Monday, February 5, 2018

AI SPECIAL ....Artificial intelligence: The time to act is now PART 1

Artificial intelligence: The time to act is now PART 1

·          
Artificial intelligence will soon change how we conduct our daily lives. Are companies prepared to capture value from the oncoming wave of innovation?
The evolution of AI
Artificial intelligence (AI) was born in the 1950s, when the English polymath Alan Turing created a test to determine a machine’s ability to mimic human cognitive functions, including perception, reasoning, learning, and problem solving. AI grew with the rise of machine learning (ML)—wherein systems absorb and “learn” from data. They then use this knowledge base to make better predictions and decisions over time. In 2010, the advent of deep neural networks ushered in the deep learning (DL) era.
All ML and DL solutions require two steps: training and inference. Take the software in autonomous cars. To help systems detect obstacles in the road, developers present images to the neural net—for instance, those of dogs or pedestrians—and perform recognition tests. Network parameters are then refined until the neural net displays high accuracy in visual detection. After the network has viewed millions of images and is fully trained, it enables recognition of dogs and pedestrians during the inference phase.
Training now accounts for about 95 percent of AI-related workloads in the public cloud because most AI applications are still relatively immature and require huge amounts of data to refine them. As AI models mature, inference will gain more share in the cloud. In fact, DL inference could account for 30 to 40 percent of public-cloud workloads over the next three to five years, with training dropping to 60 to 70 percent. Inference will also gain share with the rise of edge computing (which takes place within devices), as innovation enables low-power, high-performance inference chips.
Pity the radiology department at your local hospital. Yes, they have a fine MRI machine and powerful software to generate the images. But that’s where the machines bog down. The radiologist has to find and read the patient’s file, examine the images, and make a determination. What if artificial intelligence (AI) could jump-start that process by enabling real-time and more accurate diagnoses or guidance, beyond what human eyes can see?
Thanks to technological advances over the past few years, manufacturers are close to offering such leading-edge MRI solutions. In fact, they’re exploring new AI applications that span virtually every major industry, from industrials to the public sector. With better algorithms and increased stores of data, the error rate for computer calculations is now often similar to or better than those of human beings for image recognition and several other cognitive functions. Hardware performance has also improved drastically, allowing machines to process this unprecedented amount of data. That has been a major driver of the improvement in the accuracy of AI models.
Within AI, deep learning (DL) represents the area of greatest untapped potential. This technology relies on complex neural networks that process information using various architectures, comprised of layers and nodes, that approximate the functions of neurons in a brain. Each set of nodes in the network performs a different pattern analysis, allowing DL to deliver far more sophisticated insights than earlier AI tools. With this increased sophistication comes greater needs for leading-edge hardware and software.
Well aware of AI’s massive potential, leading high-tech companies have taken early steps to win in this market. But the industry is still nascent and a clear recipe for success hasn’t emerged. So how can companies capture value and see a return on their huge AI investments?
Our research, as well as interactions with end customers of AI, suggests that six tenets will ring true once the dust settles. First off, value capture will initially be limited in the consumer space, and companies will achieve most value by focusing on enterprise “microverticals”—specific use cases within select industries. Our analysis of the technology stack also suggests that opportunities will vary by layer and that the most successful companies will pursue end-to-end solutions, often through partnerships or acquisitions. For certain hardware players, AI might represent a reversal of fortune, after years of waning interest from investors who gravitated toward software, combined with heavy commoditization that depressed margins. We believe that the advent of AI opens significant opportunities, with solutions in both the cloud and the edge generating strong end-customer demand. But our most important takeaway is that companies need to act quickly. Those that make big bets now and overhaul their traditional strategies will emerge as the winners.
The nuts and bolts of the AI market
Despite the hype about AI, the market can intimidate even the most fearless analysts and investors. No standard definition of the technology stack has emerged in the industry, making it difficult to understand the crowded competitive field. Of the hundreds of companies jockeying for market share, who is offering what?
To bring some clarity to the seemingly chaotic supply landscape, we divided the machine-learning (ML) and DL technology stack into nine layers, across services, training, platform, interface, and hardware. Some companies are competing in multiple layers, while others are concentrating on only one or two. As we’ll discuss later, companies that limit their focus to specific layers may find themselves at a disadvantage.
Edge and cloud solutions
Traditionally, most AI applications have resided in the cloud—a network of remote servers—for both training and inference. However, inference at the edge will become increasingly common for applications where latency in the order of microseconds is mission critical. With self-driving cars, for instance, the decision of braking or accelerating must occur with near-zero latency, making inference on the edge the optimal option. Edge computing will also emerge as the favored choice for applications where privacy issues and data bandwidth are paramount, such as AI-enabled CT-scan diagnostics. The growth of edge computing will create new opportunities for all players along the technology stack, particularly for hardware developers.
Our core beliefs about the future of AI
AI is positioned to disrupt our world. McKinsey Global Institute estimates that rapid advances in automation and artificial intelligence will have a significant impact on the way we work and our productivity. To capture value in this growing market, companies are experimenting with different strategies, technologies, and opportunities, all of which require large investments. While much uncertainty still persists, companies that heed the following points will be better positioned to win.
1. Value capture will initially be limited in the consumer sector
The first consumer AI offerings share a common trait: they enhance products but don’t directly contribute to the bottom line. Most of these come from large and well-known tech players, including some online translation and photo-tagging services, or digital voice assistants on mobile phones. Such product enhancements definitely appeal to consumers—they may, for instance, increase the amount of time someone spends on a web site—but they don’t produce a direct uptick in sales or revenue. If smaller companies create similar offerings, they often find that sales are limited or nonexistent because consumers gravitate to free solutions. Large players also have access to a significantly larger pool of consumer data—the lifeblood of AI—which allows them to develop more accurate and insightful AI solutions for consumers. With the free products from large players winning most market share, AI value capture will be limited in the consumer sector, in the immediate term.
This may not be the case in the future, however, since newer, fee-based offerings are entering the market, including in-home assistants. The next wave of consumer AI will see even more innovation as automakers and others introduce new products. Take autonomous cars. Some consumers may be content with vehicles in which AI enables autonomous braking, but others will want more features, such as complete self-driving capabilities, even if they must pay a premium.
2. Enterprise winners will focus on microverticals in promising industries
Our early analysis of data from McKinsey Global Institute, combined with expert interviews and research, revealed nearly 600 discrete uses for AI across major industries. Of these, about 400 require some level of ML and 300 require DL capabilities. Many of the most interesting AI applications are still in the pilot stage and haven’t been deployed at scale yet. Here are a few AI applications that could see high demand over the next few years because of their strong visual-perception and processing capabilities:
·         Governments can use AI to scan video and identify suspicious activity in public places, or apply AI algorithms to detect potential cyberattacks. Many military applications, including drones, also rely on AI. Beyond security, AI is finding a role in traffic control, including sensors and cameras that allow light signals to change their timing and sequence based on the number of cars on the road.
·         As with the public sector, banks are beginning to use AI to detect suspicious behavior, such as patterns suggestive of money laundering. AI algorithms can also help process transactions and make decisions, often with greater accuracy than human employees. For instance, AI algorithms might reveal that certain overlooked characteristics increase the odds that a particular transaction is fraudulent.
·         Within retail, AI is already helping with theft detection and it could bring further enhancements to automated checkouts. Several retailers are piloting systems that use cameras and sensors to detect when shoppers take or return items from the store. After the customers leave the store, their accounts are charged for the total. Other retailers use in-store video to optimize sales associates’ coverage. If cameras detect a shopper lingering before a display, the system notifies an associate to provide assistance. In the future, we could see even more enhancements in this area, including AI systems that identify customers with high purchase potential by looking at various characteristics—facial expression (as a signifier of mood), clothing, and number of companions. They could then alert associates about the location of these shoppers within the store.

Companies face a difficult task when deciding which opportunities to pursue, among the hundreds available, but they can narrow their options through a structured approach. The first step involves picking an industry focus. It’s true that a company’s expertise and capabilities will influence this decision, but players should also consider industry characteristics, including the sector’s size. Also important is the potential for disruption within an industry, which we estimated by looking at the number of AI use cases, start-up equity funding, and the total economic impact of AI, defined as the extent to which solutions reduced costs, increased productivity, or otherwise benefited the bottom line in a retrospective analysis of various applications. The greater the economic benefit, the more likely that customers will pay for an AI solution.
Just as AI value varies by industry, so does maturity. For instancethe industrial sector could gain big from AI, but member companies are not as ready to embrace these solutions as their counterparts in the automotive industry. For producers of AI products and services, this means that value capture will be staggered, with some industries initially producing higher returns than others.
When we considered value at stake in combination with maturity, it became clear that several industries now offer the strongest opportunities for AI: public sector, banking, retail, and automotive . While the public sector’s prominence may seem surprising in an age where governments are cutting budgets, many officials see the value of AI in improving efficiency and efficacy, and they are willing to provide funding. As they plan their AI strategies, suppliers should focus their investments on potential consumers of AI solutions who are willing to be the first domino.
Microverticals. Once companies have chosen one industry, or a few, as their focus, they must go a step further by selecting particular use cases—which we call microverticals—where they will concentrate. Buyers aren’t interested in AI just because it’s an exciting new technology—instead, they want AI to generate a solid return on investment (ROI) by solving specific problems, saving them money, or increasing sales. For instance, a manufacturing plant that wants to reduce machine downtime won’t simply look for an AI provider that’s well known in the industrial space; it will instead seek a company with proven predictive-maintenance expertise and solutions. If an AI provider tried to offer a horizontal solution—one that customers could apply across a variety of unrelated use cases—the value proposition would not be as compelling. End customers would question whether the solution’s ROI could justify its greater expense, especially if it applied to several use cases that they considered unimportant or irrelevant.
CONTINUES

By Gaurav Batra, Andrea Queirolo, and Nick Santhanam January 2018

https://www.mckinsey.com/industries/advanced-electronics/our-insights/artificial-intelligence-the-time-to-act-is-now?cid=other-eml-alt-mip-mck-oth-1801&hlkid=2a9d89c8c3ca4764a09bb204647e16b4&hctky=1627601&hdpid=094bf139-fd76-400b-8896-e6caf6e19d81



No comments: