Skip to content

Instantly share code, notes, and snippets.

@dmurfet

dmurfet/dlapplied.md

Last active Apr 20, 2020
Embed
What would you like to do?
Commercial applications of deep learning

Why is deep learning important?

It is still unclear what the long-term impacts of this technology will be. Large changes in productivity have occurred in history, and the potential of deep learning is comparable to other general purpose technologies (steam, electricity, chemical manufacturing, etc) responsible for those changes. While there are many real-world applications of today's deep learning in computer vision, natural language, and perhaps soon in robotics, these impacts would have to increase by several orders of magnitude to be reasonably compared with the general purpose technologies which drove previous industrial revolutions. However, as anybody familiar with the history of the industrial revolutions knows, once it is obvious to everybody that things are working you may not have time to catch up.

It is therefore worth noting that rich governments (US, China) and corporations (Google, Facebook, Amazon, Microsoft, Baidu, Alibaba, Tencent, ...) are making billion dollar bets that these technologies will be part of a fourth industrial revolution taking off in the next few decades. Below is a list of references that can help you understand the current landscape of these investments.

Do note that deep learning and artificial intelligence are not the same thing although they are often conflated in these kinds of reports, with varying levels of harm.

  • (Australia) Australia AI roadmap: solving problems, growing the economy and improving our quality of life. Prepared by CSIRO's Data61 for the Australian Government, this report is intended to help guide future investment in AI and machine learning (November 2019). Explains applications of artificial intelligence to natural resources and environment, health aging and disability and cities towns and infrastructure.

  • (Australia) ACOLA on the effective and ethical development of AI.

  • (France, Europe) For a Meaningful Artificial Intelligence: Towards a French and European Strategy. Written by Cedric Villani 2010 Fields medallist.

  • (China) China's New Generation Artificial Intelligence Development Plan (full translation by New America).

  • (United States) CSET: The Question of Comparative Advantage in Artificial Intelligence: Enduring Strengths and Emerging Challenges for the United States. Extract: who is leading in artificial intelligence (AI) and machine learning (ML)? How should leadership in AI be evaluated or measured? Which aspects of comparative advantage in AI possess the greatest strategic importance? These questions are critical to address as nations around the world embrace the potential of AI through a range of policy initiatives.

  • (United States) Artificial Intelligence Index Report 2019 from Stanford's Human-Centered Artificial Intelligence. Tracks a range of useful metrics for understanding the field of AI.

  • (UK) How robots change the world: What automation really means for jobs and productivity. Trends suggest the global stock of robots will multiply even faster in the next 20 years, reaching as many as 20 million by 2030, with 14 million in China alone. The implications are immense, and the emerging challenges for policy-makers are equally daunting in scale..

  • (United States) AI Nationalism by Ian Hogarth (from June 2018 but still very much worth reading).

What are the current applications?

Some notes on current and near-future applications of deep learning. References are:

The main successes of deep learning so far are in supervised learning (e.g. computer vision, speech recognition, machine translation) generative models (e.g. GANs) and reinforcement learning (e.g. AlphaGo). According to the McKinskey AI Survey, the main areas where "AI" is seeing impact are marketing and sales, product and service development, supply-chain management, manufacturing, service operations, strategy and corporate finance, risk and HR.

  • Marketing and sales: customer-service analytics, customer segmentation, channel management, prediction of likelihood to buy, pricing and promotion, closed-loop marketing, marketing-budget allocation, churn reduction, and next product to buy.

  • Product and service development: product-feature optimization, product-development-cycle optimization, creation of new AI-based enhancements, and creation of new AI-based products.

  • Supply-chain management: logistics-network optimization, sales and parts forecasting, warehouse optimization, inventory and parts optimization, spend analytics, and sales and demand forecasting.

  • Manufacturing: predictive maintenance and yield, energy, and throughput optimization.

  • Service operations: service-operations optimization, contact-center automation, and predictive service and intervention.

  • Strategy and corporate finance: capital allocation, treasury management, and M&A support.

  • Risk: risk modeling/analytics, and fraud/debt analytics.

  • HR: performance management and organization-design, workforce-deployment, and talent-management optimization.

Examples

  • Machine translation
  • Speech recognition
  • Speech synthesis (podcast)
  • Image production using GANs (Rosebud.AI)
  • Game content production using GANs
  • Music production (underscore)
  • NLP for document analysis
  • NLP for summarisation
  • NLP for smart search
  • NLP for basic reasoning
  • NLP for conversational agents
  • Computer vision for infrastructure maintenance
  • Drug discover
  • Chemistry
  • Robotics (Covariant.AI)
  • To perfume
  • To glass
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
You can’t perform that action at this time.