Weekend leisure? Why don't you take a look at a selection of articles covering the IT industry's latest hot topics chosen for you by our team.
Does this need any explanation? Yes, this is an IT orientated week summary, but events like this spark interest in STEM which effect we will see in the next generation. In short we have made a huge step towards being able to land on another planet - flew and landed a reusable rocket. Why is it so important? I'm not the right person to explain this, but "I know a guy". Book yourself 52 minutes and listen to how it felt to be a ten year old boy living in the USA witnessing a man landing on the Moon.
It has been already established that machine learning will be the main topic of this year. The second will be security. How big is data security becoming? IBM suggests 6 times more than the previous year when counting number of leaked records.
Last week I have written about a Bitcoin debate of hard-fork possibility. The 21 team conducted a survey among 61 blockchain world influencers. It is not to guess that opinions were… divided. The problem of governing a distributed platform is hard to solve. Bitcoin network has reached certain level of maturity and besides technological factors, also economic factors come into the picture when deciding on the future roadmap. Even in case of Ethereum which has once-a-year-hard-fork policy it is not seamless. It is worth reading what wise heads say about the topic.
Machine Learning is one of the most important technology trends in recent years. At ITMAGINATION we put much attention to this area as well. But what really is ML? What does it mean that algorithms need to learn? Why do they need to do so? Avinash Sharma V gives a little bit of the answers. At the same time he points out the importance of data quality for Machine Learning algorithms.
Can a piece of software help to improve projects’ profitability by optimization of staffing? With machine learning in place it’s possible. Read the blog post written by Cortana Intelligence and ML Team from Microsoft to find out how Baker Tilly Virchow Krause, LLP, a full-service accounting and advisory company, is using Microsoft Data Platform to achieve 4-5% improvement on projects’ profits.
Most AI systems today rely on supervised learning - one provide labelled data, and get a program which then can perform analogous computation for new data. The problem is that if we provide bad data to the input, the supervised learning will obviously fail (the program will not compute anything meaningful). In this blog post the Explosion AI team describes why and how to take care about appropriate data labelling.