{"id":8967,"date":"2020-07-20T08:37:01","date_gmt":"2020-07-20T08:37:01","guid":{"rendered":"https:\/\/www.experfy.com\/blog\/?p=8967"},"modified":"2023-11-28T07:25:41","modified_gmt":"2023-11-28T07:25:41","slug":"how-to-create-an-ai-artificial-intelligence-model","status":"publish","type":"post","link":"https:\/\/www.experfy.com\/blog\/ai-ml\/how-to-create-an-ai-artificial-intelligence-model\/","title":{"rendered":"How To Create An AI (Artificial Intelligence) Model"},"content":{"rendered":"\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Lemonade is one of this year\u2019s hottest IPOs and a key reason for this is the company\u2019s heavy investments in AI (Artificial Intelligence).\u00a0The company has used this technology to develop bots to handle the purchase of policies and the managing of claims.\u00a0<\/p>\n\n\n\n

Then how does a company like this create AI models?\u00a0What is the process?\u00a0Well, as should be no surprise, it is complex and susceptible to failure.\u00a0<\/p>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

But then again, there are some key principles to keep in mind.\u00a0So let\u2019s take a look:<\/p>\n\n\n\n

Selection<\/strong>:\u00a0There are hundreds of algorithms to choose from.\u00a0In some cases, the best approach is to use several (this is known as ensemble modelling).\u00a0<\/p>\n\n\n\n

\u201cSelecting the right model starts with gaining a thorough understanding of what the organization wishes to achieve,\u201d said Shadi Sifain, who is the senior manager of data science and predictive analytics at\u00a0Paychex<\/a>. \u201cSelecting the right model often also involves balancing a number of requirements including model performance, accuracy, interpretability, and compute power among other factors,”<\/p>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

It\u2019s important to realize that you need the right kind of data for certain models.\u00a0If anything, this is one of the biggest challenges in the AI development process.\u00a0\u201cOn average, the data preparation process takes 2X or in some cases 3X longer that just the design of the machine learning algorithm,\u201d said Valeria Sadovykh, who is the Emerging Technology Global Delivery Lead at\u00a0PwC Labs<\/a>.<\/p>\n\n\n\n

So in the early phases of a project, you need to get a good sense of the data.\u00a0\u201cConduct an exploratory analysis,\u201d said Dan Simion, who is the VP of AI & Analytics at\u00a0Capgemini North America<\/a>.\u00a0\u201cVisualize the data in 2-dimensions and 3-dimensions, then run simple, descriptive statistics to understand the data more effectively.\u00a0Next, check for anomalies and missing data. Then clean the data to get a better picture of the sample size.\u201d<\/p>\n\n\n\n

But there is no perfect model, as there will always be trade-offs.<\/p>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

\u201cThere is an old theorem in the machine learning and pattern recognition community called the No Free Lunch Theorem, which states that there is no single model that is best on all tasks,\u201d said Dr. Jason Corso, who is a Professor of Electrical Engineering and Computer Science at the University of Michigan and the co-founder and CEO of\u00a0Voxel51<\/a>.\u00a0\u201cSo, understanding the relationships between the assumptions a model makes and the assumptions a task makes is key.\u201d<\/p>\n\n\n\n

Training<\/strong>:\u00a0Once you have an algorithm \u2013 or a set of them \u2013 you want to perform tests against the dataset.\u00a0The best practice is to divide the dataset into at least two parts.\u00a0About 70% to 80% is for testing and tuning of the model.\u00a0The remaining will then be used for validation.\u00a0Through this process, there will be a look at the accuracy rates.<\/p>\n\n\n\n

The good news is that there are many AI platforms that can help streamline the process.\u00a0There are open source offerings, such as TensorFlow, PyTorch, KNIME, Anaconda and Keras, as well as proprietary applications like Alteryx, Databricks, DataRobot, MathWorks and SAS.\u00a0And of course, there are rich AI systems from Amazon, Microsoft and Google.\u00a0<\/p>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\n\n

\u201cThe key is to look for open source tools which allow for easy and quick experimentation,\u201d said Monica Livingston, who is the Director of AI Sales at\u00a0Intel<\/a>.\u00a0\u201cIf you prefer to purchase 3rd party solutions, there are many ISVs offering AI-based solutions for tasks like image recognition, chat bots, defect detection and so on.\u201d<\/p>\n\n\n\n

Feature Engineering<\/strong>:\u00a0This is the process of finding the variables that are the best predictors for a model.\u00a0This is where the expertise of a data scientist is essential.\u00a0But there is also often a need to have domain experts help out.\u00a0<\/p>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\n

\u201cTo perform feature engineering, the practitioner building the model is required to have a good understanding of the problem at hand\u2014such as having a preconceived notion of possible effective predictors even before discovering them through the data,\u201d said Jason Cottrell, who is the CEO of\u00a0Myplanet<\/a>. \u201cFor example, in the case of predicting defaults for loan applicants, an effective predictor could be monthly income flow from the applicant.\u201d<\/p>\n\n\n\n

But finding the right features can be nearly impossible in some situations.\u00a0This could be the case with computer vision, such as when used with autonomous vehicles.\u00a0Yet using sophisticated deep learning can be a solution.\u00a0<\/p>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

\u201cThese days, neural networks are used to learn features, as they are better at understanding statistics than humans,\u201d said Eric Yeh, who is a computer scientist at the Artificial Intelligence Center at\u00a0SRI International<\/a>. \u201cHowever, they are not necessarily a panacea and might develop features that were not intended as well. The famous example is the image classifier which was developed to detect tanks and jeeps. Instead, it learned to detect night and day since all jeep photos were taken in the day and all tank photos were taken in the museum at night.\u201d<\/p>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"

How to create AI models? What is the process? Well, as should be no surprise, it is complex and susceptible to failure. But then again, there are some key principles to keep in mind. So let\u2019s take a look.<\/p>\n","protected":false},"author":667,"featured_media":8968,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"content-type":"","footnotes":""},"categories":[183],"tags":[476,477,478],"ppma_author":[3440],"class_list":["post-8967","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-ml","tag-ai-model","tag-create","tag-key-principles"],"authors":[{"term_id":3440,"user_id":667,"is_guest":0,"slug":"tom-taulli","display_name":"Tom Taulli","avatar_url":"https:\/\/www.experfy.com\/blog\/wp-content\/uploads\/2020\/04\/medium_63f5f357-f6ae-4999-b3f4-ed30406eb86c-150x150.jpg","user_url":"http:\/\/www.taulli.com\/","last_name":"Taulli","first_name":"Tom","job_title":"","description":"Tom Taulli is an author, speaker, and advisor \u2013 with a focus primarily on technology. He has co-founded a variety of\u00a0companies, including Hypermart.net (sold to InfoSpace), WebIPO and BizEquity. He is also a contributor to Forbes.com, Bloomberg.com, Kiplinger and BusinessWeek. \u00a0He has also written a variety of books, mostly on tech and finance.\u00a0 His latest is\u00a0Artificial Intelligence Basics: A Non-Technical Introduction<\/a>. \u00a0As of now, he is an advisor to some awesome startups."}],"_links":{"self":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts\/8967","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/users\/667"}],"replies":[{"embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/comments?post=8967"}],"version-history":[{"count":4,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts\/8967\/revisions"}],"predecessor-version":[{"id":34420,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts\/8967\/revisions\/34420"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/media\/8968"}],"wp:attachment":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/media?parent=8967"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/categories?post=8967"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/tags?post=8967"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/ppma_author?post=8967"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}