Deep Learning (DL) has become more than just a buzzword in the Artificial Intelligence (AI) community – it is reshaping global business through the prolific use of autonomous, self-teaching systems, which can build models by directly studying images, text, audio, or video data.
Such systems can use that data for future pattern recognition.
According to many technical professionals, businesses can reap the full benefits of AI only when the appropriate levels of competency is developed in advanced data technologies such as Machine Learning (ML) and Deep Learning for extracting reliable business insights. This unilateral opinion among professionals implies that the skill gaps have to be identified and training must be in place to make the best use of available technologies and tools.
The Gartner article Artificial Intelligence and the Enterprise indicates an urgent need to develop teams of highly skilled Data Science expert “who can manage the complexity of data, analytical methods and machine learning associated with AI, and help apply it with workers, customers and constituents.”
AI Analytics Solutions: How Much is Really Deep?
In How to Tell When Vendors Are Hyping AI Capabilities, the author discusses how an increasing number of technology vendors in the market continue to claim that they have value-added AI capabilities in their solutions. But, the real value of AI in the enterprise is getting side-tracked by market hype.
The purchase decision for business consumers gets tougher when they have to compare and choose the most suitable AI solution (whether based on ML or DL) for their specific Analytics or BI needs. Customers need to be more focused and ask questions about managing risks, monitoring performance, and extracting the right business benefits from a given solution.
The Deep Learning revolution began from a need to build “high-accuracy predictive models” from unstructured data such as images, voice, and natural language. The ultimate push came from the four giants in the IT industry (Facebook, Google, Microsoft, and IBM) who were all out to win the AI in the enterprise race by leveraging their Deep Learning technology development strategy.
Now the revolution has reached the proportion of a gigantic disruptive wave, which has also created further development opportunities for solution vendors. The Forrester Report Deep Learning: an AI Revolution Started for Courageous Enterprises shows how the Deep Learning storm in the market inspired enterprises to embrace advanced AI research for their Analytics and BI ever-growing requirements.
According to recent Tractica report on Deep Learning, the DL software market will expand from “$655 million in 2016 to $34.9 billion worldwide by 2025.” Moreover, this report suggested that the top 10 Deep Learning use cases in terms of potential for revenue generation are:
“(1) Static image recognition, classification, and tagging; (2) Machine/vehicular object detection/identification/avoidance; (3) Patient data processing; (4) Algorithmic trading strategy performance improvement; (5) Converting paperwork into digital data; (6) Medical image analysis; (7) Localization and mapping; (8) Sentiment analysis; (9) Social media publishing and management; (10) Intelligent recruitment and HR systems.”
Factors that Differentiate Deep Learning from Other AI Technologies
One of the primary drivers of Deep Learning is that it can crunch much more data at very high speeds. DL techniques have become necessary for successful pattern recognition in large unstructured data. So, two major factors that differentiate Deep Learning from other AI technologies: Largeness of training data and direct analysis of unstructured data.
In traditional Data Modeling, a “labeled data set” is used to train a model with an algorithm, and then the model is expected to accurately predict new data sets in future, based on that learning. Deep Learning takes this learning process one step ahead by directly working with images, audio, or video data without the data going through any kind of initial preparation.
The Data Scientist just tells the DL algorithm what to look for and then, the algorithm does it all. This unique capability of DL is known as “feature engineering,” which helps the DL algorithm to directly focus on the right features or distinguishing elements of particular data without any intervention from a technical expert.
Thus, in case of DL, Data Science staff members are not required for training data models to recognize pattern in images, audio, or video data. However, the major drawback in DL is going through an unlimited number of permutations to make the feature engineering process accurate.
Another drawback is that for Deep Learning processes to work, very high-end supercomputers handling billions of high-speed mathematical computations are required. One answer to this processing need has been provided by NVIDIA’s GPU system and open source Deep Learning libraries. NDVIDIA has helped to make DL research and development affordable for enterprises.
Applications in Businesses: Deep Learning Use Cases
Deep Learning algorithms are becoming more widely used in every industry sector from online retail to photography; some use cases are more popular and have attracted extra attention of global media than others. Some widely publicized Deep Learning applications include:
- Speech recognition used by Amazon Alexa, Google, Apple Siri, or Microsoft Cortana.
- Image recognition used for analyzing documents and pictures residing on large databases.
- Natural Language Processing (NLP) used for negative sampling, sentiment analysis, machine translation, or contextual entity linking.
- Automated drug discovery and toxicology used for drug design and development work, as well as for predictive diagnosis of diseases.
- CRM activities used for automated marketing practices.
- Recommendation engines used in a variety of applications.
- Predictions in gene ontology and gene-function relationships.
- Health predictions based on data collected from wearables and EMRs.
The Computer World article Deep Learning Use Cases for ASEAN describes how DL algorithms can be used to aid traffic management in ASEAN member countries.
Deep Learning Success in the Enterprise
Deep Learning use cases have been widely used for knowledge discovery and Predictive Analytics. For example, Google uses DL to build powerful voice- and image-recognition algorithms. Netflix and Amazon use DL in their recommendation engines, and MIT researchers use DL for Predictive Analytics.
According to the NVIDIA article Deep Learning Success Stories, Deep Learning case studies are easily found in many enterprises. Soon, DL solutions will be used in ways no one thought was possible. Right now, we can witness successful DL applications behind self-driving cars, automated web services like recommendation engines, and smart assistants.
Here are some business-specific, Deep Learning use cases:
- Canary: a NY-based DL startup has their vision set on the world’s first smart home security device, which comprises of an HD video camera and sensors for tracking temperature, sound, vibration, air quality, and movement. This device can be controlled by a smartphone. A product video is available here.
- Atomwise: Another startup applies Deep Learning technology to drug discovery. This solution uses DL networks to help discover new medicines and to explore avenues for repurposing known and tested drugs for use against new diseases.
- ViSenze: Provides tools to simplify image search and categorization through an API. The solution uses DL networks to power image recognition and categorization.
- Bay Labs: A startup devoted to medical imaging technology has used DL for medical diagnosis and disease management. They have their eyes on both the developed and developing nations with a firm vision to improve the quality of global healthcare.
- Knit Health: Promises to help people with sleeping disorders. They have combined computer vision and DL technologies to provide “personalized insights and suggestions” related to sleep management.
- CarePredict: Provides a senior care platform to enable timely intervention and monitoring of medical conditions that may have been overlooked by close family members or friends. The ultimate goal of this solution is to provide timely detection of preventable health conditions. A video of CarePredict is available here.
- BenchSci: Is an Machine Learning research tool, which aids biomedical researchers to locate the best biological compounds for their experiments. This solution grew out of a highly publicized need for finding a tool to quickly glean millions of research publications for locating the most suitable antibodies for particular experiments. A video is available here.
You will find more examples of Deep Learning use cases in Deep Learning Startups, Use Cases, and Books,
Deep Learning has pervaded the global business landscape, capturing the undivided attention of industry giants like IBM, Facebook, Google, Microsoft, Twitter, PayPal, or Yahoo, among others. It’s quite clear that large or small companies alike are making heavy investments into Deep Learning technologies, as they all think such advances will be core drivers of enterprise growth far into the future.
Photo Credit: hanss/Shutterstock.com