Introduction to Subsets of Artificial Intelligence
Artificial Intelligence as the name suggest is artificially created intelligence for machines so as to make them smart and do the tasks which humans can do. We know, computers are quicker when we talk about analytical abilities, but it does not have ability to make decisions on their own. The job of making computers to take decision is Artificial Intelligence. Artificial Intelligence consists of many sets whose common goal is to make decision based on data and provide output. We will discuss some important Subsets of artificial intelligence like Machine learning, Deep learning, Natural language programming and Virtual assistance like chatbots.
Top 4 Subsets of Artificial Intelligence
We will discuss all the subsets one by one in this section:
1. Machine Learning
“Machine learning is a Field of study where computer learn from available data/historical data without being explicitly programmed”
In Machine learning, we won’t program the code explicitly for each type of problem i.e. machine will try to figure out how to approach for the problem. In machine learning the approach we are going to use is different, instead of giving direct instruction a special algorithm is used which will recognize the pattern and based off of those patterns, it will predict best possible output.
Machine learning can further be categorized into 3 categories:
- Supervised Learning: Here input and output both labels are known. The goal is to train the model on this i/p and o/p data and predict the outputs for the input labels.
- Unsupervised Learning: Here output label is not known, and goal is to make sense of data available either through clustering or association.
- Reinforcement Learning: Reinforcement learning is a feedback reward algorithm and has an agent which is trained on some command and performs action. For each action the agent gets reward feedback and keeps improving its performance.
2. Deep Learning
As we know Machine learning is subset of Artificial Intelligence, Deep learning is a subset of machine learning. The only difference in deep learning model is that with experience model becomes better without any specific guidance. In deep learning model is awarded and penalised based on feedback and then it will adjust weights accordingly for the input parameters.
Deep learning requires large amount of data to build models, Deep learning models are based on neural networks. Neural networks used in deep learning algorithms are replicas of neurons of human brain. In human brain neuron relate to each other forming a deep neural network, in deep learning also the artificial neurons are connected with each other to form a deep neural network and that is why this type of learning is called as deep learning. In AI the replica of human neuron is perceptron, which are connected together to form deep neural networks. A perceptron has input nodes (dendrites in human brain), an actuation function to make a small decision and output nodes (axon in human brain).
Deep neural network consists of 3 sections:
- Input Layer: The first layer in the network and takes raw input, and after processing propagates it to next layer of neurons.
- Hidden Layers: These layers are intermediate and depending on the complexity of the problem the number of layers varies from one to hundreds. The information passed from input layer is processed in each layer of neurons and passed to the next.
- Output Layer: This is the last layer in the neuron which gives output to end user.
3. Natural Language Processing
- Natural language is the language which is spoken by human in general. Natural Language Programming can be treated as a subfield of linguistics concerned with interaction between computer machine and human. In NLP we are trying to make computers understand like a human.
- There are two approaches to do NLP, one is Rule Based NLP and other is Statistical NLP. In Rule based the rules are defined based on grammar rules of a language. In statistical NLP machine is fed with a large amount of data called Corpora of real-world communication examples. The algorithms then try to learn from the data and model is build which can understand any instruction based on previous experience and make sense of it.
- Statistical NLP is better, because humans do not stick to grammar rules in general communication, and also way of speaking also differs from person to person. In rule-based NLP it becomes very complex to define rules for all and some rules contradicts others which makes it even more difficult.
4. Virtual Assistants / Chat Bots
- Now a days, we are using a number of online tools, applications and websites for daily purpose for eg, banking, online shopping, online courses. Now if we face any issue, we contact to support team of that online tool, application or website. Support person from the other end will provide us the solution. Here many a times the situation is that a number of customers have similar queries.
- To attend to those queries, now we have chat bots which are trained on previous queries from the customers, for what query what could be the best possible solution. We see such chatbots now everywhere for eg, in Amazon Website we see options in help such as the query is related to which order and regarding what delivery or coupon or payment and upon giving all information, we get automatic replies generated by chat bot.
- Moreover, we have Virtual assistant such as Alexa, Siri and Google home available to us, which can follow our command precisely and provide as the related information. You can tell them to remind any event or call someone or play your favourite music and they will follow the same.
Artificial Intelligence as the name suggest is artificially created intelligence for machines so as to make them smart and do the tasks which humans can do. AI has numerous subsets, which consists of mathematics, science and so on. We have discussed some important subsets like machine learning, deep learning, Natural language processing and virtual assistants.
This is a guide to Subsets of Artificial Intelligence. Here we discuss an introduction and top 4 Subsets in Artificial Intelligence. You can also go to our other related articles –