It is something of an understatement to brand AI as a disruptive technology, given its ever-growing prevalence across business and industry. Analysts at Gartner predict that AI will “enhance decision making, reinvent business models and ecosystems and remake the customer experience” from 2018 until at least 2025. Yet a major spanner in the works is the yawning skills gap in this sector.
With enterprise professionals constantly urged to incorporate the disruptive tech into legacy infrastructure by any means necessary, upskilling personnel is never usually money down the drain.
Given true AI is decades away, the below course suggestions relate to computer science and logic-based disciplines, as industry experts told CBR that data science and cognitive programming capabilities are the most in-demand for AI projects in the workplace.
1. Machine Learning Nanodegree, Kaggle
Educators at Kaggle have put together this six-months online course to gain advanced skills in making predictive models. Hosted by web learning platform Udacity, the Nanodegree is billed “to provide students the foundation required to start participating in this exciting new field”, meaning it should be beginner-friendly yet rigorous.
Taking learners through nine discrete projects, the nanodegree develops abilities around decision-making algorithms for multiple business use cases, including house pricing, image classification and smartcab simulation using reinforcement learning techniques.
New customers start with a 7-day free trial and then pay £150 GBP / month, bringing the total cost to no more than £900 for the full six months. However, it is worth noting that the final two projects relate to CV preparation and interview/assessment techniques, rather than purely maths-based work. This may not be of use to all AI skills seekers. Udacity also offer corporate training in machine learning, data analysis, AI, self-driving cars and virtual reality (past clients include AT&T and GE).
2. Machine Learning, Stanford University
One of its most-subscribed options, this 11-week programme from e-learning platform Coursera goes through a variety of core machine learning processes and practices. Users can expect to get to grips with backpropogation algorithms for neural networks, error analysis in machine learning system design and the ins and outs of support vector machines as well as undertaking an introduction to MATLAB.
This highly popular Machine Learning course created by Andrew Ng, professor at Stanford University and formerly head of Baidu AI Group/Google Brain, is easily accessible online. That said, Week One alone contains 24 reading units, though the literature load lightens up as the programme progresses and text-based learning is broken up by video teaching.
Coursework is set up to allow flexibility, meaning if a learner misses a deadline then s/he can switch to a later session under the same programme. At a cost of just £58 (or free if the learner does not want grading or a certificate), this programme leans towards the more casual end of distance-learning courses, perhaps best suited to busy, independently-minded IT enthusiasts. Coursera warns “the Course Certificate does not represent official academic credit from the partner institution offering the course.”
3. Deep Learning Specialization, NVIDIA
Aimed at anyone who wants to “break into AI”, this Specialization by deeplearning.ai and the NVIDIA Deep Learning Institute promises to ground a participant in the foundations of Deep Learning. The online course has an industry focus, centring on building models based on application-specific topics using a Machine Learning “flight simulator”.
Those who sign up will try their hand at Convolutional networks, RNNs, LSTM, Adam, Dropout, BatchNorm and Xavier/He initialisation. The course promises that by its end, a learner will understand how to build neural networks and lead successful machine learning projects. Business use case studies covered include autonomous driving, natural language processing and industry applications in healthcare and beyond. Along the way, the course will teach Python and TensorFlow as needed. Educators advise some related experience is required – not for absolute beginners.
Automation: The GDPR silver bullet?
9 in 10 look forward to managing AI robot co-workers
5 tech trends transforming detection & prevention of financial crime
4. IBM Open Badge Programme, IBM Skills Gateway
Customers of the Open Badge Programme will learn the basics of machine learning and AI in the context of business applications. By the end, a user will have gained understanding of a range of business use cases and developed a prototype to demonstrate their new skills. Naturally, skills programmes are geared towards use of cognitive computing with IBM Watson APIs and IBM Cloud.
Three levels of awards can be earned in the form of a digital “badge” of accreditation: Explorer, Instructor and Author. Each gives credit for taking on skills in chatbots, image recognition and discovery services using IBM Watson, with the latter two rewarding those who contribute teaching programmes of their own. Courses are aimed at enterprise developers with an academic background in computer science. However, beginners’ courses are available in some data analytics disciplines.
5. Artificial Intelligence MicroMasters, Columbia University
Another offering in the range of mini degrees is the MicroMasters created by Columbia University, available to learners all over the world on the edX web portal. Recognised by industry leaders GE, IBM, Volvo, Ford, Adobe, PwC and more, the course guarantees a solid understanding of the guiding principles of AI as well as opportunities to apply concepts of machine learning to real life problems and applications.
Customers will gain expertise in the design of neural networks as well as broad applications of AI in robotics, vision and physical simulation. A Machine Learning option is also available. The programme is free to try or the full credential costs $946 (£686) plus the attainment can count towards 25% of the coursework for a full Masters degree in Computer Science at Columbia.
