Within the past few years, the study of data science and machine learning has exploded into its own job field. However, the tech subgenre has been galavanting to the mainstream for nearly 3 centuries. It all started sometime in the 1740s with Bayes’ Theorem.
Today, the demand for data scientists is at its peak and is only continuing to surge. By the end of this year, there will have been 2.7 million data scientist job openings. The million-dollar question is:
What do data scientists do?
To be direct, data science is the process of analyzing data. Explaining with more complexity, data scientists use heterogeneous data – data that were composed of different forms or dissimilar components – to solve rigorous problems.
To do so, data scientists use their master skills in computer science, high-level mathematics, and more. These aforementioned skills are especially unique to data scientists developing industrial machine learning technologies, programs, and Enterprise AI.
In short, data scientists take three steps in analyzing their data: preparing, testing, and capturing. Of course, 2-3 individual tasks are necessary to complete each step.
To prepare for data analysis, data must be first captured. In other words, the first task in a data scientist’s work is to simply collect their data. Scientists can extract or acquire them. To further prepare for data analysis, the scientist then maintains their data, safely storing and staging it.
Some data storage methods include scalable storage optimized with artificial intelligence. Finally, the data undergo processing which involves mining and classification.
Once data scientists are done with preparation, having completed the previously mentioned tasks, they will move on to physically testing their findings. Think back to your public school science fair days:
What must you do before testing your data?
Draft a hypothesis or an educated guess. It involves developing a theory to test with the data model. After this, the data is finally ready for analysis. This is the stage where new findings based on the initially collected data are discovered. It’s often done by modeling, exploring, and experimenting with data to reach desired outcomes and to decide what the data means.
All that’s left to do at this point is to communicate the results. Reconnecting data science to science fair procedures, how would you express your final discoveries? Perhaps a visual aid, such as a tri-fold poster board.
While data scientists may not necessarily take this route of expression, they create an easily understood picture of the model’s predictions. It’s easily translatable to an audience of laymen. The final task in data analysis is to apply your results. It’s to help end-users understand how to use the predictions to take effective actions within their business.
Who are data scientists?
More importantly, how do you become one?
Jenn Gamble, Director of Noodle.ai – the leading machine learning software giant – spoke on the subject, saying, “You don’t necessarily need a Ph.D. to do data science – you need an aptitude for math and a creative, problem-solving mentality.”
By 2025, we will be creating data worth 175 billion terabytes on a daily basis, so the primary way to fully understand and analyze the world’s surging data is to hire more data scientists with access to advanced tools.
Some of the most popular tools in the industry include the R programming language, python, PyTorch, hadoop, and Apache Spark. Among the most crucial roles needing fulfillment in the machine learning job economy are data engineers, AI hardware specialists, and software engineers.
Data engineers create and maintain the methods which bring in data, needing skills in Scikit-learn, AForge.NET, and/or Java programming language. Software engineers analyze business data and design software to fit needs, needing skills in Java programming language, SQL, and/or python. Lastly, AI hardware specialists create and program AI to perform specific tasks, needing skills in machine learning, python, Saas, and Java programming language.
Data science provides opportunities for people to express their creativity.
It gives them the means to create technology that can initiate changes worldwide. Think about all of the space exploration, autonomous vehicles, personalized medicine, and personalized education that have been created within the past few years. They are the work of data scientists.
This isn’t all, however.
Data scientists have also created technologies capable of monitoring wildlife migration and optimize energy. The essentiality of data science is no question. In fact, between just 2011 and 2012, job listings for “data scientist” increased by 15,000%.
Find out more about what data science is below.
The UK has a strong history when it comes to chips and processors, but the global chip market has seen some ups and downs of late. Today comes some big news that underscores how investors are doubling down on one of the big hopefuls for the next generation of chipmaking to see it through the winter. Graphcore, the Bristol-based startup that designs processors specifically for artificial intelligence applications, announced that it has raised another $150 million in funding for R&D and to continue bringing on new customers. It’s valuation is now $1.95 billion.
Graphcore has now raised over $450 million and says that it has some $300 million in cash reserves — an important detail considering the doldrums that have plagued the chipmaking market in the last few months, and could become exacerbated now with the slowdown in production due to the coronavirus outbreak.
The funding is an extension of its Series D, it said, and brings the total valuation of the company to $1.95 billion. (For reference, the original Series D in December 2018 valued Graphcore at $1.7 billion.) This latest round includes investments from Baillie Gifford, Mayfair Equity Partners and M&G Investments — all new backers — as well as participation from previous investors Merian Chrysalis, Ahren Innovation Capital, Amadeus Capital Partners and Sofina. Other backers of the startup include BMW and Microsoft, Atomico and Demis Hassabis of DeepMind.
Graphcore’s big claim to fame has been the development of what it calls its Intelligence Processing Unit (IPU) hardware and corresponding Poplar software — which is designed specifically for the kind of simultaneous, intensive calculations demanded of AI applications innovators create next generation machine intelligence solutions. Graphcore describes the IPU as the first processor to be designed specifically for AI, although a number of other companies including Nvidia, Intel and AMD have made huge investments into this area and have ramped up their pace of development to meet market demands and hopefully overtake what have been limitations in the wider area of AI processing.
This D2 round comes ahead of what it describes as strong demand for 2020, and is happening on the heels of a strong year for Graphcore, the company said, including a commercial deal with one of its previous strategic backers.
“2019 was a transformative year for Graphcore as we moved from development to a full commercial business with volume production products shipping,” said Nigel Toon, founder and CEO. “We were pleased to publicly announce our close partnership with Microsoft in November 2019, jointly announcing IPU availability for external customers on the Azure Cloud, as well as for use by Microsoft internal AI initiatives. In addition, we announced availability of the DSS8440 IPU Server in partnership with Dell Technologies and the launch of the Cirrascale IPU-Bare Metal Cloud. We also announced some of our other early access customers which include Citadel Securities, Carmot Capital, and Qwant, the European search engine company.”
See Toon speaking at our recent Disrupt conference in Berlin about the prospect for chips here:
Monday’s episode of the Tonight Show saw K-pop group BTS join Jimmy Fallon for a round of Subway Olympics, playing several games that probably won’t appear in the real Olympics any time soon. These included Protect the Duck, in which players attempted to knock a rubber duck off each other’s hand, and Dance Your Shades and Gloves Off, which was exactly what it sounds like.
It was predictably chaotic, with J-Hope accidentally slapping Jin in the face at one point, but it wouldn’t be BTS playing games without some silly antics. There was even a bit of cheeky cheating, with Jimin sneakily sticking Post-Its on Jungkook well after time was up. Read more…