Fabrinet provides complex optical and electro-mechanical components, modules, and bulk optics. These components serve the markets of data communications, telecommunications, networking, medical devices, and automotive technologies.

Marc Andreessen had famously said software is eating the world. He probably had Artificial Intelligence (AI) in the back of his mind. In its simplistic form, AI enables a machine to perform human-like tasks, such as image, voice, and text recognition, natural language processing and understanding human-like perception. The journey of AI from Expert Systems in the early eighties to Heuristics analysis, machine learning, and finally to present day deep learning has been a roller coaster ride. Just a few years back, people thought neural networks were something academicians talked in their leisure time. This change, where AI is becoming more and more mainstream and affordable, has come about with the convergence of big data, availability of parallel processing advancements such as GPUs on public clouds and breakthroughs in machine learning.

Machine learning is set of algorithms that enable a computer program to recognize patterns in data sets and interpret those patterns to provide meaningful insights. Machine learning can be supervised or unsupervised. In the supervised learning you are training machine learning task for every input with corresponding target output. In supervised learning, machine is trained with labeled data and looks at data with specific parameters. Human input and bias are an ingredient of the supervised learning making it more expensive and limited in use.

In the unsupervised learning the information is classified without the help of trainers or instructors. The machine finds structure or relationships among different inputs. One example of unsupervised learning is clustering where new input data is automatically put into an appropriate cluster. The affordable processing power and storage coupled with explosion of Big Data from multitude of sources ­ such as text, images, and connected devices ­ is making it easier for machines to train and learn in the unsupervised mode.

Unsupervised learning is a precursor to deep learning where most of the benefits for an enterprise will be realized using AI. Deep learning systems can learn from iterative data computations. They just don't follow

explicitly programmed instructions. Deep learning algorithms get more intelligent and context aware with use and experience, making them a key enabler of artificial intelligence platforms.

AI in the form of Facebook targeted content feed, Google Alpha-Go, and smart assistant like Alexa have grabbed our attention. The enterprise inflection point would be when AI is incorporated into strategic business applications or business domains. This will help to raise employee productivity, improve, and automate business processes, detect fraud, build smarter factories, and connected vehicles, make better recommendations, anticipate customer sentiment, and even address cyber security. The digitally transformed businesses would be "algorithm driven business" that would use machine learning to drive process automation and improve decision making. These will be the businesses that would reap the benefits of modern day "gold rush" where deep understanding of their data paves the way for innovative business models, products, and services.

AI can enable better outcomes by eliminating human error and faster decision making that is adaptable to changing business conditions by simply tweaking an algorithm

There are major enterprise soft-ware providers that are already working on bringing AI into their core applications. They already have the advantage of having vast amount of digital data and interactions in the form of consumer profiles, transactions, and business outcomes. Once the data is anonymized; they can make current applications AI adaptive by continually capturing and learning from new data and tapping transactional and behavioral history.

One example is in the HR arena in finding the right talent match for open positions. The talent management solution driven with natural language processing and understanding and predictive language analysis will help speed up recruitment by allowing you to focus on just not on keywords but the general sentiment of the resume and social media profiles. The whole recruitment process can be done with fewer mistakes and be more equitable, accountable, and compliant.

Another enterprise application of AI is processing the enormous volume of data and information flows generated by the new generation of connected IoT networks. Because machine learning algorithms get smarter as they are ex-posed to more data, these deep learning platforms are key to finding insights in the data flows generated by Industrial IoT networks. Machine learning systems can detect the anomalies or patterns outside the norm and create self-healing behavior or alert a human for corrective action.

Augmented Reality is another area that is fueled by the advancement in AI where, the way we inter-act with everything will be rewritten and new business models will be created.

The Tesla Autopilot enabled automobiles are collecting data from millions of miles driven by their drivers in real life situations. These videos and data are fed to a deep learning engine on the cloud to create a terminology of autonomous driving. As per Elon Musk, the whole Tesla fleet operates as a network. When one car learns some-thing, they all learn it. The same model can be applied to enterprise software. The algorithms can be continuously enhanced and made to reflect on edge devices by constant over the air upgrades. They can help in building conversational interfaces into any applications using voice and text and create highly engaging user experience by constantly learning from the network.

For an enterprise, AI can enable better outcomes by eliminating human error and faster decision making that is adaptable to changing business conditions by simply tweaking an algorithm.