The Future of Deep Learning

By on

Deep learning (DL) became an overnight “star” when a robot player beat a human player in the famed game of AlphaGo. Deep learning training and learning methods have been widely acknowledged for “humanizing” machines. Many of the advanced automation capabilities now found in enterprise AI platforms are due to the rapid growth of machine learning (ML) and deep learning technologies. What’s Next for Deep Learning? attempts to answer this question that originally appeared on Quora.

A Deep Dive into Deep Learning in 2019 comments on the “ubiquitous” presence of DL in many facets of AI — be it NLP or computer vision applications. Gradually, AI and DL-enabled automated systems, tools, and solutions are penetrating and taking over all business sectors —from marketing to customer experience, from virtual reality to natural language processing (NLP) — the digital impact is everywhere.

Facebook Researchers Plagued with Privacy Dilemma

Deep Learning Will Be the End to End Encryption brings forth controversy over public demand of absolute privacy of personal data. This consumer demand is in direct conflict with Facebook’s current AI research endeavors. The AI researchers at Facebook need to “mass harvest” personal data to train learning algorithms.

Facebook realizes that the utopian concept of end-to-end encryption was indeed a myth in a research world seeking answers from piles of personal data. For future efforts, researchers are now seriously considering training algorithms on “dead data” on individual devices rather than mass harvesting personal data. In that case, Facebook engineers will install content-moderation algorithms directly on users’ phones to bypass data-privacy violations.

In a controversial post, the author of this KD Nugget post predicts that deep learning may not be the future of AI. The reason behind this, according to the author, is that in future many DL methods will not only become non-complaint, but outright illegal. The post also suggests there is a distinct possibility that future mobile apps will be devoid of DL.

Another severely limiting characteristic of DL-enabled solutions is that the learning algorithms still cannot provide detailed reasons for their choices, which can provoke users to accept decisions provided by AI tools blindly and then concoct “fake” explanations for any rejected answer. That is not very encouraging for decision-support solutions!

Democratization of Deep Learning in Five to Ten Years

Predictions for the Future of Deep Learning claims that in the next 5 to 10 years, DL will be democratized via every software-development platform. DL tools will become a standard part of the developer’s toolkit. Reusable DL components, incorporated into standard DL libraries, will carry the training characteristics of its previous models to speed up learning. As automation of deep learning tools continue, there’s an inherent risk the technology will develop into something so complex that the average developer will find themselves totally ignorant.

Predictions About Deep Learning

Towards DataScience has this to say about the imminent future of deep learning:

  • Prediction 1: Deep learning networks will demystify computer memory.
  • Prediction 2: Neural architecture search will play a key role in building datasets for DL models.
  • Prediction 3: NAS will continue to use reinforcement learning to search convolutional architectures.

An Information Age  article argues in favor of unsupervised learning methods over training data. The hope is that with time, unsupervised learning will be able to match the “accuracy and effectiveness” of supervised learning. In spite of high volumes of available data, most of it is still unusable to DL algorithms.

Deep Learning Applications of the Present and Future

Google was the pioneer in pursuing deep learning in marketing. Google’s acquisition of DeepMind Technologies shook the business world. Google’s mission is to make DL a serious solution for search marketers who care about SEO. 

The Future of Artificial Intelligence for Small Businesses showcases the bi-directional movement of AI between research labs and company operations, where businesses are using the power of automated AI tools to enhance customer experience or execute high-speed data analysis.

Machine Learning and Artificial Intelligence Trends in 2019 presents some interesting AI and ML trends for the current year. The most notable trend to follow is the real-world impact of ML technologies and tools as they begin to transform one business at a time “from chatbots and digital agents in CRM to virtual reality (VR)-powered shop-floor demos.” The future Ml technologies, which include DL, must demonstrate learning from limited training materials, and transfer learning between contexts, continuous learning, and adaptive capabilities to remain useful.

A YouTube video, The Future of Deep Learning Research, talks about back propagation, its use in deep learning research, and seven research methods that can potentially overtake back propagation in near future.

Deep Learning Future Trends in a Nutshell

Some of the primary trends that are moving deep learning into the future are:

  • Current growth of DL research and industry applications demonstrate its “ubiquitous” presence in every facet of AI — be it NLP or computer vision applications.
  • With time and research opportunities, unsupervised learning methods may deliver models that will closely mimic human behavior.
  • The apparent conflict between consumer data protection laws and research needs of high volumes of consumer data will continue.
  • Deep learning technology’s limitations in being able to “reason” is a hindrance to automated, decision-support tools.
  • Google’s acquisition of DeepMind Technologies holds promise for global marketers.
  • The future ML and DL technologies must demonstrate learning from limited training materials, and transfer learning between contexts, continuous learning, and adaptive capabilities to remain useful.
  • Though globally popular, deep learning may not be the only savior of AI solutions.
  • If deep learning technology research progresses in the current pace, developers may soon find themselves outpaced and will be forced to take intensive training.

Image used under license from Shutterstock.com

We use technologies such as cookies to understand how you use our site and to provide a better user experience. This includes personalizing content, using analytics and improving site operations. We may share your information about your use of our site with third parties in accordance with our Privacy Policy. You can change your cookie settings as described here at any time, but parts of our site may not function correctly without them. By continuing to use our site, you agree that we can save cookies on your device, unless you have disabled cookies.
I Accept