Machine learning (ML)—the artificial intelligence (AI) subfield in which machines learn from datasets and past experiences by recognizing patterns and generating predictions—is a $21 billion global industry projected to become a $209 billion industry by 2029. Here are some real-world applications of machine learning that have become part of our everyday lives.
Machine learning in marketing and sales
According to Forbes, marketing and sales teams prioritize AI and ML more than any other enterprise department. Marketers use ML for lead generation, data analytics, online searches and search engine optimization (SEO). For example, many use it to contact users who leave products in their cart or exit their website.
ML algorithms and data science are how recommendation engines at sites like Amazon, Netflix and StitchFix make recommendations based on a user’s taste, browsing and shopping cart history. ML also helps drive personalized marketing initiatives by identifying the offerings that might meet a specific customer’s interests. Then, it can tailor marketing materials to match those interests. ML also provides the ability to closely monitor a campaign by checking open and clickthrough rates, among other metrics.
Customer service use cases
Not only can ML understand what customers are saying, but it also understands their tone and can direct them to appropriate customer service agents for customer support. Voice-based queries use Natural Language Processing (NLP) and sentiment analysis for speech recognition.
Text-based queries are usually handled by chatbots, virtual agents that most businesses provide on their e-commerce sites. Such chatbots ensure that customers don’t have to wait, and even large numbers of simultaneous customers can get immediate attention around the clock and, hopefully, a more positive customer experience. One bank using a watsonx Assistant system for customer service found the chatbot answered 96% of all customer questions correctly, quickly, consistently, and in multiple languages.
Businesses use ML to monitor social media and other activity for customer responses and reviews. ML also helps businesses forecast and decrease customer churn (the rate at which a company loses customers), a widespread use of big data.
Personal assistants and voice assistants
It’s ML that powers the tasks done by virtual personal assistants or voice assistants, such as Amazon’s Alexa and Apple’s Siri. This communication can involve speech recognition, speech-to-text conversion, NLP, or text-to-speech. When someone asks a virtual assistant a question, ML searches for the answer or recalls similar questions the person has asked before.
ML is also behind messaging bots, such as those used by Facebook Messenger and Slack. At Facebook Messenger, ML powers customer service chatbots. Companies set up chatbots there to ensure fast responses, provide carousels of images and call-to-action buttons, help customers find nearby options or track shipments, and allow secure purchases. Facebook also uses ML to monitor Messenger chats for scams or unwanted contacts, such as when an adult sends a great deal of friend or message requests to people under 18.
At Slack, ML powers video processing, transcription and live captioning that’s easily searchable by keyword and even helps predict potential employee turnover. Some companies also set up chatbots on Slack, using ML to answer questions and requests.
Filtering email
ML algorithms in Google’s Gmail automate filtering customers’ email into Primary, Social and Promotions categories while also detecting and rerouting spam to a spam folder. Going beyond email app “rules,” ML tools can also automate email management by classifying emails to route them to the right people for faster action, moving attachments to the right place, and more. For instance, email management automation tools such as Levity use ML to identify and categorize emails as they come in using text classification algorithms. This allows you to craft personalized responses based on category, which saves time, and such customization can help improve your conversion rate.
Machine learning and cybersecurity use cases
There are four ways ML is being used in cybersecurity:
- ML and facial recognition are used in authentication methods to protect an enterprise’s security.
- Antivirus programs may use AI and ML techniques to detect and block malware.
- Reinforcement learning uses ML to train models to identify and respond to cyberattacks and detect intrusions.
- ML classification algorithms are also used to label events as fraud, classify phishing attacks and more.
Machine learning in financial transactions
ML and deep learning are widely used in banking, for example, in fraud detection. Banks and other financial institutions train ML models to recognize suspicious online transactions and other atypical transactions that require further investigation. Banks and other lenders use ML classification algorithms and predictive models to determine who they will offer loans to.
Many stock market transactions use ML. AI and ML use decades of stock market data to forecast trends and suggest whether to buy or sell. ML can also conduct algorithmic trading without human intervention. Around 60-73% of stock market trading is conducted by algorithms that can trade at high volume and speed. ML algorithms can predict patterns, improve accuracy, lower costs and reduce the risk of human error.
The nonprofit tech organization Change Machine worked with IBM to build an AI-powered recommendation engine using IBM Cloud Pak® for Data that helps financial coaches find fintech products best suited to its customers’ goals. “The engagement with IBM taught us how to leverage our data in new ways and how to build a framework for creating and managing machine learning models,” said David Bautista, Director of Product Development at Change Machine.
Machine learning in healthcare
ML developments led to training machines in pattern recognition, which is now sometimes used in radiology imaging. AI-enabled computer vision is often used to analyze mammograms and for early lung cancer screening. Doctors evaluating mammograms for breast cancer miss 40% of cancers, and ML can improve on that figure. ML is also trained and used to classify tumors, find bone fractures that are hard to see with the human eye and detect neurological disorders.
ML is sometimes used to examine historical patient medical records and outcomes to create new treatment plans. In genetic research, gene modification and genome sequencing, ML is used to identify how genes impact health. ML can identify genetic markers and genes that will or will not respond to a specific treatment or drug and may cause significant side effects in certain people. These advanced analytics can lead to data-driven personalized medication or treatment recommendations.
The discovery and manufacturing of new medications, which traditionally go through involved, expensive and time-consuming tests, can be sped up using ML. Pfizer uses IBM Watson’s ML capabilities to choose the best candidates for clinical trials in its immuno-oncology research. Geisinger Health System uses AI and ML on its clinical data to help prevent sepsis mortality. They are working with IBM’s Data Science and AI Elite team to build models that predict which patients are at greatest risk for sepsis, which helps them prioritize care, decrease risky and expensive inpatient admissions and lower sepsis mortality rate.
Machine learning and transportation
ML informs a lot of our transportation these days. For instance, Google Maps uses ML algorithms to check current traffic conditions, determine the fastest route, suggest places to “explore nearby” and estimate arrival times.
Ride-sharing applications like Uber and Lyft use ML to match riders and drivers, set prices, examine traffic and, like Google Maps, analyze real-time traffic conditions to optimize the driving route and predict an estimated arrival time.
Computer vision fuels self-driving cars. An unsupervised ML algorithm lets self-driving cars gather data from cameras and sensors to understand what’s happening around them and enables real-time decision-making on actions to take.
Machine learning in smartphones
ML powers a lot of what happens with our smartphones. ML algorithms govern the facial recognition we rely on to turn on our phones. They power the voice assistants that set alarms and compose messages. These include Apple’s Siri, Amazon’s Alexa, Google Assistant and Microsoft’s Cortana, which use NLP to recognize what we say and respond appropriately.
Companies also take advantage of ML in smartphone cameras. ML analyzes and enhances photos using image classifiers, detects objects (or faces) in the images, and can even use artificial neural networks to enhance or expand a photo by predicting what lies beyond its borders.
Machine learning and apps
We see lots of ML use on social media platforms today:
- Social media, such as Facebook, automates friend-tagging suggestions by using ML face detection and image recognition to identify a face in its database. Then, it suggests the social media user tag that individual.
- LinkedIn uses ML to filter items in a newsfeed, make employment recommendations and suggest that someone connect with others.
- Spotify uses ML models to generate its song recommendations.
- Google Translate uses NLP to translate words across more than 130 languages. In some languages, it can provide translations via photo, voice in “conversation mode” and through live video images using augmented reality mode.
AI can help strategize, modernize, build and manage existing applications, too, leading to more efficiency and creating opportunities for innovation. Sonoma County, California, consulted with IBM to match homeless citizens with available resources in an integrated system called ACCESS Sonoma. “Because IBM designed this open architecture that literally could be lifted and shifted, we loaded 91,000 clients and linked them across four key systems in four months,” said Carolyn Staats, Director of Innovation, Sonoma County Central IT. “That’s an amazing timeline.” They placed 35% of homeless people in housing, four times higher than the national rate, and in two years, the County reduced its number of homeless people by nine percent.
Machine learning and IBM
At IBM, we are combining the power of ML and AI in IBM watsonx, our new studio for foundation models, generative AI and ML.
Watsonx is a next-generation data and AI platform built to help organizations multiply the power of AI for business. The platform has three powerful components: the watsonx.ai studio for new foundation models, generative AI and ML; the watsonx.data fit-for-purpose store for the flexibility of a data lake and the performance of a data warehouse; and the watsonx.governance toolkit to enable AI workflows built with responsibility, transparency and explainability.
Together, watsonx offers organizations the ability to:
- Train, tune and deploy AI across your business with watsonx.ai
- Scale AI workloads anywhere, for all your data, with watsonx.data
- Enable responsible, transparent and explainable data and AI workflows with watsonx.governance
