Artificial Intelligence
Things to think about while learning Artificial Intelligence.
1. Think about ways you can apply what you're learning to your daily work. 2. What conversation should you be having with your team, your leads, your clients? 3. What ways can AI help you think of new innovative ways of working or new business ideas? 4. What can you learn from the client stories?
How Is AI Applied?
A business can choose to implement artificial intelligence in three different ways. 1. They can build their own custom solution, 2. They can buy pre‑constructed building blocks of a solution, or 3. they can partner with other businesses that have already built the relevant capabilities.
AI winter
A period of reduced funding and interest in artificial intelligence research. The field has experienced several cycles of hype, followed by disappointment and criticism, followed by funding cuts, followed by renewed interest years or decades later. There were two major instances in 1974-80 and 1987-93.
Natural language processing (NLP)
A technology that converts human language (structured or unstructured) into data that can be translated then manipulated by computer systems; branch of artificial intelligence. that area of artificial intelligence that deals with recognizing, understanding, analyzing, and even emulating the typical ways that humans communicate using either voice or text or both
Applied intelligence
A term that Accenture coined, encapsulates more than just artificial intelligence. It's about data, automation, capturing the business process, the experience of the design, and tailoring it all for specific industries and clients to drive outcomes. Applied intelligence is this sum of that whole package of capabilities all wrapped together. how you generate the impact with AI, artificial intelligence, but also with analytics, automation, and then all powered by data.
The AI effect
And as soon as we've ever figured out how to do that thing, the general response becomes okay, but that's not really intelligence, is it? The computer isn't actually thinking, it's just executing a formula, it's just a computation. They begin as some kind of artificial intelligence research, but as soon as any part of it became implemented, practical, and widespread, nobody will call it AI anymore
What Is Accenture's Role with AI?
Applied intelligence is the name that we at Accenture have given to our unique approach to packaging artificial intelligence and how to generate real impact and big outcomes with it. Our deep industry expertise is then combined with: • Advanced Analytics, • automation, and • AI services powered by data to help our customers to transform their businesses with new agility from front office to back office. Sense, comprehend, act, and learn. Human plus machine. Smarter tech, smarter systems paired with people to make people more effective. The Accenture difference is that we plug our unique capabilities and tools into integrated solutions powered by our partners. We use our deep industry and functional experience to provide an unbeatable offering that transforms our customers' businesses.
Narrow AI (Weak AI)
Artificial intelligence in which a program is written to accomplish a specific task. extremely powerful, something that can typically outperform a person but only in one specific narrow ability
How Does AI Work?
Artificial intelligence is machine learning powered by high quality data. 1. You start with an algorithm and feed it data. Lots and lots of data. 2. The algorithm analyzes the data and teaches an AI model what to do with it. 3. The AI model adapts, recognizing and analyzes any data it gets in the future based on this training. Three types of learning algorithms are used to train an AI model. 1. A supervised learning algorithm takes a label dataset, let's say cat images, and learns how to recognize a cat. Using what it learned about cats, it can spot one in other pictures. 2. An unsupervised learning algorithm takes an unlabeled dataset, say unlabeled images of cats and dogs, and sorts the images with similar characteristics into groups without knowing that one group is cats and the other is dogs. 3. A reinforcement learning algorithm works by trial and error, using a feedback loop of positive and negative reinforcement. It's like teaching your dog to do tricks. If your pet performs a trick successfully, he's rewarded with a treat. Unsuccessful tricks are not rewarded. The algorithm builds a set of successful tricks and another of unsuccessful tricks.
How Does AI Combine with Other Technologies?
Businesses can redesign the future by combining artificial intelligence with powerful technologies. • AI and Cloud. Cloud gives AI virtually unlimited processing power and the storage capacity to house the huge data sets needed to support machine learning. • AI and automation. Robotic Process Automation, or RPA, uses a computer to execute simple, repetitive tasks. AI enables intelligent RPA, allowing the computers to do more complex tasks and expand the range of use. • AI and the Internet of Things. IoT networks collect vast amounts of data as people use their devices. This is a great opportunity to incorporate AI and machine learning to analyze that data for insights. • AI and Blockchain. When used together, AI and Blockchain allow organizations to find significant amounts of trapped value by securely accessing shared tamperproof data. • AI and Analytics. Analytics works closely with AI to discover new, powerful business insights and previously unexamined data. Machine learning lets us analyze more data at higher levels to give us predictive insights at a speed, scale, and depth that was impossible before. Artificial intelligence is expanding insight exponentially.
neural networks (Artificial Neural Network)
DL algorithms. algorithms are actually modeling the brain a little. The idea that our brains have all these neurons and synaptic connections, neurons with multiple connections to other neurons. And in an artificial neural network, that idea is somewhat modeled. There are simulated neurons, individual nodes that could be connected to and send messages to multiple other simulated neurons. made of multiple layers. There's always an input layer and an output layer and typically at least one hidden layer in between them.
Automated Sentiment Analysis
In a situation where you have online reviews or comments, to have a process that could scan and identify whether they're positive or neutral or negative, even perhaps making judgments about emotion, Are any of these customers happy? Are any of these customers angry?, and deal with any major issues before they get out of hand
What is the purpose of Applied Intelligence for Accenture clients?
It gives our clients a roadmap to start with automation, building in the intelligence and the analytics for insight, and then layer on the artificial intelligence, again, driving the business outcomes.
Human + Machine Perspective
It's not about the person, the human, or the machine, or versus the machine, it's about how the two come together and create this new opportunity. smarter capabilities of AI, but we strongly believe that smarter technologies, smarter systems, when paired with people, make people more effective
Artificial General Intelligence (AGI), strong AI, full AI
Looks for a universal algorithm for learning and acting in any environment. generally intelligent, it's capable of many skills and capable of learning new skills by itself. Now just to be clear, this is still hypothetical. It's theoretical. We don't have AGI yet.
Why Does AI Matter?
The number one reason most businesses use artificial intelligence is to increase task efficiency, which can greatly reduce costs. four reasons why AI matters. 1. One, working smarter and more effectively. As AI tackles the lower‑level, repetitive work, people are freed up for higher‑level tasks that are more fulfilling and fuel business growth. 2. Two, improved accuracy and better decision making. Automation brings new levels of consistency, speed, and scalability to business production. 3. Three, better human experience through enhanced interaction. Technologies such as personalized chatbots create superior experiences for people through more personalized and efficient conversations. 4. Four, invention of new, intelligent products and services. AI supports agility and rapid experimentation, helping you discover new products and services with a speed and excellence that wasn't achievable before. Artificial intelligence, the future of business growth.
Insights: the overwhelmingly vast majority of companies do not need to write and never need to write machine learning algorithms
find a machine learning algorithm somebody else has already written, and we can feed our data into it. It's extremely common to use a cloud‑based hosted machine learning platform. combination of having these pre‑prepared machine learning frameworks ready to go, together with the computing power you need to run them in the cloud where you only have to pay for what you use, and it's reduced the barrier to entry
the black box of the machine learning.
one of the very real issues is that with many algorithms in machine learning, when we train a model, what you get from it is an answer. What you don't get is an explanation of that answer.
Expert Systems
rule-based systems that encode human knowledge in the form of if/then rules. combinations of software and even dedicated hardware
Machine Learning (ML)
subset of AI. the extraction of knowledge from data based on algorithms created from training data. we try and get to the results that we want not by providing a bunch of rules to follow, but by providing examples to learn from. Machine learning can only work if you have data to learn from and not just a little bit. You want a lot of data, and you want good quality data. most important part of this entire machine learning process is preparing the data, understanding what it is we're looking for, and then filtering that data, cleaning it, labeling it. If your data is full of garbage, full of invalid values and missing values and conflicting information, it doesn't matter how good the machine learning algorithms are
Deep Learning (DL)
subset of ML. uses algorithms that are more to do with the brain.
deep neural network (DNN)
A type of "machine learning" in artificial intelligence in which a computer is programmed to learn something (here object recognition). First the network is "trained" using input for which the answer is known ("that is a cow"). Subsequently, the network can provide answers from input that it has never seen before. it's more than one hidden layer
What Does AI Do?
Artificial intelligence enhances our lives personally and professionally by helping us be better at what we do and how we do it.
data monetization
How do we help clients monetize their data, new data services and such?
Chatbots (conversational interfaces)
Software that mimics written or spoken human speech to simulate a conversation or interaction with a real person; used commonly in customer service because people prefer chat over other methods
social determinants of health
The conditions in which people are born, grow, live, work, and age, shaped by the distribution of money, power, and resources at global, national, and local levels. are factors that you use to control health outside of the four walls of the hospital.
artificial intelligence
a subdiscipline of computer science that attempts to simulate human thinking. systems that approximate our human ability, you know, human‑like abilities of intelligence. is about systems that can sense, respond, comprehend, act, and then learn and improve as we go. So sense, comprehend, act, learn. And that combination of capabilities is really what defines artificial intelligence.
What Is AI?
artificial intelligence is the process of enabling machines to make human‑like decisions
machine learning algorithm (learner)
provide both positive and negative examples.
real‑world applied business‑focused artificial intelligence
there're two things we're usually not trying to do, model a brain and create a robot. many of the practical successes of recent years have taken completely different strategies. They're often more to do with statistics than biology
fireworks projects
where they've had specific data pipelines built for specific use cases for disease management, and they've had some success with it. But the success is not pervasive. It has not become a sustained innovation sort of a scenario for them where. They have not been able to make a tremendous impact thus far.
What is at Accenture's core is industry depth?
• You know, we know industries better than anybody in depth, • Every part of an industry and • A process and how it works. • We understand the function and capabilities. • We understand supply chain. • We understand customer experience in the areas Athina talked about. • And we understand change and how to drive big change and transformation for clients and how to also create experiences, develop the technology, and run the operations for clients.
