What does Machine Learning Mean for Today’s Entrepreneur?
What is Machine Learning?
Machine learning is a form of artificial intelligence in which a computer is able to learn and improve its programs and processes based on its experience. With machine learning, a computer program accesses data on its own and uses it without being explicitly programmed for these tasks.
In the process of machine learning, a computer starts with data gained through experience, instruction or observation and looks for patterns so it can improve the choices it makes going forward. As computers begin to learn automatically, they can continue to improve their abilities with limited or even no human intervention.
As a result of the mainstreaming of machine learning over the past year or so, tasks that once had to be done manually can now be automated. In addition, major companies and startups alike are creating services and products through machine learning.
How are Entrepreneurs Using Machine Learning?
Machine learning is starting to pervade numerous firms, and not just digital natives such as Google or Amazon. Approximately $20 to $30 billion was invested in AI and machine learning in 2016 alone, with about 60 percent of that coming from outside investors, and more is pouring into AI ventures currently. [Bughin, Hazan, et al. Artificial Intelligence: The Next Digital Frontier. McKinsey Global Institute, McKinsey & Company. 2017]
Take a look at some of the varied entrepreneurial fields using machine learning: • FinTech. PayPal is currently using machine learning to push back against money laundering operations. Onfido is applying machine learning to the world of background checks. Sift’s AI is focused on preventing online fraud and chargebacks. • Consumer Services. Forkable’s Lunchbot uses machine learning algorithms to figure out what office workers want for lunch, then delivers it. Target continues to innovate its consumer marketing, using machine learning to target advertising narrowly towards its customers. Nova uses machine learning to write personalized sales emails, tweaking the content as the computer learns what works best. • IT. Dark Trace’s machine learning algorithms listen to your network traffic, using the data to pinpoint emergent threats. Deep Instinct is similarly focused on detecting malware through machine learning algorithms. • Publishing.Pinterest uses machine learning to suggest highly personalized pins to individual users. • Agriculture. Blue River Technology is using computer vision paired with machine learning to diagnose ailing plants and to propose and deliver treatments.
• Health Care. Much of the focus in the health care field is on detecting disease. Medecision can predict hospitalization in diabetes patients, allowing preventative care in time to help them. Zebra Medical Vision is focused on predicting and preventing various diseases.
• Legal Services. TrademarkVision helps startups to work their way through the potential legal confusion involved with getting new ideas off the ground. In any of these fields, the basic machine learning workflow follows similar steps.
1. Collecting Data. Machine learning relies on data — lots of data. The better the quality and the higher the quantity of the data, the more the machine learning process is likely to be successful.
2. Preparing Data. Any data set will contain outliers and missing data. Analysis of the data’s nuances helps streamline the path the computer will have to take.
3. Building the Model. Developing the appropriate algorithms to crunch the data is a key element of machine learning. The prepared data is used to train the model.
4. Evaluating the Model. Once the machine learning model has been built and tried, it must be tested for accuracy. Using new data at this point can help to verify the algorithm’s precision.
5. Improving the Model’s Performance. New algorithms may be needed, or new data may need to be tested to verify the efficiency and accuracy of the model before turning the computer loose on real data.
What Challenges Do Entrepreneurs Face Regarding Machine Learning? Surprisingly, one of the challenges many entrepreneurs face regarding machine learning is dealing with data. There’s simply so much of it. Data is collected on every detail of every customer interaction, and the growth in raw data is only going to get exponentially stronger. Experts predict a 4300 percent increase in data generated over the next three years. Sifting through all that data and determining what to focus on — or what to ask computers to focus on — is an increasingly vital challenge.
In addition, even with the AI field burgeoning, there aren’t enough experts to go around. As a result, the costs of hiring true AI and machine learning experts is increasing rapidly. One response to these increased expenses is the phenomenon of “acqui-hires,” in which a large company will swallow up a smaller company just to gain access to its AI experts.
Add to this the fact that the entire machine learning field is growing with extreme speed, and there are plenty of hurdles for entrepreneurs to overcome.
What’s the Future of Machine Learning for Entrepreneurs?
Entrepreneurs looking to start new companies in the AI or machine learning space need to focus on a few specific areas. • Reducing human labor. Entrepreneurs should look for ways to automate tasks that have traditionally been done by humans and that are considered hard to automate. Amazon’s constant efforts to improve door-to-door delivery are an example of this type of automation.
• Real-time optimization. Where do people need immediate, real-time data to make decisions? Uber takes advantage of machine learning in choosing routes, and both Waze and Google Maps put this same functionality in the hands of everyday drivers.
• New products and services. As machine learning continues to deepen its development, products or services that were previously not cost-effective will become easy. Looking for this “white space” is a potentially profitable path for entrepreneurs.
• Radical personalization. Large companies like Target and Amazon are already optimizing data to personalize customer engagement, and Google Adwords provides valuable data to help much smaller entrepreneurs step into this space as well. • Predictive analysis. This is already happening in the automotive and health care industries, where machine learning algorithms can let you know when your car — or your body — is about to break down. Entrepreneurs who can translate these machine learning abilities to other specific analytical ventures may find plenty of willing users. As you explore entrepreneurial opportunities in machine learning, look for the expertise to help you make wise and forward-looking decisions, while staying abreast of what the competition is doing and planning a sound business strategy.