The most popular and comprehensive Open Source ECM platform
Let us know more
Data Science and Machine Learning: Trends Driving Wider Acceptance of ML Decisions
While Machine Learning is making a big impact across different industries, the fact is that less than ten percent of all companies say that they’ve made any investment in it at all. ML technology is evolving quickly,and as it does, the impact and application of ML techniques are expected to increase rapidly.
Lisa Morgan, freelance writer at InformationWeek, discussed in an interview with David Schatsky, managing director at Deloitte, five key trends that are expected to make Machine Learning an important force in the future:
- Data Capture and Cleansing will be automated. Other than having the skills to know how to analyze the data, perhaps the biggest headache for data scientists is prepping their data for analysis. Capture and cleansing takes a lot of time but it is a process that is expected to gradually become automated, and once automated, data scientists will be able to focus their attention to analysis.
- Improved ML Training techniques. ML typically requires massive amounts of data to train with in order to identify patterns and important parameters that define the data. New techniques are being developed that either require less data for training or which can use ‘synthetic data‘.
- Faster Hardware. Chip vendors like Nvidia are rapidly introducing new chip generations of GPU and ASIC-based hardware that are targeting specific ML techniques and applications. Better hardware will make computations go more quickly, allowing ML algorithms to train and run more speedily.
- Transparency. Many ML models currently produce results which can’t be traced back to provide concrete reasons for why or how those results were arrived at. Today’s black-box algorithms for ML are expected to be replaced by algorithms that allow better understanding of how the ML results were arrived at, and this is expected to increase the trust that people will be able to place in ML-derived decisions.
- Local Computations. Hardware will not only be faster, but hardware is also expected to shrink in size and cost. Computations today that might require large-scale computations in the cloud or across multiple servers will be able to shrink to a size that can easily be processed locally, making ML more secure and more convenient.
Dick, great post. Following up on that, is there data science/ML software that you recommend? I have a capture solution (Ephesoft), but need to use that information more fully. It seems like there are a ton of ML options out there, and I don’t know enough to know which platforms are a head above the rest. One of several that I am looking at is https://www.bisok.com/data-science-workbench/ . Thank you, your help is appreciated