Tag: Oracle Policy Modeling

Back to Basics : Extensions #2

Back to Basics : Extensions #2

Following on from the previous post, we delve more deeply into the JavaScript Extensions world.

Interview Execution in a Browser

So how does an Interview Extension work? Let’s begin with some basic information about how your Oracle Policy Automation Interview runs in your Browser. If you happened to be viewing an Interview right now, and you were to open the Console (F12 in Google Chrome, or Microsoft Edge. Check your Browser documentation for the equivalent key or menu option), you might be able to view something like the following screenshot. Check the steps under the image as you may need to refer to them in your own case.

Extensions

In this screenshot I have launched a Project using the Debugger. Remember that if you hold down F5 while clicking the Debug button, you will open the Interview in the Debugger and in your default Browser.

  1. Your web server may of course be a different address.
  2. The web-determinations folder will not have the same numeric suffix as in this screenshot, indeed will most likely not have a suffix at all. This is a feature of the Debugging session.
  3. The js file is most likely in the staticresource folder, however if you are in a more integrated environment it may be in a different subfolder, or a different folder altogether. But it will be present.
  4. The contents of the js file can be read more easily by selecting (in Google Chrome in this case) the option to pretty print the code.

Interviews.js

This file is the foundation of the Interview experience provided by Oracle Policy Automation. It contains all the code necessary to make the user experience function correctly. Inside this file, however, there is a built-in capacity to accept extensions that change the behaviour of the Interview.

In your Console, search in the file for the following text – “customLabel:” (without the quotation marks, but with the colon). You should find one instance of that text, as shown in the screenshot below.

Extensions

  1. Search for the text
  2. Find the text in the file.

Take a moment to perform a second search in the same file, for the text shown below. Use the screenshot as your guide.

  1. Ensure you are looking at interviews.js
  2. Search for this text
  3. View the style definition for textInputStyle.

Accepted Extension Types

Notice in the first example, that customLabel is only one of a series of items in the first list. These are the recognized types of Control extension that we, as Oracle Policy Automation Project workers, are permitted to develop.

In the second search you found that there was a style defined for controls called textInputs. Although not quite as obvious perhaps as the first example, an Oracle Policy Automation Project might want to override the Style(s) used in a Project, in order to comply with corporate guidelines for example: and this system will help us do just that. Style Extensions use keywords in the same way to indicate which elements you wish to style.

About Extensions

It is not important at this stage to understand how these extensions are created or used. It is, however fundamentally important to understand that you will be extending Oracle Policy Automation Interviews by adding one or more of these acceptable extensions, and that they will be run in the browser in the same way as the standard JavaScript is already. These interviews can then be better adapted to your IT environment. As an example, read how Styling Extensions enable integration visually with Oracle Content & Experience Cloud.

More on this subject, with some worked examples, shortly.

JavaScript Extension : Geomapping with Leaflet JS

JavaScript Extension : Geomapping with Leaflet JS

Ever since the dawn of time (or at least since the arrival of JavaScript Extensions in Oracle Policy Automation), for whatever reason, the subject of Google Maps (geocoding, reverse geocoding, etc) has been one of the most read and downloaded on this site. But now, given that Google Maps (like Google APIs in general) requires registrations, license keys, and is subject to all sorts of restrictions  and throttling, developers have needed something else to impress their customers and colleagues with eye candy.

Geomapping with Leaflet JS and MapBox

So it’s appropriate therefore that this little post is all about Leaflet JS, “an open-source JavaScript library
for mobile-friendly interactive maps”. The Leaflet JS website is stuffed full of interesting examples and tutorials, so here goes for a short example of using this library in Oracle Policy Automation.

The setup is slightly different than working with Google Maps, but everything starts with the obvious step : download a library! In my case, I downloaded the library and CSS styles and dropped them into the resources folder. The files are tiny.

For the demonstration, my Oracle Policy Automation project contains only a handful of attributes, specifically enough to be able to create an address (street, town, zipcode, country) and a customer name (for a marker on the map).

The second element to handle in your code, is the image tile provider for your map. Working with Leaflet requires two things : Leaflet and a tile provider – in my case mapbox proved to be a good choice. It looks lovely. It requires a registration and access token, but it is very easy to obtain and you can try for free. Leaflet is entirely agnostic however, so if you want to use another tiling source, you can.

Apart from these steps, most of the code looks familiar : get the attribute values, pass them to the geocoding engine and then extract the latitude and longitude from the returned JSON object. Once you have that information,we use onreadystatechange to check that the page is loaded, and then draw the map, using latitude and longitude, and drop a custom marker at the location with a shadow and a popup window associated with the marker.

Because this is based on OpenStreetMap, we need to be good citizens and ensure we give credit where credit is due. Let’s look at a simple example:

/**
 * Richard Napier The OPA Hub Website June 2019
 * Educational Example of Custom Label with Leaflet JS and MapBox
 * I will remember this is for demonstration purposes only
 */
OraclePolicyAutomation.AddExtension({
	customLabel: function (control, interview) {

		if (control.getProperty("name") == "xMap") {
			return {
				mount: function (el) {

					// Add CSS First
					var head = document.getElementsByTagName('head')[0];
					var linkcss = document.createElement('link');
					linkcss.id = "leafletcss";
					linkcss.rel = 'stylesheet';
					linkcss.type = 'text/css';
					linkcss.href = '${resources-root}/leaflet.css';
					linkcss.setAttribute('crossorigin', "anonymous");
					head.appendChild(linkcss);
					var div = document.createElement("div");
					div.id = "lblmap";
					div.style.visibility = 'visible';
					div.style.height = '370px';
					div.style.width = '400px';
					el.appendChild(div);

					var street = interview.getValue("street");
					var city = interview.getValue("city");
					var zip = interview.getValue("zip");
					var country = interview.getValue("country");
					var customer = interview.getValue("customer");
					var latitude;
					var longitude;
					var uri = street + "," + city + "," + zip + "," + country;
					var res = encodeURI(uri);
					var root = "https://api.mapbox.com/geocoding/v5/mapbox.places/" + res + ".json?access_token=pXXX";

					$.ajax({
						url: root,
						method: "GET",
						async: false,
						dataType: 'json'
					}).then(function (data) {
						//console.log("Cleaning Up Response");
						//for (var key in data.features) {
						if (data.features[0]) {
							var val = data.features[0];
							longitude = val.center[0];
							latitude = val.center[1];
							//console.log(" Returned " + latitude + " " + longitude);
							$.getScript("${resources-root}/leaflet.js", function () {

								//console.log("Script loaded");
								window.onerror = function (errorMsg, url, lineNumber, column, errorObj) {
									console.log('Error: ' + errorMsg + ' Script: ' + url + ' Line: ' + lineNumber
										 + ' Column: ' + column + ' StackTrace: ' + errorObj);
								}

								//console.log(document.readyState);
								//console.log(latitude + " " + longitude);
								//	document.onreadystatechange  = function () {
								//console.log("ReadyState Change Event");
								if (document.readyState === "complete") {
									//console.log("ReadyState is Complete");
									var siteIcon = L.icon({
											iconUrl: '${resources-root}/images/Logo_Marker.png',
											shadowUrl: '${resources-root}/images/Logo_Marker_Shadow.png',
											iconSize: [30, 40],
											shadowSize: [30, 40],
											iconAnchor: [0, 40],
											shadowAnchor: [0, 38],
											popupAnchor: [15, -20]
										});
									var myLatLng = new L.LatLng(latitude, longitude);
									var mymap = L.map('lblmap').setView(myLatLng, 13);
									L.tileLayer('https://api.tiles.mapbox.com/v4/{id}/{z}/{x}/{y}.png?access_token={accessToken}', {
										attribution: 'Map data © OpenStreetMap contributors, CC-BY-SA, Imagery © Mapbox',
										maxZoom: 18,
										id: 'mapbox.streets',
										accessToken: 'pk.eyJ1IjoidGhlb3BhaHViIiwiYSI6ImNqdDJ3Y2liNDFvMjE0M3BpampzOG16cDYifQ.0ppcKabM2WhuaPT-kCGwYw'
									}).addTo(mymap);

									var marker = L.marker(new L.LatLng(latitude, longitude), {
											icon: siteIcon
										}).addTo(mymap);
									marker.bindPopup(customer + " is based here.");
								};

							

							});
						}
					

					});

					// Assumes JS is loaded

				},
				unmount: function (el) {
					if (control.getProperty("name") == "xMap") {
						
						 var myMap = document.getElementById("lblmap");
                        myMap.parentNode.removeChild(myMap);


					}

				}

			}

		}

	}
})

Lines 0-40 : Pretty standard stuff, setting up the CSS file and the encoded URL for the REST call which will get us the latitude and longitude.
Lines 41- 52 : Used to call the service and get / parse the response, giving us the latitude and longitude
Lines 53 – 64 : We need to be sure that we wait until the leaflet JavaScript file is loaded before doing anything else
Lines 67 – 75 : Set up a Custom Marker
Lines 77 – 83 : Add a Map Layer and Credits
Lines 84 – 87 : Add a Marker using a custom image and shadow
Lines 85 – 88 : Add a Popup to the Marker click, and use some Oracle Policy Automation data in the popup.

The rest is just clean-up and go home (the unmount).

Geomapping with Leaflet JS : The Result

The end result is a nicely presented map, with a marker at the correct address, and a cute popup window when the marker is clicked.

Geomapping with Leaflet JS MapBox Example

We’ll see what else we can do with this nice alternative library in the coming weeks. In the meantime, you will find the sample Project in the OPA Hub Shop.

Back to OPA Basics : Oracle Policy Modeling Features

Back to OPA Basics : Oracle Policy Modeling Features

Welcome to another in our periodic back to OPA basics series. At the moment I am watching a lot of new starters join a set of experienced developers. And funnily enough, both groups sometimes are stuck in their routine. For the new people, they fall back on what they know from other rules engines. For the more experienced people, that are familiar with Oracle Policy Modeling, they repeat what they learned long ago and do not necessarily see anything that has been added to the application in the intervening time since they first started using it.

So here are my top five cool time savers and useful things you can find in Oracle Policy Modeling today.

  • The Rule Assistant

I still find it strange that many people don’t use the Assistant in Word. If there was ever a tool to avoid having to remember the arcane spelling and phraseology of an Oracle Policy Automation attribute, this is it!

Back to OPA Basics - Rule Assistant

  • The Convert to Test Case Export

Watching people filling in every cell in their Test Case spreadsheet, when they probably have already saved many of the initial scenarios as debug sessions, I think they should remember the fact that they can move a unit test into a Test Case, and vice-versa.

Back to OPA Basics - Export as Test Case

  • The Find Unused Attribute Filter

I tend to use this one when the Rule Assistant has not been used very much, so we are looking for duplicate / mispelled attributes in our Project (see Rule Assistant, above).

Back to OPA Basics - Unused Attributes

  • The Inclusion Report

OK, so I’m probably cheating as far as this one is concerned, but it is a real time saver. In 19A, the introduction of the Inclusion Report has saved me time already. Great for beginning an impact assessment when some sort of surgery is required on Project structure. Find out more here.

Back to OPA Basics - Inclusion Report

  • The Export Entity Data Model Option

This is a tiny little option hidden away in the toolbar but I’m often asked by non-OPA people for the data model and I find this export really simple and quick to use.

Tiny Button - Export Data Model

Well, that’s our top five for now. What other tricks do you use to get the most out of Oracle Policy Modeling?

Let us know in the comments!

 

Business Process : Force PDF Download in OPA

Business Process : Force PDF Download in OPA

After last week’s business case put forward about Calendar Blackout dates, this week we can thank Arnaud D and Sylvie A from Switzerland for their excellent use case. This week we will be discussing the idea of letting a user progress to the end of an Interview, only if they have clicked to download a PDF file. So, can we Force PDF Download in OPA?

There is not really a technically feasible way to prove that a user actually read a PDF. But if we can force them to download it, at least we know they have gotten that far!

The business case is easily applied to many industries, for example where terms and conditions or a contract are generated and tacit approval needed before the user continues. Or, more simply, you need some way to encourage people to read the BI Publisher thing you spent all those weeks developing.

It is also a good opportunity for us to investigate the new custom Button extensions available in Oracle Policy Automation 19A. So this won’t work in 18D or earlier. The approach we are going to take will involve two different extensions. Let’s go!

Input Extension

The input extension is needed because we wish to capture the value of a Boolean attribute for our demonstration. Navigating forwards will only be allowed if this Boolean is true.

/*  Generated by the OPA Hub Website 29/03/2019 19:35
Educational Example of Custom Input Extension for Oracle Policy Automation
I will remember this is for demonstration purposes only.
 */
OraclePolicyAutomation.AddExtension({
	fullCustomInput: function (control, interview) {
		if (control.getProperty("name") == "xInput") {
			return {
				mount: function (el) {
					console.log("Starting customInput Mount");
					var div = document.createElement("input");
					div.id = "myInput";
					div.value = interview.getValue("my_flag");
div.setAttribute("style", "visibility: hidden");
el.appendChild(div);
console.log("Ending customInput Mount");
$(".opa-document").click(function(){
console.log("Clicked Link");
control.setValue(true);
});
},
update: function (el) {
console.log("Starting customInput Update");

console.log("Ending customInput Update");
},
validate: function (el) {
console.log("Starting customInput Validate");

console.log("Ending customInput Validate");
},
unmount: function (el) {
if (control.getProperty("name") == "xInput") {
console.log("Starting customInput UnMount");

interview.setInputValue("my_flag", $("#myInput").val());
var myInput = document.getElementById("myInput");
myInput.parentNode.removeChild(myInput);
console.log("Ending customInput UnMount");
}
}
}
}
}
})

So what’s going on here? Basically we are putting a click event on a PDF download. Notice line 17 : this adds a click event.  So as soon as the custom Input is loaded, an event is added to another control using jQuery. When the link is clicked, we update our flag to true. We use the style selector “.opa-document” to find the link. If you have more than one link, you will need a more precise method of finding it. Other than that, the custom input itself is invisible, and simply reads the current value of the flag into the control.

Button Extension

The other extension, for a nextButton, uses jQuery to display a dialog when you try and leave the screen before clicking on the link (the flag is false). Otherwise it lets you navigate without issues.

OraclePolicyAutomation.AddExtension({
	customNextButton: function (interview) {
		let button,
		caption;

		return {
			mount: function (el) {
				const screen = interview.currentScreen();

				button = document.createElement("div");
				button.setAttribute("title", "Print Your PDF First");
				button.setAttribute("class", "opa-back");
				button.setAttribute("role", "button");
				button.setAttribute("style", "	text-overflow: ellipsis; white-space: nowrap; line-height: 1; cursor: pointer; padding: 0.7em 1.2em; border: none; text-decoration: none; -webkit-appearance: none; flex: 0 0 auto; text-align: center; color: rgb(255, 255, 255); border-radius: 6px; background: rgb(20, 116, 191); font-size: 16px; font-style: normal; font-weight: normal; margin-left: 6px; visibility: visible; outline: none;");

				caption = document.createTextNode(screen.getNextCaption());
				button.appendChild(caption);

				var dialog = document.createElement("div");
				dialog.setAttribute("id", "dialog");
				dialog.setAttribute("title", "Warning");

				var dialogtext = "Please download your document first.";
				el.appendChild(dialog);


				button.onclick = function (evt) {
					evt.preventDefault();
					if (interview.getValue("my_flag") == true) {
						interview.submit();
					} else {
						$("#dialog").dialog();
						$("#dialog").text(dialogtext);
						console.log("Not Yet ");
						console.log("Flag Value is " + $("#myInput").val());
					}
				}

				el.appendChild(button);
			},
			update: function (el) {
				const screen = interview.currentScreen();

				caption.textContent = screen.getNextCaption();

				if (screen.hasNextButton())
					button.removeAttribute("disabled");
				else
					button.setAttribute("disabled", "disabled");
			},
			unmount: function (el) {}
		}
	}
});

So what’s happening? The early part of the mount is simply painting a nice button on the screen. We could of course not display the button at all, but the User Experience might be better if we display it always, just prohibit using it in certain circumstances. Starting around line 19 we create the dialog and make it ready for use.

Line 27 includes the button click routine which either displays the dialog, complete with a message, or just lets you navigate as normal.

The other lines in the unmount are just taken from the example in the online help, where if this is the last Screen, no next button is available. But our business case is based on it not being the last.

Force PDF Download : Animated GIF

In the animated GIF above, you can see what it looks like.

The Project Zip is in the shop. Have fun!

Oracle Policy Automation May 2017 – New Features #4

Oracle Policy Automation May 2017 – New Features #4

Oracle Policy Automation May 2017 – New Features #4

Following on from the previous post, in this third video about the Oracle Policy Automation May 2017 – New Features, we look at a functionality that is actually making a comeback. In Oracle Policy Modeling version 10, it was possible to hide or show values in lists (for example, a drop down control on an Interview Screen) based on a Boolean attribute value. This was shelved in version 12 but now it makes a welcome return in May 2017.

The functionality is easy to implement, and simply requires a Value List, a Drop Down control or similar, and some logic to power the decision whether to display a value or not. The display of each value, including uncertain can be powered by the same or a different Boolean attribute.

The functionality extends the Interview experience by allowing for a better user experience and an easier manner than jQuery or some such client-side smoke and mirrors to implement a state model or logical transition model within the Oracle Policy Modeling environment. The video that follows this article walks through a simple example and hopefully shows you the way. Catch up with the other tutorials and videos in this series in the links below

Whats New May 2017 Part One

Whats New May 2017 Part Two

Whats New May 2017 Part Three

The Walkthrough Video

Oracle Policy Automation OPA 12 Browser Detection

Oracle Policy Automation OPA 12 Browser Detection

Oracle Policy Automation OPA 12 Browser Detection

Oracle Policy Automation OPA 12 Browser DetectionI wanted to go back over a situation I had a couple of weeks ago, in respect of needing to detect the user agent of the Browser that had an OPA 12 IFRAME in it, for reasons I will describe in a moment. As many of you know, the latest versions of Oracle Policy Modeling are moving away from the concept of bundling JavaScript files and other externals (such as CSS Files) with the Project by means of the /interview-theme/resources folder. In fact in the latest versions in “latest version” mode there is no support for JavaScript or CSS added by this mechanism, nor by Custom Control. So “August 2016” mode to the rescue when it comes to OPA 12 Browser Detection.

We could either just drop the JavaScript and CAA in the resources folder above, or in the case I met this week, this was not an option, it had to be a PHP file thus it was a Custom Control.

Putting aside the questions like “Where is OPA going with support for JavaScript?” (which would seem to point in the direction of RuleScript) it brings to the fore the need to achieve certain things – like browser user-agent detection – within the now somewhat less flexible framework. Let’s look at what I am talking about.

Internet Explorer 11 and the IFRAME : OPA 12 Browser Detection

Internet Explorer 11, especially in Windows 10 and on Surface tablets, behaves differently from the majority of browsers (why is this not a surprise?) in respect of opening PDF files from within an IFRAME. The PDF file is not displayed, an error is logged in the console and (in the case of Oracle Policy Automation) the session is lost.

So one obvious approach would be to prepare two links, one to download the PDF and another to open the Oracle Policy Automation Form in, for example, HTML format since that is not subject to the same security block. And then, depending on the browser user agent, hide one of the links so for the end user it is totally transparent:

  • For IE11, display the HTML version of the OPA Form
  • For all other browsers display the PDF link to the OPA Form

The advantage of this approach is improved maintenance – there is only one form, just two delivery mechanisms. Now we need to show or hide the relevant link depending on the browser.

Getting the User Agent : OPA 12 Browser Detection

Now we are faced with the User Agent issue. As you probably know, Custom Controls in PHP are still supported or at least functional as long as it is an Oracle Policy Modeling Control of type Label and your Project is set to “August 2016 mode” . So this is how we might proceed. If we use a Custom Control in PHP, the $_SERVER[‘HTTP_USER_AGENT’] variable will contain the User Agent ready for use in PHP. But it will probably return a value that is not that serviceable or even correct. But the advantages of being able to access the $_POST array (with all of the interesting metadata and things in it) mean that sometimes PHP is the way you need to go.

PHP Version : OPA 12 Browser Detection

So, JavaScript via PHP to the rescue (and yes, I know there are other ways of doing this, but I just want to illustrate a problem and point you in one direction for a solution, even if it perhaps sits right on the edge of server / client code separation). Here is a sample PHP snippet which you can attribute to a Label Custom Control. Host the PHP on your web server or wherever you have PHP.

Notice the “Trident” match : as a simple demonstration I used the text string “Trident” to detect IE11. It is not meant to be a completely bulletproof detection just an example to be replace which whatever string you are looking for. For reference I include the HTTP User Agent reported by PHP versus the JavaScript navigator.userAgent.

<?php
echo '<div style="display : none;" name="ode_browser_check" id="ode_browser_check" class="ode_submit" >' . $_SERVER['HTTP_USER_AGENT'] . '</div>';
echo "<script type=\"text/javascript\">
var myData = navigator.userAgent;
if(myData.match(/Trident/i) )
{
 (function($) {
 $(document).ready(function(){
 $(\"#id_tohide\").remove();
 $(\"#id_toshow\").show();
 $(\"#id_ofcontainerdiv > a\").attr(\"target\", \"_blank\");
 });
 });
})(jQuery);
} else {
 (function($) {
 $(document).ready(function(){
 $(\"#id_tohide\").remove();
 $(\"#id_toshow\").show();
 });
 });
})(jQuery);
};";
echo '</script>';
?>

The example above assumes you have two HTML IDs, one for the link to hide, one to show. If you are using the standard OPA Form link, you will need to add the “_blank” target to the link that points to the HTML output version, as in the example otherwise IE11 will open it in the same window and you lose your session. With these tags and code in place, your IE11 users will see one link, the other browsers will see another link.

JavaScript Version : OPA 12 Browser Detection

If you are in a position where a JavaScript file is all that is needed then you can do pretty much the same thing of course just by reverting to normal JavaScript syntax.

if(navigator.userAgent.match(/Trident/i) )
{
 (function($) {
 $(document).ready(function(){
$(\"#id_tohide\").remove();
 $(\"#id_toshow\").show();
 $(\"#id_ofcontainerdiv > a\").attr(\"target\", \"_blank\");
 });
})(jQuery);
} else {
 (function($) {
$(\"#id_tohide\").remove();
$(\"#id_toshow\").show();
 });
})(jQuery);
}

Note : All of the above will, by definition,  not work in “latest version” mode since Custom Controls / JavaScript  are not supported in “latest version” mode.

Note : In light of all of this, it makes me wonder how much longer Oracle Policy Automation deployment on a website  or on Oracle Service Cloud Customer Portal or Agent Desktop, can depend on such a nasty, legacy HTML tag as the IFRAME.

Have fun.

How Oracle Policy Automation (OPA) Helps In Enterprise AI

Enterprise AI – an Intelligent Data Warehouse/ERP system?

AI enables computers to do some things better than humans especially when it comes to finding insights from large amounts of Unstructured or semi-structured data. Technologies like Machine learning , Natural language processing (NLP) , Speech recognition, and computer vision drive the AI layer. More specifically, AI applies to an algorithm which is learning on its own.

To understand this, we have to ask ourselves: How do we train a Big Data algorithm?

There are two ways:

  • Start with the Rules and apply them to Data (Top down) OR
  • Start with the data and find the rules from the Data (Bottom up)     

The above Rules can be implemented / Executed using Oracle Policy Modeling and Oracle Policy Automation (OPA)

The Top-down approach involved writing enough rules for all possible circumstances. But this approach is obviously limited by the number of rules and by its finite rules base. The Bottom-up approach applies for two cases. Firstly, when rules can be derived from instances of positive and negative examples(SPAM /NO SPAN). This is traditional machine learning when the Algorithm can be trained. But, the more extreme case is : Where there are no examples to train the algorithm.

What do we mean by ‘no examples’?

a)   There is no schema

b)   Linearity(sequence) and hierarchy is not known

c)   The output is not known(non-deterministic)

d)  Problem domain is not finite

Hence, this is not an easy problem to solve. However, there is a payoff in the enterprise if AI algorithms can be created to learn and self-train manual, repetitive tasks – especially when the tasks involve both structured and unstructured data.

Oracle Policy Automation - AIHow can we visualize the AI layer?

One simple way is to think of it as an ‘Intelligent Data warehouse’ i.e. an extension to either the Data warehouse or the ERP system

For instance, an organization would transcribe call centre agents’ interactions with customers create a more intelligent workflow, bot etc using Deep learning algorithms.

Enterprise AI layer – What it mean to the Enterprise

So, if we imagine such a conceptual AI layer for the enterprise, what does it mean in terms of new services that can be offered? Here are some examples

Bots : Bots are a great example of the use of AI to automate repetitive tasks like scheduling meetings. Bots are often the starting point of engagement for AI especially in Retail and Financial services

Inferring from textual/voice narrative: Security applications to detect suspicious behaviour, Algorithms that can draw connections between how patients describe their symptoms etc

Detecting patterns from vast amounts of data: Using log files to predict future failures, predicting cyberseurity attacks etc

Creating a knowledge base from large datasets: for example an AI program that can read all of Wikipedia or Github.

Creating content on scale: Using Robots to replace Writers or even to compose Pop songs

Predicting future workflows: Using existing patterns to predict future workflows

Mass personalization: in advertising

Video and image analytics: Collision Avoidance for Drones, Autonomous vehicles, Agricultural Crop Health Analysis etc

These applications provide competitive advantage, Differentiation, Customer loyalty and mass personalization. They have simple business models (such as deployed as premium features /new products /cost reduction )

The Enterprise AI layer and IoT

So, the final question is: What does the Enterprise layer mean for IoT?

IoT has tremendous potential but faces an inherent problem. Currently, IoT is implemented in verticals/ silos and these silos do not talk to each other. To realize the full potential of IoT, an over-arching layer above individual verticals could ‘connect the dots’. Coming from the Telco industry, these ideas are not new i.e. the winners of the mobile/Telco ecosystem were iPhone and Android – which succeeded in doing exactly that.

Firstly, the AI layer could help in deriving actionable insights from billions of data points which come from IoT devices across verticals. This is the obvious benefit as IoT data from various verticals can act as an input to the AI layer. Deep learning algorithms play an important role in IoT analytics because Machine data is sparse and / or has a temporal element to it. Devices may behave differently at different conditions. Hence, capturing all scenarios for data pre-processing/training stage of an algorithm is difficult. Deep learning algorithms can help to mitigate these risks by enabling algorithms to learn on their own. This concept of machines learning on their own can be extended to ‘machines teaching other machines’. This idea is not so far-fetched and is already happening, A Fanuc robot teaches itself to perform a task overnight by observation and through reinforcement learning. Fanuc’s robot uses reinforcement learning to train itself. After eight hours or so it gets to 90 percent accuracy or above, which is almost the same as if an expert were to program it. The process can be accelerated if several robots work in parallel and then share what they have learned. This form of distributed learning is called cloud robotics

We can extend the idea of ‘machines teaching other machines’ more generically within the Enterprise. Any entity in an enterprise can train other ‘peer’ entities in the Enterprise. That could be buildings learning from other buildings – or planes or oil rigs. We see early examples of this approach in Salesforce.com and Einstein. Longer term, Reinforcement learning is the key technology that drives IoT and AI layer for the Enterprise – but initially any technologies that implement self learning algorithms would help for this task

Conclusion

In this brief article, we proposed a logical concept called the AI layer for the Enterprise. We could see such a layer as an extension to the Data Warehouse or the ERP system. This has tangible and practical benefits for the Enterprise with a clear business model. The AI layer could also incorporate the IoT datasets and unite the disparate ecosystem. This will not be easy. But it is worth it because the payoffs for creating such an AI layer around the Enterprise are huge! The Enterprise AI layer theme is a key part of the Data Science for Internet of Things.

Note : this article was reprinted in The OPA Hub Website with the kind permission of the author. You can find him on LinkedIn and the original article can be found here.

Worldwide
Logo by Southpaw Projects LLC