the energy efficiency in its buildings, above all in the data centers, by leveraging innovations in server virtualization and air-flow management.

A clear indicator of just how successful that approach was is the power usage effectiveness (PUE) value of its global data center in St. Leon-Rot, which is currently 1.36. As such, the energy required to maintain the SAP facilities is now just one-third of the energy consumed by the servers. This is an ideal value for a data center operating in the highest availability class.

Cornerstone No. 2: Intelligent Mobility

Following the successful establishment of a green cloud, SAP’s climate management is now focusing on reducing the remaining sources of emissions. More than four-fifths of its current emissions are generated in the mobility sector. Among the culprits is commuter traffic, which produced 49 kilotons of CO2 in 2017. That being said, according to SAP’s most recent internal commuting survey, carbon emissions per SAP employee went down by 6.6% compared to the previous year.

SAP supports this development through a number of measures ranging from promoting bicycle use, rail transport, and e-mobility, to the operation of its TwoGo ride-sharing solution, which is open to employees and external drivers alike. Employees are also taking an increasing number of home office days — up 16% in 2017 — which is expected to have further reduction effects.

Air travel currently accounts for more than half of all mobility-related CO2 emissions. To reduce its carbon footprint in this area, SAP encourages the use of telepresence, video conferencing systems, video telephony, and other virtual collaboration technologies. Thanks to such measures, the number of flights remained constant in 2017 despite continued business growth. A reduction in the absolute number of business flights is not expected; as the world’s biggest provider of business software, SAP relies heavily on close contact with its customers and partners, and will therefore continue to depend on a certain minimum level of flights going forward.

Carbon Neutral as of 2025

Notwithstanding the above, SAP has set itself the goal of being fully carbon neutral as of 2025. Again, avoidance and reduction will be key factors in achieving this ambition. Whenever SAP is unable to reduce emissions on its own, it turns to the third pillar in its operating strategy: compensation. This fallback option means investing in CO2 offsets, such as the Livelihoods Fund’s forest restoration program. In return, SAP receives carbon credits from the sponsored offset projects. In 2017, SAP’s offset efforts resulted in total compensation of 160 kilotons of CO2.

SAP primarily uses the renowned World Wide Fund for Nature (WWF) GOLD Standard to select its offset projects. The advantage of the WWF “seal of approval” is that it not only discloses the carbon footprint of the projects, but also makes clear how the projects impact the social environment and the neighboring ecosystems.

By 2050: 85% Less CO2 Along the Entire Value Chain

SAP’s goal to be carbon-neutral by 2025 addresses the company’s internal operational processes. What customers do with SAP products is another matter. In 2017, these so-called product-in-use emissions totaled 9.7 megatons CO2. Thus, looking at the downstream emissions, climate management will be a much longer-term undertaking than minimizing SAP’s internal carbon footprint.

And SAP is tackling the problem head on. In June 2017, it committed itself to reducing all CO2 emissions of the reference year 2016 by 85% by 2050 as part of the Science Based Targets initiative. The first major interim stage is 2025, by which time a minimum 40% reduction is expected.

“Given our cloud strategy, this roadmap is definitely feasible,” says Schmid. “If all goes to plan and we really are cloud-only in 2050, the entire operation of our solutions will take place in SAP’s carbon-neutral data centers. And that in turn will mean less product-in-use emissions.”

The journey continues. http://bit.ly/2G4PRjn #SAP #SAPCloud #AI