Consumers have spoken, artificial intelligence is a profitable industry. From Amazon to Google to Apple, major tech companies have made inroads, crafting intelligent software — housed in sleek, accessible hardware — that has drawn massive customer attention.

This trend is set to soon move out of home devices, like Echo and Google Home, and onto the streets, where self-driving cars leverage major breakthroughs in computer vision so passengers can ride easy, knowing their vehicles will “see” and react to objects and road signs in real time without their input. In fact, cars with these features are already popular with consumers, and by 2020 10 million cars with self-driving attributes will be on roadways.

But while there are plenty of ways for consumers to leverage AI, enterprises are asking themselves how they can get in on this wave of innovation. And a big part of the answer lies at the crossroads of computer vision and an emerging field known as transfer learning.

Transfer learning has the potential to unlock dozens of new AI use cases in the enterprise by reusing existing, state-of-the-art deep learning models.

In short, transfer learning is an approach to taking existing AI models and applying them to new data. In this case, we’ll talk about computer vision models that make deductions based on images and visual data, applying that learning on numerical sets of data. It means that businesses could use the very advanced deep learning models that perform computer vision functions, like for self-driving cars, and apply that level of sophistication to a whole new set of, non-image datasets on a spreadsheet.

The way transfer learning works is that the algorithm functions much like the human brain does when looking at a small dataset in Excel. We use our eyes to scan through information and mentally detect patterns. Computer vision applied to numerical data does the same thing, through convolutional neural networks, which look for both high- and low-detail features of an image to help classify what is pictured.

Turning numbers into images

Through transfer learning, a model built originally for computer vision use cases can take a layered approach to curate the new numerical data it’s fed, improving its guesses on patterns in data from a separate domain. Researchers have proven this out, showing that transfer learning that is built using vision-based source data can be applied to a target domain of data that doesn’t contain visual information, like sensor data from Internet of things devices, for example.

For instance, a 2-D sensor’s numerical readings can be converted into pressure distribution image heat maps. Then the convolutional neural network’s last layer is peeled back, and transfer learning aids the computer vision model to reinterpret the data as a visual.

In terms of use cases, often, for many deep learning applications to work, computer vision included, it needs a much larger dataset to gain new insights from. That is something many companies have as the world ramps up to having more than 4.4 zettabytes of data. However, using transfer learning, enterprises can extrapolate patterns and real business insights from smaller data sets as well.

Companies can apply existing and sophisticated computer vision models to more traditional use cases like fraud detection and prevention, preventative maintenance, and marketing attribution. These are all areas where enterprises can apply image-based deep learning models to more general, or numerical, data sets. Much of this results in net-new insights that are not achievable with traditional machine learning methods or advanced analytical approaches.

It is important for business leaders to know that deep learning, and specifically computer vision, has much wider applications in the enterprise beyond simply vision-based use cases like autonomous driving and identifying facial micro-expressions. In fact, applying computer vision to enterprise problems can unlock dozens of new use cases and business outcomes. It’s by no means easy, but deep learning does provide a path ahead. Not every company can afford to spin up 1,000 new hires in data science, like Google or Amazon — and they don’t have to.

Not only are the state-of-the-art models mentioned available, but they can also be embedded as pre-built functions through popular libraries like TensorFlow and Keras. Therefore, methods like computer vision combined with transfer learning can help level the playing field for any businesses today that want to use AI to get more out of their data.

The post The Secret to AI Could be Little-Known Transfer Learning appeared first on Artificial Intelligence.

While artificial intelligence (AI), machine learning and deep learning are often thought of as being interchangeable, they do in fact relate to very different concepts. It all began in the 1950s with AI and the idea that a computer could be made to simulate human learning and intelligence.

A subclass of that is machine learning, whereby a computer can take large amounts of data and use it begin to recognize patterns, make predictions on new data, and essentially ‘learn’ for itself. The drawback is that machine learning requires that parameters be set for what the computer needs to recognize, and those inputs can be time-consuming. And so we go one step further, into deep learning.

For example, Ripjar offers a service under the heading of ‘Analysis at the Speed of Thought’ that utilizes deep learning combined with natural language processing to analyze an organization’s internal data, in addition to information from sources like news feeds, web pages, and social media posts. These data streams are captured and monitored in real-time, in more than 160 languages, in order to provide cybersecurity, reputation management, compliance, etc. Without the capabilities of deep learning, the inputs required to get results would prove incredibly difficult. In essence, deep learning is enabling the practical application on of machine learning. So how does it work?

Inspired by the structure and activity of neurons within the human brain, deep neural networks (DNN) form the basis of deep learning. Through these algorithms, computers are able to identify features in significantly sized datasets and progress that information on through layers of the neural network, refining as it goes. This leads to a hierarchical representation of the problem.

Developer Considerations for Machine Learning 

There are many reasons why startups might struggle to fulfill their potential for financial and technological success. Among the many unique challenges they face from initial concept through to expansion, a lack of scalability can be one of the most difficult to overcome. In this section, we’ll focus on the capabilities and practical application of machine and deep learning, the frameworks and technologies you need to know about, and the ways that the community can help from the very beginning.

If you’re trying to decide whether or not to begin a machine or deep learning project, there are several points that should first be considered:

• Cost
• Need
• Organizational readiness
• Industry readiness
• Competition
• Regulations and compliance
• The pace of innovation

It may sound obvious, but the majority of startups that fail to find traction in the market do so because they’ve identified a need that doesn’t really exist – or at least not enough to be monetized.

Cost can often be the deciding factor. Can your organization afford to embark on this journey, and will your potential customers be able to afford what you’re offering? Be realistic when making these assessments. Once that’s out of the way, the second issue is one of need. It may sound obvious, but the majority of startups that fail to find traction in the market do so because they’ve identified a need that doesn’t really exist—or at least not enough to be monetized.

Readiness is a question you must ask of yourself and the industry. Is your organization ready (and able) to devote time and resources to integrating machine and deep learning into the pipeline, and is the industry ready to adopt your new solution or service? Another thing to consider is the competition. It’s an exciting time for startups, and the potential is huge, but tech heavyweights like Google and Microsoft are also looking to cash in on deep learning. It’s worth keeping that in mind when positioning yourself in the market with a specialty.

For the past five years or so, the pace of innovation within machine and deep learning has quickened significantly. Will your organization be able to keep up?

If they occur, regulation and compliance issues can slow everything down so much that it no longer becomes worth the effort. Finally, is it scalable? For the past five years or so, the pace of innovation within machine and deep learning has quickened significantly. Will your organization be able to keep up?

Where to Begin 

If you’re approaching machine or deep learning with no real experience in the design, development and employment of deep neural networks, you’re in good company. Very few organizations—and even fewer startups—come staffed with a full roster of data scientists, ready to build a platform on an enterprise scale.

One of the first points it’s important to recognize is just how accessible machine and deep learning truly are—though that shouldn’t be confused with thinking that these are easy fields to be in. Having the computing power and necessary people skills at your disposal won’t guarantee results. After giving careful consideration to the issues highlighted in the overview, the first step is to focus on the tools and infrastructure while remembering that machine and deep learning successes comes from more than the algorithms.

How to Choose a Framework

Frameworks, applications, libraries and toolkits—journeying through the world of deep learning can be daunting. The ease with which you’ll be able to build and run your application is first determined by the framework you choose. With that in mind, the five best-known frameworks are as follows:

• • 1. Caffe
    2. Tensorflow
    3. Torch
    4. Apache Mahout
    5. Microsoft Cognitive Toolkit (CNTK)

These are five of the frameworks, but you may still be wondering how to choose between them. The answer is that it really depends on what your goals are. If in doubt, it can be helpful to go with one of the more popular or supported frameworks like Caffe or Torch. The full guide covers descriptions and specifics on each of these frameworks to assist you in choosing the perfect framework for your needs.

Deploying the right kit can be critical, and the main thing is the significant advantages that GPU acceleration provides. GPUs and deep learning go together like a marriage made in heaven. The multi-layered nature of the deep neural networks means that they run best on highly parallel processors. Deep learning training and inference will, therefore, be achieved much faster on GPUs—any GPUs—from small workstations to some serious hardware. In fact, you can start developing on any GPU-based system.

The insideHPC Special Report, “Riding the Wave of Machine Learning & Deep Learning,”explains it well: ‘the high compute capability and high memory bandwidth make GPUs an ideal candidate to accelerate deep learning applications, especially when powered with NVIDIA’s Deep Learning so ware development kit (SDK) that includes CUDA® Deep Neural Network library (cuDNN), a GPU-accelerated library of primitives for deep neural networks, TensorRTTM, a high performance neural network inference engine for production deployment of deep learning applications, and CuBLAS a fast GPU-accelerated implementa on of the standard basic linear algebra subroutines.’

The NVIDIA cuBLAS library is a fast GPU-accelerated implementa on of the standard basic linear algebra subrou nes (BLAS). Using cuBLAS APIs, you can speed up your applications by deploying compute-intensive operations to a single GPU or scale up and distribute work across multi-GPU configurations efficiently.

The full guide also offers information on how developers can receive help from community resources, as well as what questions you should be asking while exploring the field of machine learning.

The post What Developers Need to Consider When Exploring Machine Learning appeared first on Artificial Intelligence.