Tag: Gartner

Are Your New Remote Workers Visible to Security Operations?


Due to coronavirus, companies are seeing an unprecedented amount of remote work. Whether mandated by the government or the organization, businesses are asking many (if not all) employees to work from home. While this move creates obvious challenges for IT in terms of infrastructure and capacity, it’s also creating challenges for security teams as they push to scale remote work on a rapid and global level.

Many are utilizing remote working systems that have not been operationally tested as part of their core security operations monitoring. For many, the likely result is fewer security alerts and issues because the corporate infrastructure will not be subject to the same levels of usage in areas such as internet browsing, and users may be working from web-based applications on non-company-sanctioned assets.


Don’t stop asking questions (even if it seems like you should)

For security operations center (SOC) analysts, this may seem like a heavenly situation: Fewer false positives to deal with, lower likelihood of security policies being broken and more time to deal with all those things they want to concentrate on but never had the time for before.

But fewer alerts does not equate with being more secure; rather it might mean you are more blinded by the lack of visibility. Lack of visibility does not equate to a lack of security vulnerabilities. Security leaders must consider these new risks to our organizations:

  • Will we even know that data and systems are being compromised?
  • Are we now dependent on a wide range of key remote working solutions that don’t have proper resilience?
  • Once all this is over, will we know where all our sensitive data resides?
  • Are we still compliant with the IT security regulations that we need to be?

Gather your security steering committee 

It’s time to gather your team (assuming you haven’t already done so) around the (possibly virtual) table. Much like the government response to the socioeconomic challenges at the moment, we have to manage these challenges one day at a time.

Where the government will have top medical experts, your table should have not just management, but a variety of security skill sets. The goal should be to help security operations prioritize a strategic response to this potential crisis so that we are not negligent. We don’t want to adopt an attitude where we just let security issues happen. Once gathered, answer a couple of the key questions:

  1. Are we still looking in the right direction? Are the use cases, security data sources, endpoint agents, etc., all focused on the areas that will keep our business in business, and are there any massive gaps?
  2. Do we have a plan to revert to normal working when all this is over? How are we recording where our data is going, and how do we make sure it remains secure?
  3. Are we still running the right security operations model?

The solution is not technology

The solution to these issues lies in solid processes, not technology. The goal is twofold: Establish a set of priorities for the security operations team and focus on an adjusted set of business risks. But don’t neglect to establish a path to return to normal in a nondisruptive way when the time comes.

Read more: 4 Actions to Be a Strong Leader During COVID-19 Disruption 

As business requirements shift and flex in the current environment, security use cases will require reevaluation. First, we need to account for any new data sources and new ways of working. Second, think about the protection of new key business enablers (such as remote working platforms or VPNs).

Part of this process will be to meticulously document all changes so they can be reversed at a later date, understanding and recording where everything that creates new risk now resides. Even if these are strategic changes, evaluate carefully, as most are probably tactical at this point. You’ll need to reevaluate the path taken to ensure it is robust. 

The security part of our businesses need to move quickly on this. It’s not just a question of “Can our security operations and SOC analysts work remotely?” but also “What new risk does this bring?” and “Have our security priorities changed?”

Whether these changes are purely for our internal teams or whether we have to engage our security service providers about moving faster to change based on new requirements, it’s clear that organizations need to complete a due diligence exercise to make sure that what they are doing to protect the organization matches the objectives set to keep cyberrisk low. Adjusting in alignment with what is high priority and what is feasible is an agile change.

The post Are Your New Remote Workers Visible to Security Operations? appeared first on Smarter With Gartner.

Vía IT – Smarter With Gartner https://ift.tt/2UW5cYm

3 Keys to Leading Large-Scale Virtual Meetings


In today’s geographically dispersed workplace, leaders no longer present to employees in the same room at the same time for town halls, all hands and all-company meetings. Yet these large-scale events are essential for executive leaders to share their vision and strategy, address urgent issues and celebrate wins. The solution? Live video.

Companies vary widely in their ability to produce and use high-quality, live video assets in their communications. “There’s a spectrum of engagement,” says Adam Preset, Senior Director Analyst, Gartner. “On one end, you have leaders and teams eager to deliver quality live video events and workers who are able to join and interact in real time. On the other, you have traditional organizations that are still using massive audio conference bridges to keep people informed at scale.”

The middle of a crisis or other period when people really need to hear from leaders is not the ideal time to figure out a new technology solution

Most organizations are in the middle: they’d like to do more with live video, but they feel ill-equipped to meet audience expectations for engaging content, anywhere consumption and quality production. To bridge the gap between desire and execution of large video-based meetings, focus on three key areas:

  • Preparing virtual meeting content
  • Identifying ideal virtual meeting environments
  • Procuring virtual meeting talent and technology solutions


Preparing virtual meeting content

Large-scale events offer a rare opportunity to send a clear message, but only if leaders carefully hone the content for optimum engagement. On-screen leaders should prepare and practice their talk track so that it sounds polished, but not wooden. Leverage the same presentation techniques that keep in-person audiences engaged, like planned pauses, vocal emphasis and physical gestures. Use images, audio and video clips to complement and illustrate the ideas.

This sounds like a lot of work — and it is. Leaders might be tempted to go instead with voice narration over slide-share images, but the more static approach feels less personal and will promote less engagement.

Leaders should continue to participate on chat, through crowdsourcing and via social channels to hear what people care about

Fortunately, well-prepared, live content encourages audience interaction, which ultimately results in content contributions from the audience. Live video meetings can integrate tools to enable questions from the audience, live polls, real-time chat and live comments — all critical sources of audience interaction and content generation. 

“Nor should the interaction end with the close of a live stream,” says Preset. “Leaders should continue to participate on chat, through crowdsourcing and via social channels to hear what people care about after the video event ends.”

Gartner Top Priorities and Use Cases for Enterprise Video in the Digital Workplace.png

Identifying ideal virtual meeting environments

In an ideal world, all team members would be able to attend video-based all-company meetings at the same time and/or with everyone in their local office. That’s not always possible in widely dispersed organizations, or in operations that require a portion of the workforce to be on the job during the meeting. As a consequence, live video should be accessible on a range of screen types and though on-demand options to enable people in different work environments and time zones to watch when they’re available.

To accommodate this range, organizations should maximize distribution to as many platforms as possible. Options include desktops, mobile screens, conference room screens, and digital sign boards in common areas or break rooms, among others. Also include audio transcripts to ensure access.

“The environment” also includes the digital environments where employees look for company content. Organizations might push on-demand video content to workstations, or post it to a company website. Other options include posting the video in a messaging application, on social media or in workflow applications.

Read more: Gartner Top Technologies and Trends Driving the Digital Workplace

Procuring virtual meeting talent and technology solutions

Quality live video events require the right people and the right technology to execute. Simple events may need only an on-camera leader delivering content, whereas longer events with more segments may require a moderator to balance contributions, manage time, and ensure that questions and feedback are acknowledged. Off-camera teams also seed questions, solve off-screen technical problems and provide audience “service” to ensure the event runs smoothly. 

On the technology side, there are two options companies use most often to capture and deliver video for meetings: Meeting software that allows participants to join and watch a live-streamed presentation and enterprise video content management platforms that support live and on-demand delivery for both internal and external audiences. Companies might also post nonconfidential videos to sharing platforms like YouTube or Vimeo, or work with a webcasting company when the goal is to deliver a consumer experience.

The right solution might be the simplest. “The middle of a crisis or other period when people really need to hear from leaders is not the ideal time to figure out a new technology solution,” says Preset.

If you use a meeting solution for small meetings, check whether it can do large-scale ones as well. If you use a platform for training and marketing videos, see if it can do live video as well as on-demand. If you work with a unified communications vendor, service provider or system integrator already, they might have options to help you communicate at scale.

The post 3 Keys to Leading Large-Scale Virtual Meetings appeared first on Smarter With Gartner.

Vía IT – Smarter With Gartner https://ift.tt/2X5G4kF

Ask These Questions Before Deploying Remote Access Technology


Over the past two decades, remote access became a stable, but often neglected, technology. Now, coronavirus (COVID-19) changed the way employees need to work. Organizations now have old virtual personal network (VPN) technologies lacking the required licenses, updated features and adequate bandwidth to support all users working remotely and simultaneously.  

If there is an existing, workable product in place today, it still may not be optimal for providing the best experience for all users

“Corporate VPN is an aging technology as organizations shift to more cloud-based services,” says Rob Smith, Senior Director Analyst, Gartner. “However, in the wake of the global coronavirus pandemic, companies are realizing they have to fundamentally change the way they work. For security and risk management leaders, this means grappling with the best course of action to solve the challenges of large-scale modern remote access.”

Read more: With Coronavirus In Mind, Is Your Organization Ready for Remote Work?


4 questions to ask before you deploy

As companies support more work-from-home employees, they must have the right technology in place to ensure avoid poor performance and ensure secure access. Ask these four questions before deploying modern high-volume remote access products.

Question No. 1: Who is the user, and what is their job function?

All users are not equal. Some require more bandwidth than others, like executives or mission-critical employees with above-average data analysis needs. A user’s job function needs to be considered when defining any remote use case.

“Employees who simply check email will have different demands from those downloading and analyzing large sets of sensitive data,” says Smith. “Even if there is an existing, workable product in place today, it still may not be optimal for providing the best experience for all users.”

Question No. 2: What kind of device is being used, and who owns it?

Usability and security vary widely across the spectrum of available remote devices, like laptops and mobile devices. A corporate-owned PC is much easier to secure than a personally owned smartphone on which users are conducting concurrent activities and accessing websites that are potentially out of policy.

“Remote workers must ensure the same, if not a greater, level of security for all company networks and data access, documents or otherwise confidential information that might be displayed on a home office computer screen,” says Smith. “If security requirements prohibit storing data on individual personal devices, virtualization is an ideal option.” 

Question No. 3: What kind of applications and data do users need to access?

From a performance perspective,  employees using dedicated cloud applications and having an always-on VPN to the corporate network would not make as much sense as using a cloud access security broker (CASB). The way in which users access applications and data — either through on-premises or via the cloud — makes a difference when choosing remote access services.

Question No. 4: Where is the user located?

Data security, labor and privacy laws differ across countries and state/local jurisdictions, which creates an added layer of complexity when choosing offline data storage choices, and thus the remote access solution.

Read more: Coronavirus: CIO Areas of Focus During the COVID-19 Outbreak

Build a remote access end-user policy

After determining use cases and technology, build an end-user remote access policy with buy-in from all business units. If this is an urgent issue, like COVID-19, the policy must be escalated to legal counsel. Ensure that simple and local language is used, and stress the importance of employees physically signing the policy document as soon as possible. 

The post Ask These Questions Before Deploying Remote Access Technology appeared first on Smarter With Gartner.

Vía IT – Smarter With Gartner https://ift.tt/2UPnH0h

Create Contingency Plans for Events Canceled Due to COVID-19


Events rank as one of the most effective ways of generating leads for technology marketers. Technology and service providers spend an average of 11% of the marketing programs budget on third-party tradeshows, according to the Gartner 2019 Tech Marketing Benchmarks Survey. For hosted events, spend averages 6% of the program’s budget. 

But COVID-19 forced a change in how businesses approach events. In the current climate of limited (or totally canceled) travel, frozen budgets and daily changes on global health reports, determining the best approach as a technology marketer can be challenging. 

Cancellations or postponement of these events can be significantly disruptive to exhibitors or hosts. Eighteen percent of technology and service provider respondents ranked events as the best-performing calls to action of marketing-qualified leads, second only to content assets (20%). 


As the number of tradeshow cancellations increases daily, marketers need to implement contingency plans to ensure that awareness of and demand generation tied to events continues, regardless of event status.  

How product marketers should respond will vary depending on whether they are marketing a product at an event or hosting the event.

As an exhibitor 

Teams must be prepared to quickly pivot marketing strategy and build campaigns to address this shift. Examples include: For buyers who have already scheduled meetings with you, activate your sales development representatives (SDRs) and sales teams. Replace the event meetings with web meetings. Ideally, reach out within 24 hours of the event being canceled, as you’ll be competing with other exhibitors.

If you had product launches planned as part of an event, shift or increase spend in digital advertising to replace the launch. Use paid social to target buyers with web meetings or those on account-based marketing lists. You can also enlist your PR firm to expand media engagement stories to shift the plan for any planned product launches. Consider webinar content that conveys the key messages you planned to share at the tradeshow, including speaking sessions and in-booth presentations.

Finally, reach out to show hosts to understand promotional opportunities you may have with buyers that registered for the event. Are there opportunities to reach the audience through the host’s digital channels? For example, a tradeshow host or association may have a website and regular communications with attendees. Is there an opportunity to co-market through these channels? Are plans in place to reach the audience via streaming the event? 

As an event host 

If your marketing plans include a hosted event, such as user conferences, customer roundtables or field marketing events, ensure your teams have developed contingency plans.

Reach out to attendees and exhibitors to understand their plans. This will be highly dependent on when and where the event is planned to take place. Find out: Have their companies implemented travel restrictions? Will there be global or border travel restrictions in place? For how long? Are they currently planning on attending/exhibiting?

Develop ad creative and email campaigns to announce the cancellation or postponement to registrants now, and be prepared to quickly swap these messages into campaigns when/if necessary. If the organization cancels, t messaging delays are reduced. 

Marketers with contingency plans in place are less likely to experience disruptions

For large events or user conferences, establish a virtual event in place of the physical event. Virtual event platforms offer the ability to build exhibit “floors” with virtual booths for attendees to download relevant content and engage in online chats with booth staff. These platforms also enable you to deliver virtual experiences, including live streaming and on-demand general session keynotes and breakout presentations. For smaller or regional events, quickly pivot to a webinar approach to deliver your content.

Due to the fluid nature of the current environment, it’s necessary to build contingency plans for the foreseeable future. Marketers with contingency plans in place are less likely to experience disruptions to their awareness, demand generation and product launch efforts.

The post Create Contingency Plans for Events Canceled Due to COVID-19 appeared first on Smarter With Gartner.

Vía IT – Smarter With Gartner https://ift.tt/33MhgPS

Create the Right Type of Crisis Management Team


When an oxygen tank exploded during the Apollo 13 mission to land on the moon, NASA quickly formed what it later called a “tiger” team — a small group of specialists with different areas of expertise to investigate the issue and figure out how to get the four astronauts home. Their work guided the actions of the flight crew to turn a major crisis into a shining moment in the history of the U.S. space program.

Tiger teams fulfill a crucial risk management function, but they aren’t enough to fully protect organizations from experiencing negative financial, operational or reputational outcomes due to adverse events, such as a natural disaster, major security hack or global pandemic. That’s because the skills needed to predict and educate about adverse events are different from the skills needed in a moment of crisis.

The tiger team takes a first pass to identify procedures, and the SWAT team assesses their feasibility and recommends adjustments

“Anticipating adverse events is a qualitatively different task than responding to adverse impacts when they occur,” says Christie Struckman, VP Analyst, Gartner. “Many organizations use the same individuals in both scenarios, but not all team members are best suited for both situations.”

Instead of one team to fulfill all aspects of adverse event prediction and response, there should be two: A tiger team to anticipate and educate, and a SWAT team — derived from the “special weapons and tactics” units from military and police forces — to handle on-the-ground actions in crisis situations.


Tiger vs. SWAT: Who does what?

Tiger teams and SWAT teams each handle aspects of the six-stage adverse event assessment and mitigation process.

Different teams lead during different phases. For example, tiger teams lead the process of identifying possible adverse events and educating the organization, prioritizing which events to plan for and documenting the response. SWAT teams, for their part, lead the event response. 

The two teams collaborate on planning: The tiger team takes a first pass to identify procedures, and the SWAT team assesses their feasibility and recommends adjustments. And both teams particulate in evaluating the results after the fact and looking for ways to improve.

Both tiger and SWAT teams need people with concrete skills and competencies. Less concrete are the various ways of thinking — also known as cognitive styles — that are critical to meeting the needs of a tiger or a SWAT team.

Gartner outlines steps for planning and responding to adverse events with team ownership.

For example, people who tend to wait until they have all the relevant information before making a decision are well-suited to the methodical, research-driven work of a tiger team. People who are willing to act on limited knowledge, on the other hand, are essential SWAT team members. The former have a low tolerance for ambiguity; the latter, high.

Leaders can assess five cognitive dimensions in their tiger and SWAT team candidates to determine their cognitive style and create balanced teams. For each dimension, a person’s style falls on a continuum between two poles. The dimensions and their poles are:

  • Problem solving: Better vs. different. Does the person approach problems by looking to improve on what exists (better), or by reframing the problem altogether (different)?
  • Perspective: Part vs. whole. Does the person break an issue down into component parts and examine each one independently (part perspective, also known as analytical perspective), or does the person view events in terms of the entire system of which they are part and draw intuitive connections and associations (whole perspective, also known as intuitive perspective)?
  • Tolerance for ambiguity: High vs. low. Does the person gather as much information as they can before drawing conclusions (low tolerance), or are they willing to make a faster call based on partial insight (high tolerance)
  • Focus: People vs. task. Does the person thrive on collaborating with others (people focused), or on completing work and seeing results (task focused)?
  • Focus: Internal vs. external. Does the person think through solutions alone and then present them to others for feedback (internal), or does the person seek out group brainstorming and interaction to articulate and hone ideas (external)?

As with tolerance for ambiguity, the work of tiger and SWAT teams in many cases aligns with certain cognitive styles. People who automatically question and reframe problems are excellent tiger team members, skilled at identifying possible adverse events. The middle of a crisis is not the time to completely rethink a plan, however, so improvers make better SWAT members.

Yet staffing teams is not a binary exercise of placing people with one style on the tiger team and another on the SWAT team. “The most effective teams have a balance of cognitive styles,” says Struckman.

The post Create the Right Type of Crisis Management Team appeared first on Smarter With Gartner.

Vía IT – Smarter With Gartner https://ift.tt/2JbXvaT

4 Actions to Be a Good Leader During COVID-19 Disruption


COVID-19 is causing a humanitarian crisis of global proportions, with hundreds of thousands of lives disrupted. Sadly, we’re not nearing the end of the crisis.

In my day-to-day, I work with executive leaders and their teams on preparing for enterprise transformation. It’s no surprise that I’ve been musing for a while on what makes a great leader. In recent days, I’ve observed how COVID-19 is a test case for good and bad leadership. In my opinion, the leader’s primary responsibility is to keep the team safe, cohesive and productive. But what should a leader be focused on in the midst of a global disruption?

Leaders at any level can take four specific actions. These actions aren’t the only ones, but some of them might be missed in the rush to create an effective response.

No. 1: Create a central and clear list of priorities

Centrally set out a clear set of priorities, and locally implement them with a high degree of local manager autonomy. In a dynamic situation, overly centralizing decision making hobbles the organization’s ability to respond quickly and effectively. Effective leaders empower managers to make the best decisions they can, bearing in mind a clear set of enterprise priorities, such as keeping employees safe and behaving ethically toward customers. Clear thresholds should be put in place for when a local manager needs to ask for authorization from the center.

Instead of a clear, hierarchical list of priorities, mediocre leaders set out a buffet of priorities, all of which seem to be equal. It looks something like: “In this time of need, we need to be efficient, effective, innovative, secure, fast, agile and high quality.” The worst situations are where leaders provide unclear and contradictory priorities to their people.

Hard choices need to be made, never more so than during a crisis. Keep those choices, or priorities, hyperclear

Disruptions inevitably lead to an overload of sometimes-contradictory information. In the worst cases, employees are being given unclear or incoherent priorities. That’s why a crystal-clear set of priorities matters in times of upheaval, but is so hard to achieve.

The reality is that the buffet approach is not helpful. It’s like saying, “I want to be a blonde and a brunette and a redhead when I grow up.” Hard choices need to be made, never more so than during a crisis. Keep those choices, or priorities, hyperclear. Then allow for a high degree of decentralized decision making. Respecting and implementing those priorities locally, with the information available at the scene, is essential to be able to respond to emerging events.

No. 2: Pursue a nonbinary approach to problem solving

This is the opposite of what I call the “teenager approach to problem solving,” inspired by the book “Decisive,” by Chip and Dan Heath. Teenagers tend to adopt a binary approach to solving problems. “Should we break up or stay together?” or “I hate you, mum. I love you, mum!” or “I hate school, should I just drop out?”

There are almost always more options besides “do it/don’t do it” if teams can be creative about how they solve problems. For example, many leaders will be faced with the decision of whether to fire people and risk losing talent (not to mention further disrupting lives), or keep their people and risk their margins. But what if there were more options?

If you think hard enough about the problem, the options are not limited to “We need to fire people or risk our business.” More options could be considered, such as reducing full-time work to 80% for a portion of the workforce, or exploring early retirement options for some. Getting teams together to be creative and nonbinary about solving the rash of problems that crop up in a crisis can create lasting cohesion, not to mention more productivity and value for stressed-out customers and stakeholders.

[swg_ad id=”34349″]

No. 3: Be honest, empathetic, clean and simple

Hone a clear, honest, empathic and simple approach to communication as the COVID-19 situation evolves. This is obvious, I know. Yet I include it because it is so important and also because we live in an era of rapid dissemination of information that is of questionable veracity, from multiple sources.

Be the trusted source. Remember that in any communication plan, it matters less that you communicated something and more that your audience understood it. As a colleague of mine is fond of saying, it shouldn’t be called a communication plan — it should be called an understanding plan. The emphasis should be on what your listeners take in, especially in a situation that is volatile and unpredictable.

The 10x10x10 rule applies here: Say something 10 times in 10 different ways for people to retain 10%

Communicating well includes not being a victim of panic or hype, and laying out clear actions for what to do and when. The 10x10x10 rule applies here: Say something 10 times in 10 different ways for people to retain 10%. You’re probably not communicating enough unless you feel like you’re going blue in the face repeating the priorities and areas to focus on, as things change and morph.

Based on my anecdotal conversations with many different clients, I think a majority of organizations are doing this pretty well so far. The public sector bears the heaviest burden, as schools, regions and whole countries enter lockdown.

No. 4: Write down the stories

Collect stories of teams coming together to overcome the adversity caused by COVID-19. People are capable of amazing things in a time of crisis. Capture the detail of these stories; don’t summarize them too much. So, if Chris and Sandeep from the IT team were up until 5 a.m. in their respective home offices to get that VPN system up and running for hundreds of employees, and between them they drank 17 cups of coffee to get there, include it.

Humans respond to the detail of the story, so capture that. Over the past few days, I’ve heard of scientists pooling resources and heroically searching for a vaccine; I’ve heard of hilarious conference calls where both pets and children intervene as teams of people adjust to working from home; I’ve heard of absolutely heroic IT efforts, as that department rises to the challenge of getting entire companies to work remotely, overnight. These stories matter.

Read more: Coronavirus: CIO Areas of Focus During the COVID-19 Outbreak

As I said at the beginning, COVID-19 is a humanitarian tragedy that continues to disrupt millions of lives. Good leadership is crucial and there is perhaps one silver lining, which is that this crisis represents an opportunity for leaders to create more team cohesion and innovation in the face of adversity.

The post 4 Actions to Be a Good Leader During COVID-19 Disruption appeared first on Smarter With Gartner.

Vía IT – Smarter With Gartner https://ift.tt/2UcazC4

Understand 3 Key Types of Machine Learning


From fraud detection to image recognition to self-driving cars, machine learning (ML) and artificial intelligence (AI) will revolutionize entire industries. Together, ML and AI change the way we interact with data and use it to enable digital growth.

ML is a subset of AI that enables machines to develop problem-solving models by identifying patterns in data instead of leveraging explicit programming. The learning refers to the training process — the algorithms identify patterns in data and then use those patterns to tweak the model, aiming to provide a more accurate output each time. ML can be supervised, unsupervised or reinforced.

Through 2022, supervised learning will remain the type of ML utilized most by enterprise IT leaders

“Most of the current economic value gained from ML is based on supervised learning use cases,” says Saniye Alaybeyi, Senior Director Analyst, Gartner. “Yet unsupervised learning may be a better fit for certain problems — for example, when the goal is clustering entities and labeled data isn’t available. Reinforcement learning is still limited in its enterprise deployments, but its superior precision and targeting is promising for the future.”

Alaybeyi examines the three types of ML used in enterprise AI programs today and the business problems that each can solve.

Supervised learning

Through 2022, supervised learning will remain the type of ML utilized most by enterprise IT leaders. Supervised learning is effective in many business scenarios, such as fraud detection, sales forecasting and inventory optimization. 

Supervised learning works by feeding known historical input and output data into ML algorithms. In each step, after processing each input-output pair, the algorithm alters the model to create an output that is as close as possible to the desired result.

Supervised learning can be used to make predictions, recognize data or classify it

For example, a model could be fed data from thousands of bank transactions, with each transaction labeled as fraudulent or not. The model will identify patterns that led to a “fraudulent” or “not fraudulent” output, and over time, learn to more accurately predict whether a given transaction is fraudulent. 

Input and output data can be derived from historical data, through simulations or through human data labeling. In cases involving unstructured data, like images, video, audio or text, certain properties or categorizations can serve as output data. Supervised learning can be used to make predictions, recognize data or classify it. 

Example use cases for supervised learning include:

  • Identifying risk factors for diseases and planning preventive measures
  • Classifying whether or not an email is spam
  • Predicting housing prices
  • Predicting whether or not people will vote for a given candidate.
  • Finding out whether a loan applicant is low- or high-risk
  • Predicting the failure of mechanical parts in industrial equipment

Read more: The CIO’s Guide to Artificial Intelligence


Unsupervised learning

Unsupervised learning is used to develop predictive models from data that consists of input data without historical labeled responses. For example, a list of customers or a set of unlabeled photos could serve as input data in an unsupervised learning use case.

The most common applications of unsupervised learning are clustering and association problems. Clustering produces a model that groups objects based on certain properties, such as color. Association takes those clusters and identifies rules that exist between them.

Example use cases for unsupervised learning include:

  • Grouping customers by purchase behavior
  • Identifying associations in customer data; for instance, people who buy a certain style or shoe may also be interested in a certain style of bag. 
  • Classifying people based on different interests
  • Grouping inventories by manufacturing and sales metrics

Unsupervised learning can also be used to prepare data for subsequent supervised learning. This is done by identifying patterns or features that can be used to categorize, compress and reduce the dimensionality of data.

Reinforcement learning

Reinforcement learning (RL) is based on rewarding desired behaviors or punishing undesired ones. Instead of one input producing one output, the algorithm produces a variety of outputs and is trained to select the right one based on certain variables. So, for example, a computer program could be trained to win a video game by identifying patterns in the actions that lead to it scoring more points than the other players.

Although it has been around for decades, RL has recently seen a renewed interest. RL requires less management than supervised learning, making it easier to work with unlabeled datasets. There have been some recent successes in RL implementations in the gaming world. However, practical RL applications are still emerging. 

Recognize the potential opportunities for RL, but only employ it in limited scenarios

Most current data science and ML platforms don’t have native RL capabilities, and it requires much higher computing power than most enterprises have available. Right now, RL is only applicable in areas that can be fully simulated, that are quite stationary or where massive amounts of relevant data are available. 

Example use cases for reinforcement learning include:

  • Load balancing, for example, by teaching the algorithm to minimize the number of jobs waiting based on available compute resources
  • Solving traffic jam problems by dynamically controlling traffic lights
  • Training a robot to learn policies by mapping input from raw video images and replicating the actions it has “seen”
  • Teaching a car to park itself, thus reducing time-consuming and trial-and-error work

AI leaders must develop significantly better simulation capabilities before RL can enter mass adoption. Recognize the potential opportunities for RL, but only employ it in limited scenarios.

The post Understand 3 Key Types of Machine Learning appeared first on Smarter With Gartner.

Vía IT – Smarter With Gartner https://ift.tt/38Zybzx