Most AI heads admit deep learning still out of their grasp
Despite its perceived transformational potential, deep learning remains a machine learning ‘holy grail’ for the majority of AI specialists.
That’s the conclusion of research from Peltarion, based on the responses of 350 AI decision-makers from the UK and the Nordics with direct responsibility for “shepherding AI” at companies with more than 1000 employees.
The resulting report sought to explore the understanding of deep learning versus other types of machine learning practices, and the barriers holding businesses back from taking it from an ideal to reality.
What the findings seemed to show is that while businesses are appointing individuals to oversee AI, and there is indeed confidence in the technology’s potential, there is still a clear knowledge gap in the workings of the tech— even among the so-called specialists who make the decisions to invest and develop it.
According to the report, 99 percent of respondents thought that deep learning would transform their industry— 32 percent said it would “totally” transform it. At the same time, just 60 percent were confident about what deep learning is and how it works, and just 1 percent had deployed it extensively.
Deep learning, according to AI expert Bernard Marr, is “a subset of machine learning where artificial neural networks, algorithms inspired by the human brain, learn from large amounts of data.
“Similarly to how we learn from experience, the deep learning algorithm would perform a task repeatedly, each time tweaking it a little to improve the outcome.
“We refer to ‘deep learning’ because the neural networks have various (deep) layers that enable learning.”
Unlike other forms of machine learning, deep learning can work with unsupervised, or unlabelled data; “Just about any problem that requires “thought” to figure out is a problem deep learning can learn to solve,” Marr adds.
In terms of the wide range of applications with disruptive, or transformational, potential, deep learning is already used in voice assistants, such as Siri and Cortana; the medical sector is employing it for drug discovery and disease diagnosis; it’s also a crucial component in the development of autonomous vehicle technology, to name just a few use cases.
“It’s clear that deep learning is a truly transformative technology that has the potential to change the world,” said Luka Crnkovic-Friis, Co-Founder and CEO of Peltarion.
“But the path to reaching that potential is inhibited by lack of familiarity with deep learning.”
Complexity was cited as the most common perceived issue (70 percent) standing in the way of investment into the machine learning subset.
The specialist skills required (44 percent), lack of scalability (43 percent), a lack of understanding around deep learning models (41 percent), and lack of available data (41 percent) and the ability to collect and segment it, were also called out as key blockers.
A further issue was in integration. With approximately 191 different IT applications, systems, and services in use across organizations on average, the challenge of working the technology into this complex stack was another concern.
Despite an apparent lack of confidence, however, 80 percent of respondents had budget allocated to developing deep learning projects— up to 98 percent of respondents are planning to start investing part of their R&D budgets on deep learning initiatives over the next three years.
“With investment growing, we can expect to see more industries benefiting from this under-explored, yet incredibly powerful subset of AI. However, the barriers to adoption must be overcome before businesses can reap the benefits,” said Luka Crnkovic-Friis.
“In order to increase adoption of deep learning, companies need access to the right tools and skills,” Crnkovic-Friis concludes. “Operationalizing AI, and deep learning specifically, will be key in doing this.
“Not only should experts offer guidance, spreading the knowledge of how it can be used within their companies, but deep learning should be operationalized to increase the speed of model development and experimentation, ease integration and deployments and make deep learning more ‘AI Ready’.
“Once a few of these projects are up and running, the costs, on-site skills and infrastructure required to keep deep learning operational and launch new projects gets lower each time.”