In machine learning (ML), machines—computer programs—learn and improve based on the assessment of historical data without being directed to do so. This process allows them to improve the accuracy of predictions or decisions they make.

ML is part of the wider field of artificial intelligence. But, unlike AI which seeks to mimic human intelligence, ML is focused on a limited range of specific tasks. The ML concept is already being used in areas like drug discovery¹. For example, last year GSK² shared details of its use of ML in vaccine development. Likewise, in July the Gates Foundation awarded A-Alpha Bio $800,000 to use machine learning to optimize protein therapeutics for infectious diseases.³

But ML also has potential in the process development suite and on the factory floor according to Moritz von Stosch⁴, chief innovation office at analytics and process modeling firm DataHow.

“ML can be used for development, monitoring, control and optimization. ML is better at learning complex relationships—for example between process parameters and process performance—than humans and it can make better predictions of what might be happening for slightly different scenarios.”

Quality data costs

Data quality is key to any ML strategy and this is the biggest challenge for the biopharmaceutical industry, von Stosch says.

“Data quality is typically poor, data pre-processing anecdotally requiring 80% of the total effort in any machine learning project. Data quantity is rather low as in bioprocess development the generation of data costs money,” he continues. “Generally in bioprocess development we face the curse of dimensionality because of the large number of parameters and getting informative data for all possible parameter combinations is impossible, wherefore we need to add process knowledge to ML.”

Combining process knowledge to ML—which von Stosch calls⁵ “hybrid modeling”—aims to generate more insights than MLs alone to increase process understanding and develop models that can better forecast system behavior.

As well as high quality data, biopharmaceutical companies thinking about using ML need to make clear which parameters they want to model.

“The key” von Stosch said is “framing the problem, such that it is concise and can be solved by a machine. He explained that providing the learning algorithm with the correct parameters is critical. “For instance, if you train the machine with data of sugar content of grape juice and the alcohol content after it has been fermented, then the machine is able to predict the alcohol content if you provide with the sugar content of a novel grape juice,” he pointed out. “However, it will not be able to predict the alcohol content based on the mass of grapes that was used to make the juice.”

The post Utilizing Machine Learning for Better Bioprocess Development appeared first on Artificial Intelligence.

Machine learning does exactly what it says on the tin. It is a method by which a computer program can “automatically learn and improve from experience without being explicitly programmed.”

Now, what does that mean? Let’s breakdown a few concepts.

PREDICTIVE MODELING AND LEARNING

Consider how a living creature learns something. Say there’s a mail carrier that always brings a pocketful of dog treats on his rounds. Whenever he comes to a house with a dog, he drops one of those treats into the mail slot along with the mail. The dog inside the house recognizes the scent of the mailman, and knows that he comes to the house around 2 p.m. every day.

After a few days of receiving treats in the mail at the same time, the dog begins to understand the pattern: 2 p.m. = mailman = treats. The dog adjusts its behavior accordingly, getting excited and sitting near the door at the same time each day, then barking like crazy when the mailman gets there.

Naturally, there are anomalies in the dataset. Sometimes a different mailman comes who doesn’t bring treats, or sometimes the mailman is running behind schedule. And there’s no mail on Sundays.

The dog is not deterred by exceptions to the pattern, however, because the predictive model it has created is correct more often than not. If the dog were very intelligent, it might know to check that the correct mailman was coming before getting excited, and recognize that nobody comes every seventh day, making this predictive model even more accurate.

In broad strokes, a computer program using machine learning follows the same method. It analyzes data and searches for an underlying pattern or trend to develop a predictive model that learns from the data it’s fed.

ALGORITHMS

Any machine learning program needs a “training dataset” to teach it what kind of information it can expect, and from which to begin noticing the kind of information the programmer is looking for.

The difference between a dog and a computer program is, of course, the volume and complexity of the input.

Machine learning algorithms can process massive amounts of data and predict outcomes and patterns based on that information. Over time, the predictive model becomes more accurate as the program improves itself, no outside tampering required.

There are three broad categories of algorithms, which are defined by what kind of training datasets they are given: supervised, unsupervised, and semi-supervised.

Each of these approaches has advantages and disadvantages, depending on what the program is intended to accomplish.

SUPERVISED LEARNING

Overview: Supervised machine learning algorithms are trained on datasets where a given input leads to a specific output according to a mapping function.

How it works: According to Jason Brownlee, “the goal is to approximate the mapping function so well that when you have new input data (X) that you can predict the output variables (Y) for that data.”

The programmer is in this instance like a teacher who gives their student a quiz. The teacher knows the correct answers, and grades the student each time they are quizzed. The student keeps taking quizzes until they pass consistently.

Supervised problems can be grouped into two types:

• Classification is sorting output into categories. An example of a classification problem is a spam filter for your email. The program reads the emails, and classifies them as spam or not-spam based on their content.

• Regression problems, on the other hand, return an output that can be measured. For example, a program that calculates how many gallons of gas a car requires on a road trip given the distance and model of the car would require a regression algorithm.

UNSUPERVISED LEARNING

Overview: Unsupervised machine learning has no correct output for the given input. Unlike supervised machine learning, there is no expected answer, and there is no teacher, just the program plugging away at the data on its own.

The goal of this type of machine learning is to analyze the data as a whole, and discover facts about the underlying structure.

How it works: When an unsupervised algorithm analyzes data, it is usually for one of two purposes:

• In a clustering problem the goal is to find particular groups within a dataset. Discovering your customers’ age and income distribution is an example of a clustering problem, as the program can show you which age and income groups are most common.

• An association problem is more focused on finding rules or patterns that govern a dataset. When you analyze a customer’s flow through your website, checking which links they are most drawn to, that’s rule association.

SEMI-SUPERVISED LEARNING

Overview: Semi-supervised machine learning is, unsurprisingly, a combination of the first two types.

How it works: Techniques of supervised and unsupervised machine learning can be used in the same problem.

For example, one could make predictions about a dataset using an unsupervised algorithm and feed the results to a supervised algorithm.

Semi-supervised machine learning doesn’t have any defined subcategories, but is most useful when your dataset is a mix of labeled and unlabeled data points.

Real-world problems, like classifying a collection of physical photographs, may be best solved by semi-supervised machine learning.

MACHINE LEARNING USE CASES

The possible applications and advantages of this technology are numerous.

Very generally, we need machine learning if we want to accomplish a task that requires human-like adaptability, or is too large to scale. It also allows us to create an analytical model that is free of human bias, at least in theory.

Tasks that humans can learn to do automatically—such as understanding spoken words, judging road conditions, and recognizing people in a photograph—don’t come easily to a typical computer program because it would need to learn from experiences as a human does.

Machine learning is designed to mimic human intelligence within set parameters. Every iteration helps the program improve its accuracy and ability to perform whatever task it’s meant to do.

SCALING INTELLIGENCE

Human brains are marvelous data processors, but with limits. A human being could never do what a search engine does, for example, because there’s more information on the internet than a person can process.

A machine learning program can accomplish a task that most humans could do, such as search a web page for keywords, but do it on a scale that only computers can process.

The post WHAT IS MACHINE LEARNING? A QUICK GUIDE TO BASIC CONCEPTS appeared first on Artificial Intelligence.

The next time you get sick, your care may involve a form of the technology people use to navigate road trips or pick the right vacuum cleaner online.

Artificial intelligence is spreading into health care, often as software or a computer program capable of learning from large amounts of data and making predictions to guide care or help patients.

It already detects an eye disease tied to diabetes and does other behind-the-scenes work like helping doctors interpret MRI scans and other imaging tests for some forms of cancer.

Now, parts of the health system are starting to use it directly with patients. During some clinic and telemedicine appointments, AI-powered software asks patients initial questions about their symptoms that physicians or nurses normally pose.

And an AI program featuring a talking image of the Greek philosopher Aristotle is starting to help University of Southern California students cope with stress.

Researchers say this push into medicine is at an early stage, but they expect the technology to grow by helping people stay healthy, assisting doctors with tasks and doing more behind-the-scenes work. They also think patients will get used to AI in their care just like they’ve gotten accustomed to using the technology when they travel or shop.

But they say there are limits. Even the most advanced software has yet to master important parts of care like a doctor’s ability to feel compassion or use common sense.

“Our mission isn’t to replace human beings where only human beings can do the job,” said University of Southern California research professor Albert Rizzo.

Rizzo and his team have been working on a program that uses AI and a virtual reality character named “Ellie” that was originally designed to determine whether veterans returning from a deployment might need therapy.

Ellie appears on computer monitors and leads a person through initial questions. Ellie makes eye contact, nods and uses hand gestures like a human therapist. It even pauses if the person gives a short answer, to push them to say more.

“After the first or second question, you kind of forget that it’s a robot,” said Cheyenne Quilter, a West Point cadet helping to test the program.

Ellie does not diagnose or treat. Instead, human therapists used recordings of its sessions to help determine what the patient might need.

“This is not AI trying to be your therapist,” said another researcher, Gale Lucas. “This is AI trying to predict who is most likely to be suffering.”

The team that developed Ellie also has put together a newer AI-based program to help students manage stress and stay healthy.

Ask Ari is making its debut at USC this semester to give students easy access to advice on dealing with loneliness, getting better sleep or handling other complications that crop up in college life.

Ari does not replace a therapist, but its designers say it will connect students through their phones or laptops to reliable help whenever they need it

USC senior Jason Lewis didn’t think the program would have much for him when he helped test it because he wasn’t seeking counseling. But he found that Ari covered many topics he could relate to, including information on how social media affects people.

“Everybody thinks they are alone in their thoughts and problems,” he said. “Ari definitely counters that isolation.”

Aside from addressing mental health needs, artificial intelligence also is at work in more common forms of medicine.

The tech company AdviNOW Medical and 98point6, which provides treatment through secure text messaging, both use artificial intelligence to question patients at the beginning of an appointment.

AdviNOW CEO James Bates said their AI program decides what questions to ask and what information it needs. It passes that information and a suggested diagnosis to a physician who then treats the patient remotely through telemedicine.

The company currently uses the technology in a handful of Safeway and Albertsons grocery store clinics in Arizona and Idaho. But it expects to expand to about 1,000 clinics by the end of next year.

Eventually, the company wants to have AI diagnose and treat some minor illnesses, Bates said

Researchers say much of AI’s potential for medicine lies in what it can do behind the scenes by examining large amounts of data or images to spot problems or predict how a disease will develop, sometimes quicker than a doctor.

Future uses might include programs like one that hospitals currently use to tell doctors which patients are more likely to get sepsis, said Darren Dworkin, chief information officer at California’s Cedars-Sinai medical center. Those warnings can help doctors prevent the deadly illness or treat it quickly.

“It’s basically that little tap on the shoulder that we all want to get of, ‘Hey, perhaps you should look over here,’” Dworkin said.

Dr. Eric Topol predicts in his book “Deep Medicine” that artificial intelligence will change medicine, in part by freeing doctors to spend more time with patients. But he also notes that the technology will not take over care.

Even the most advanced program cannot replicate empathy, Topol said. Patients stick to their treatment and prescriptions more and do better if they know their doctor is pulling for them.

Artificial intelligence also can’t process everything a doctor considers when deciding on treatment, noted Harvard Medical School’s Dr. Isaac Kohane. That might include a patient’s tolerance for pain or the desire to live a few more months to attend a child’s wedding or graduation.

“Good doctors are the ones who understand us and our goals as human beings,” he said.

The post Paging Dr. Robot: Artificial intelligence moves into care appeared first on Artificial Intelligence.

Artificial intelligence is touching almost every industry in the ongoing era of digital transformation. But what does it mean? Put simply, AI and machine learning — a closely related idea — are computer programs that emulate and learn human behavior to then make autonomous decisions. AI can be especially transformative in retail, helping almost every aspect of a retailer’s business — from utilizing data to drive sales, to automating the supply chain, or to hyper-personalize the customer experience.

While AI can sound like some far-off idea, it’s closer than you may think. Juniper Research, for example, predicts retailers will spend $7.3 billion on AI by 2022, compared with the approximately $2 billion that was spent in 2018.

AI is quickly becoming more appealing than conventional software, which generally requires someone to think about all the possible ways you might want it to do something, and then write programs or code for each one of those tasks.

Having systems that utilize AI to continually learn and train themselves allows technology to be much more powerful and adaptable to new or changing experiences. While there are many benefits to investing in AI technology, one of the most important for retailers will be the positive impact on customer experience.

Below are some of the ways AI can impact how merchants provide their customers with personalized experiences:

•    Mobile shopping assistants powered by AI — One example of an in-store deployment of AI is Macy’s On Call, a shopping assistant powered by IBM’s Watson. Customers can access the bot via an app and then interact with the assistant as they would with a human store associate. The more conversations the bot has, the smarter it becomes. There are many examples of how a real-time intelligent reaction from a bot can help ensure a positive customer experience — one being, if a bot could detect signs of stress or anger on the customer’s face and then alert a human employee to go help immediately.

•    Customer service driven by bots – Stores are also implementing physical robots for a variety of reasons, and the use cases for in-store robots only grows as AI becomes more advanced. Can’t find a product? The robot can help. Are more products available in the back? The bot can give you that answer. Softbank created Pepper the AI robot, which can chat with and direct customers, accept secure payments and even just create a novelty experience that can draw in someone passing by. Pepper helped the company report strong numbers in terms of increased revenue and foot traffic after the launch of several pilots.  

•    AI’s impact online – With what the U.S. Department of Commerce estimates to be $513.61 billion dollars spent online every day, AI isn’t just about improving in-store shopping. There are plenty of ways to incorporate AI into your customer’s online shopping experience as well. Through the power of AI, companies can make recommendations to customers shopping online by harnessing information such as past purchases or browsing history. This helps to cut down on the time it takes for a customer to find items that fit in with what they are looking for.

•    Adjusting prices, intelligently — AI can also make a difference in price setting. Algorithms can provide retailers statistically-likely outcomes of various pricing strategies, allowing them to offer the best promotional deals to increase sales and to make — and keep — their customers happy. Through the collection and analysis of vast amounts of internal and competitor data, AI can help support the optimal pricing decision. eBay uses AI-powered pricing and inventory algorithms to help sellers determine the most appropriate prices, and Kroger applies AI in a similar way as well. Having this real-time data analysis provides retailers with the ability to stay flexible, as well as to alter prices and other promotions instantly based on insights from shoppers.

•    Bridging the gap between virtual and physical channels — AI caters to the omnichannel shopper — it isn’t limited to either the physical or online shopping experience. Imagine this: A friend is admiring your shoes. She uploads a photo of them to a retail app on her mobile device, which utilizes image recognition AI to identify and locate the product. The app can then bring her to the product on a store’s website, while also recommending nearby locations that carries them. AI can connect customers across the digital and offline channels of stores, giving them a seamless experience as well as purchasing options.

Personalization of the customer experience across sales channels is something that will only increase in popularity as consumers continue to demand it. In order to stay relevant and successful, retailers need to be aware of this, and use tools like AI to ensure the experience is a positive one. AI is a game-changer, and utilizing it can transform the retail experience for customers.

The post AI: Putting the shopper experience on tech steroids appeared first on Artificial Intelligence.