Physician, heal thyself Luke 4:23

My Thermos bottle keeps hot drinks hot and cold drinks cold. How does it know?

An electrical engineer would probably design a Thermos with a toggle switch (HOT and COLD), or a big temperature dial, or if you work in Cupertino an LCD display, touchpad, RTOS, and proprietary cable interface. Thankfully, real vacuum flasks take care of themselves with no user input at all. They just work.

It would sure be nice if new AI-enabled IoT devices could do the same thing. Instead of learning all about AI and ML (and the differences between the two), and learning how to code neural nets, and how to train them, and what type of data they require, and how to provision the hardware, etc., itd be great if they just somehow knew what to do. Now that would be real machine learning.

Guess what? A small French company thinks it has developed that very trick. It uses machine learning to teach machine learning. To machines. Without a lot of user input. It takes the mystery, mastery, and mythology out of ML, while allowing engineers and programmers to create smart devices with little or no training.

The company is Cartesiam and the product is called NanoEdge AI Studio. Its a software-only tool that cranks out learning and inference code for ARM Cortex-Mbased devices, sort of like an IDE for ML. The user interface is pretty to look at and has only a few virtual knobs and dials that you get to twist. All the rest is automatic. Under the right circumstances, its even free.

Cartesiams thesis is that ML is hard, and that developing embedded AI requires special skills that most of us dont have. You could hire a qualified data scientist to analyze your system and develop a good model, but such specialists are hard to find and expensive when theyre available. Plus, your new hire will probably need a year or so to complete their analysis and thats before you start coding or even know what sort of hardware youll need.

Instead, Cartesiam figures that most smart IoT devices have certain things in common and dont need their own full-time, dedicated data scientist to figure things out, just like you dont need a compiler expert to write C code or a physicist to draw a schematic. Let the tool do the work.

The company uses preventive motor maintenance as an example. Say you want to predict when a motor will wear out and fail. You could simply schedule replacement every few thousand hours (the equivalent of a regular 5000-mile oil change in your car), or you could be smart and instrument the motor and try to sense impending failures. But what sensors would you use, and how exactly would they detect a failure? What does a motor failure look like, anyway?

With NanoEdge AI Studio, you give it some samples of good data and some samples of bad data, and let it learn the difference. It then builds a model based on your criteria and emits code that you link into your system. Done.

You get to tweak the knobs for MCU type, RAM size, and type of sensor. In this case, a vibration sensor/accelerometer would be appropriate, and the data samples can be gathered in real-time or canned; it doesnt matter. You can also dial-in the level of accuracy and the level of confidence in the model. These last two trade off precision for memory footprint.

NanoEdge Studio includes a software simulator, so you can test out your code without burning any ROMs or downloading to a prototype board. That should make it quicker to test out various inference models to get the right balance. Cartesiam says it can produce more than 500 million different ML libraries, so its not simply a cut-and-paste tool.

As another example, Cartesiam described one customer designing a safety alarm for swimming pools. They spent days tossing small children into variously shaped pools to collect data, and then several months analyzing the data to tease out the distinguishing characteristics of a good splash versus one that should trigger the alarm. NanoEdge AI Studio accomplished the latter task in minutes and was just as accurate. Yet another customer uses it to detect when a vacuum cleaner bag needs emptying. Such is the world of smart device design.

The overarching theme here is that users dont have to know much of anything about machine learning, neural nets, inference, and other arcana. Just throw data at it and let the tool figure it out. Like any EDA tool, it trades abstraction for productivity.

In todays environment, thats a good tradeoff. Experienced data scientists are few and far between. Moreover, you probably wont need his/her talents long-term. When the project is complete and youve got your detailed model, what then?

NanoEdge AI Studio is free to try but deploying actual code in production costs money. Cartesiam describes the royalty as tens of cents to a few dollars, depending on volume. Sounds cheaper than hiring an ML specialist.

Related

View post:
Machine Learning at the Push of a Button - EE Journal

By: Mark Krupnik, PhD, Founder and CEO, Retalon

Since this article will probably come out during Income tax season, let me start with the following example: Suppose we would like to build a program that calculates income tax for people. According to US federal income tax rules: For single filers, all income less than $9,875 is subject to a 10% tax rate. Therefore, if you have $9,900 in taxable income, the first$9,875 is subject to the 10% rate and the remaining $25 is subject to the tax rate of the next bracket (12%).

This is an example of rules or an algorithm (set of instructions) for a computer.

Lets look at this from a formal, pragmatic point of view. A computer equipped with this program can achieve the goal (calculate tax) without human help. So technically, this can be classified as Artificial Intelligence.

But is it cool enough? No. Its not. That is why many people would not consider it part of AI. They may say that if we already know how to do a certain thing, then the process cannot be considered real intelligence. This is a phenomena that has become known as AI Effect. One of the first references is known as Teslers theorem that says: AI is whatever hasnt been done yet.

In the eyes of some people, the cool part of AI is associated with machine learning, and more specifically with deep learning which requires no instructions and utilizes Neural Nets to learn everything by itself, like a human brain.

The reality is that human development is a combination of multiple processes, including both: instructions, and Neural Net training, as well as many other things.

Lets take another simple example: If you work in a workshop on a complex project, you may need several tools, for instance a hammer, a screwdriver, plyers, etc. Of course, you can make up a task that can be solved by only using a hammer or only screwdriver, but for most real-life projects you will likely need to use various tools in combination to a certain extent.

In the same manner, AI also consists of several tools (such as algorithms, supervised and unsupervised machine learning, etc.). Solving a real-life problem requires a combination of these tools, and depending on the task, they can be used in different proportions or not used at all.

There are and there will always be situations where each of these methods will be preferred over others.

For example, the tax calculation task described in the beginning of this article will probably not be delegated to machine learning. There are good reasons to it, for example:

the solution of this problem does not depend on data the process should be controllable, observable, and 100% accurate (You cant just be 80% accurate on your income taxes)

However, the task to assess income tax submissions to identify potential fraud is a perfect application for ML technologies.

Equipped with a number of well labelled data inputs (age, gender, address, education, National Occupational Classification code, job title, salary, deductions, calculated tax, last year tax, and many others) and using the same type of information available from millions of other people, ML models can quickly identify outliers.

What happens next? The outliers in data are not necessarily all fraud. Data scientists will analyse anomalies and try to understand the reason for these individuals being flagged. It is quite possible that they will find some additional factors that had to be considered (feature engineering), for example a split between tax on salary, and tax on capital gain of investment. In this case, they would probably add an instruction to the computer to split this data set based on income type. At this very moment, we are not dealing with a pure ML model anymore (as the scientists just added an instruction), but rather with a combination of multiple AI tools.

ML is a great technology that can already solve many specific tasks. It will certainly expand to many areas, due to its ability to adapt to change without major effort on a human side.

At the same time, those segments that can be solved using specific instructions and require predictable outcome (financial calculations) or those involving high risk (human life, health, very expensive and risky projects) require more control and if the algorithmic approach can provide it, it will still be used.

For practical reasons, to solve any specific complex problem, the right combination of tools and methods of both types are required.

About the Author:

Mark Krupnik, PhD, is the founder and CEO ofRetalon, an award-winning provider of retail AI and predictive analytics solutions for planning, inventory optimization, merchandising, pricing and promotions.Mark is a leading expert on building and delivering state-of-the-art solutions for retailers.

See the original post here:
Is Machine Learning Always The Right Choice? - Machine Learning Times - machine learning & data science news - The Predictive Analytics Times

Dive Brief:

As technologies such as AI and machine learning revolutionize the workplace, learning and development is coming to the forefront of talent management. Preparing workers for AI and automation will lead learning trends in 2020, according to a November 2019 Udemy report. While many workplaces will train employees to sharpen their tech skills, the report said, learning professionals will also need to focus on soft skills and skills related to project management, risk management and change management.

About 120 million workers around the world will need access to retraining opportunities, a need at least partly driven by AI and automation, according to a report from IBM. This need vastly outpaces the number of organizations equipped with resources that suffice for such an effort, however.

Platforms such as O'Reilly may aid in filling this gap. Third-party training programs are growing in popularity with seemingly positive results. Managers may prefer coders with training from a boot camp, for example, a recent report from HackerRank found. But there has been at least one report that external L&D programs boast false results;New York Magazine reported Lambda School, "a 'boot camp' for people who want to quickly learn how to code," has inflated the number of job placements secured by its graduates.

Visit link:
Interest in machine learning and AI up, though slowing, one platform reports - HR Dive

You may not know it, but machine learning and AI have pervaded our everyday lives and made it not just more convenient, but introduces new quality of life changes as well.

Predicting the Lottery

Before, it was very difficult to predict the next winning set of numbers in a lottery game as youll need a proven algorithm and the computing muscle of a powerful machine but that has changed with the emergence of artificial intelligence.

Today, predicting the results of an Arizona lottery are shown on websites such as The Lotto Pro. It uses both AI and machine learning for winning number recommendations, coupled with an advanced algorithm.

The site actually predicted the right numbers for the AZ lottery 2/25/2020 draw, which speaks a lot about its accuracy.

Knowing What Youre Searching For

Typing on a search query will result in a list of predictions that you can choose to save time. This feature draws on your past searches and artificial intelligence, along with personal details, age and location.

Search engines also get better over time as they collect data such as the length of time you spent on a page and your response when presented with a list of sites with your keyword.

AI Assistants

The rise of Siri, Google Assistant and Alexa have opened up ways on how we control technology. Without needing physical input, we can make our smart devices play music, turn on the lights, open the front door and more.

More than that, you can use these AI platforms to set up alarms, reminders and schedule meetings just like you would a real assistant.

Social Media

How you get your daily feed depend on AI and machine learning. Targeted ads, deleting inappropriate and offensive tweets and friend suggestions are all handled by an algorithm thats backed by artificial intelligence.

Read the original here:
AI and Machine Learning in Everyday Life - IMC Grupo

NFL teams use machine learning.

Dont believe me? Take it from Andrew Berry, the new general manager of the Cleveland Browns, who said in a recent interview (while he was still working for the Eagles) that machine learning techniques and data are having more of an impact in all the different spaces of football operations. In a talk given while he was working for the team, former Patriots senior software engineer Sean Harrington said his job included machine learning, data warehousing and visualizations.

As more teams embrace the use of data and advanced statistical techniques to inform decision-making, they are looking to hire analysts to make sense of the information and avoid being left behind. But without connections to the technical world, finding the right people can be challenging.

Enter the Big Data Bowl, the NFLs annual competition in which data scientists and anyone else...

Go here to read the rest:
How Big Data Bowl winners used machine learning to better understand the game - The Athletic

(Dmitriy Rybin/Shutterstock)

Many companies are scrambling to find machine learning engineers who can build smart applications that run on edge devices, like mobile phones. One company thats attacking the problem in a broad way is Qeexo, which sells an AutoML platform for building and deploying ML applications to microcontrollers without writing a line of code.

Qeexo emerged from Carnegie Mellon University in 2012, just at the dawn of the big data age. According to Sang Won Lee, the companys co-founder and CEO, the original plan called for Qeexo to be a machine learning application company.

The company landed a big fish, the Chinese mobile phone manufacturer Huawei, right out the gate. Huawei liked the ML-based finger-gesture application that Qeexo (pronounced Key-tzo) developed, and the company wanted Qeexo to ensure that it could run across all of its phone lines. That was a good news-bad news situation, Lee says.

Our first commercial implementation with Huawei kept the whole company in China for two months, to finish one model with one hardware variant, Lee tells Datanami. We came back and it was difficult to keep the morale high for our ML engineers because nobody wanted to constantly go abroad to do this type of repetitive implementation.

Qeexos AutoML solution handles many aspects of ML model development and deployment for customers

It quickly dawned on Lee that, with more ML models and more hardware types, the amount of manual work would quickly get out of hand. That led him to the idea of developing an automated machine learning, or AutoML, platform that could automatically generate ML models based on the data presented to it, automatically flash it to a group of pre-selected microcontrollers.

Lee and his team of developers, which is led by CTO and co-founder Chris Harrison (who is an assistant professor at Carnegie Mellon University), developed the offering nearly five years ago, and the company has been using it ever since for its own ML services engagements.

Huawei continues to utilize Qeexos AutoML solution to generate ML applications for its handsets. In 2018, we completed 57 projects for Hauwei, and most of the projects were completed by our field engineers just using the AutoML platform, without the help of ML engineers in the US, Lee says.

Sang Won Lee is the co-founder and CEO of Qeexo

In October 2019, Qeexo released its AutoML offering as a stand-alone software offering. The product automates many steps in the ML process, from building models from collected data, comparing performance of those models, and then deploying the finished model to a microcontroller all without requiring the user to write any code.

The offering has built-in support for the most popular ML algorithms, including random forest, gradient boosted machine, and linear regression, among others. Users can also select deep learning models, like convolutional neural networks, but many microcontrollers lack the memory to handle those libraries, Lee says.

Qeexos AutoML solution automatically handles many of the engineering tasks that would otherwise require the skills of a highly trained ML engineer, including feature selection and hyperparameter optimization. These feature are built into the Qeexo offering, which also sports a built-in C compiler and generates binary code that can be deployed to microcontrollers, such as those from Renasas Electronics.

Lee says ML engineers might be able to get a little more efficiency by developing their own ML libraries, but that it wont be worth the effort for many users. There are always more improvements that you can get with ML experts digging into it and doing the research, he says. But this is giving you the convenience of being able to build a commercially viable solution without having to write a single line of code.

Today Qeexo announced its new AWS solution. Instead of training a model on a laptop, customers can use now AWS resources to train their model. It also announced more ML algorithms, including deep learning algorithms and traditional algorithms. The visualization that Qeexo provides have also been enhanced to give the user the ability to better spot outliers and trends in data. Support for microphone data has been supported. And it also added support for the Renesas RA Family of Cortex-M MCUs, which are geared toward low-power IoT edge devices.

Having Huawei as a client certainly gives Qeexo some experience with scalability under its belt. But the Mountain View, California-based company is bullish on the potential for a new class of application developers to get started using its software to imbue everyday devices with the intelligence of data.

What we really want to tell the market is that even or those microcontrollers that are already out and that have very limited memory resource and processing power, you can still have a commercially viable ML solution running on it, if you use the right tool, Lee says. You dont want to neglect all the sensor data thats connected to the microcontroller. We can provide a tool that you can use to build intelligence that can be embedded into those tools.

Related Items:

On the Radar: Cortex Labs, RadicalBit, Qeexo

AutoML Tools Emerge as Data Science Difference Makers

Cloud Tools Rev Up AI Dev Platforms

Read this article:
Machine Learning on the Edge, Hold the Code - Datanami

Machine learning is helping Penn Medicine researchers identify the size and shape of brain networks in individual children, which may be useful for understanding psychiatric disorders. In a new study publishedin the journalNeuron,a multidisciplinary team showed how brain networks unique to each child can predict cognition. The study, which used machine learning techniques to analyze the functional magnetic resonance imaging (fMRI) scans of nearly 700 children, adolescents, and young adults, is the first to show that functional neuroanatomy can vary greatly among kids, and is refined during development.

The human brain has a pattern of folds and ridges on its surface that provide physical landmarks for finding brain areas. The functional networks that govern cognition have long been studied in humans by liningup activation patternsthe software of the brainto the hardware of these physical landmarks. However, this process assumes that the functions of the brain are located on the same landmarks in each person. This works well for many simple brain systems. However, multiple recent studies in adults have shown this is not the case for more complex brain systems responsible for executive functiona set of mental processes which includes self-control and attention. In these systems, the functional networks do not always line up with the brains physical landmarks of folds and ridges. Instead, each adult has their own specific layout. Until now, it was unknown how such person-specific networks might change as kids grow up, or relate to executive function.

The exciting part of this work is that we are now able to identify the spatial layout of these functional networks in individual kids, rather than looking at everyone using the same one size fits all approach, says senior authorTheodore D. Satterthwaite, an assistant professor of psychiatry in the Perelman School of Medicine. Like adults, we found that functional neuroanatomy varies quite a lot among different kidseach child has a unique pattern. Also like adults, the networks that vary the most between kids are the same executive networks responsible for regulating the sorts of behaviors that can often land adolescents in hot water, like risk taking and impulsivity.

Read more at Penn Medicine News.

Here is the original post:
Machine learning identifies personalized brain networks in children - Penn: Office of University Communications

Artificial intelligence (AI) understood as intelligence demonstrated by machines and machine learning (ML) a subfield of AI that focuses on machines ability to automatically learn from experience without being explicitly programmed are two of the computer science technologies that will arguably have the biggest impact on our lives in the coming decades. AI and ML are not exactly new both terms date back to the mid-20th century but it is mainly in recent years that their huge potential has become evident.

The technologies are now being used more and more extensively across a whole range of fields. Artificial intelligence can be found in autonomous vehicles, search engines, online assistants and spam filtering programs, to name just a few of its numerous common applications. Applications of machine learning include computer vision, DNA sequence classification, financial market analysis, Internet fraud detection, medical diagnosis and speech recognition. Both AI and ML are currently transforming whole sectors and will continue to do so in the future.

Pushing the boundaries

Of course, behind that transformation are concrete people, outstanding engineers and scientists who continue to push the boundaries of how theoretical concepts related to computer science and other similar fields can be turned into practical use. Curious folks, always on the lookout for new possibilities. Consider Pierre-Habt Nouvellon, the CTO and Head of Machine Learning at Snipfeed, a California-based startup that has come up with an eponymous AI-based news and information recommendation engine, mainly targeted at young (Generation Z) users.

Snipfeed, which he co-founded, provides daily, highly personalized pieces of news and information (snippets) that are filtered and recommended to its users depending on their needs. The technology used by the engine ensures that Snipfeed avoids recommending articles which do not look credible. An internal tool called Fakebuster detects fake news articles by verifying content, assessing the reliability of the facts in it and the quality of the writing style and the headline.

Falling for ML

Nouvellon discovered machine learning while he was doing a Masters degree in aerospace engineering in France. He was working on a research project on theoretical supply chain problems and his supervisor advised him to see if reinforcement learning, a subfield of machine learning, could help him complete his task. I took the well-known Andre Ng course in machine learning and fell in love with it. It was the perfect mix of statistics, algebra and computer science, he recalled.

After that initial experience, Nouvellon realized that he wanted to study machine learning more thoroughly. He applied for a Masters program in computer science at the University of California, Berkeley, and was admitted. There he met some of the leading professors in the field. Once in Berkeley, Nouvellon and two of his friends started working on an AI-based tutor called Jenyai, a conversational bot on messenger that was able to answer students questions in math, history and science.

One of the applications features was a selection of news stories meant to help students better understand the discussed topics. This feature proved to be very popular with Jenyais users, which inspired Nouvellon and his partners to create Snipfeed. Nouvellon also had a chance to apply his knowledge of machine learning in a completely different field healthcare. He was involved in a medical research project done by the University of California, Berkeley, and the University of California, San Francisco.

Driven by curiosity

The projects goal was to apply machine learning algorithms to predict the occurrence of a certain disease among a major pool of patients by analyzing those patients electronic medical records and genotype data, provided by a large medical insurance group. The application of ML technologies in medicine and healthcare is set to become more and more commonplace in the near future, but such technologies will also be transforming many other sectors. The revolution will be driven by people like Nouvellon, who are always inspired by the new and unknown.

Becoming an inventor is something that he thought of from his early years. Nouvellon revealed that he had grown up in a family were science and creativity played a central role. Already when he was a child, he liked innovating, playing with such ideas as making up a new language and building a complex electric circuit. He would dream of becoming an astronaut, a chemist or a mathematician. Inventing is doing something that has not been done before and thus there is an adventurous and exciting aspect to it that I have always loved, Nouvellon admitted.

This article does not necessarily reflect the opinions of the editors or management of EconoTimes.

See original here:
The perfect mix. The story of an ML engineer - EconoTimes

2020 has started as 2019 ended, with new cyberattacks, hacking incidents and data breaches coming to light almost every day.

Cyber criminals pose a threat to all manner of organisations and businesses, and the customers and consumers who use them. Some of the numbers involved in the largest data breaches are staggering, with personal data concerning hundreds of thousands of individuals being leaked each one potentially a new victim of fraud and other cybercrime.

SEE: Cybersecurity in an IoT and mobile world (ZDNet special report) | Download the report as a PDF (TechRepublic)

Businesses are doing their best to fight off cyberattacks, but it's hard to predict what new campaigns will emerge and how they'll operate. It's even harder to discern what the next big malware threat will be: the Zeus trojan and Locky ransomware were once major threats, but now it's things like Emotet botnet, the Trickbot trojan and Ryuk ransomware.

It's difficult to defend your perimeter against unknown threats -- and that's something thatcyber criminals take advantage of.

Artificial intelligence (AI) and machine learning (ML) are playing an increasing role in cybersecurity, with security tools analysing data from millions of cyber incidents, and using it to identify potential threats -- an employee account acting strangely by clicking on phishing links, for example, or a new variant of malware.

But there is a constant battle between attackers and defenders. Cyber criminals have long tried to tweak their malware code so that security software no longer recognises it as malicious.

Spotting every variation of malware, especially when it is deliberately disguised, is hard: increasingly it's by applying AI and ML that defenders are attempting to stop even the unknown, new types of malware attack.

"Machine learning is a good fit for anti-malware solutions because machine learning is well suited to solve 'fuzzy' problems," says Josh Lemos, vice-president of research and intelligence at Cylance, aBlackBerry-owned, AI-based cybersecurity provider working out of California.

The machine-learning database can draw upon information about any form of malware that's been detected before. So when a new form of malware appears -- either a tweaked variant of existing malware, or a new kind entirely -- the system can check it against the database, examining the code and blocking the attack on the basis that similar events have previously been deemed as malicious.

That's even the case when the malicious code is bundled up with large amounts of benign or useless code in an effort to hide the nefarious intent of the payload, as often happens.

It was these machine-learning techniques that enabled Cylance to uncover -- and protect users against -- a new campaign by OceanLotus, a.k.a. APT32, a hacking group linked to Vietnam.

"As soon as they came out with a new variant in the wild, we knew exactly what it was because we had some machine-learning signatures and models designed to orient to these variants when they appear. We knew they're close enough in their genetic make-up to be from this family of threat," Lemos explains.

SEE: This latest phishing scam is spreading fake invoices loaded with malware

But uncovering new kinds of malware isn't the only way machine learning can be deployed to boost cybersecurity: an AI-based network-monitoring tool can also track what users do on a daily basis, building up a picture of their typical behaviour. By analysing this information, the AI can detect anomalies and react accordingly.

"Think about what AI is really good at -- the ability to adapt and respond to a constantly changing world", says Poppy Gustafsson, co-CEO of Darktrace, a British cybersecurity company that uses machine learning to detect threats.

"What AI enables us to do is to respond in an intelligent way, understanding the relevance and consequences of a breach or a change of behaviour, and in real time develop a proportionate response," she adds.

For example, if an employee clicks on a phishing link, the system can work out that this was not normal behaviour and could therefore be potentially malicious activity.

Using machine learning, this can be spotted almost immediately, blocking the potential damage of a malicious intrusion and preventing login credentials being stolen, malware being deployed or otherwise enabling attackers to gain access to the network.

And all of this is done without the day-to-day activity of the business being impacted, as the response is proportionate: if the potential malicious behaviour is on one machine, that doesn't require the whole network being locked down.

A key benefit of machine learning in cybersecurity is that it identifies and reacts to suspected problems almost immediately, preventing potential issues from disrupting business.

By deploying AI-based cybersecurity from Darktrace to automate some of the defence functions, the McLaren Formula One team aims to ensure that the network is going to be safe, without relying on humans having to perform the impossible the task of monitoring everything at once.

"If we can't see data coming off the car, if we're compromised, we stop racing -- so high-speed decision-making from machines makes it safer," Karen McElhatton, Group CIO at McLaren explains. "Data isn't just bits and bytes: we have video, we have chats, emails -- it's the variety of that input that's coming and the growing volume of it. It's too much for humans to be able to manage and automated tools are opening our eyes up to what we need to be watching."

That's especially the case when it comes to monitoring how employees operate on the network. Like other large organisations, McLaren employs training to help staff improve cybersecurity, but it's possible that staff will attempt to take shortcuts in an effort to do their job more efficiently -- which could potentially lead to security issues. Machine learning helps to manage this.

"We've got really clever people at McLaren, but with smart people come creative ways of getting around security, so having that intelligence response is really important to us. We can't slow decision-making or innovation down, but we need to enable them to do it safely and securely -- and that's where Darktrace helps us," McElhatton explains.

But while AI and ML do provide benefits for cybersecurity, it's important for organisations to realise that these tools aren't a replacement for human security staff.

It's possible for a machine learning-based security tool to be programmed incorrectly, for example, resulting in unexpected -- or even obvious -- things being missed by the algorithms. If the tool misses a particular kind of cyberattack because it hasn't been coded to take certain parameters into account, that's going to lead to problems.

SEE: IoT security is bad. It's time to take a different approach.

"Where AI and machine learning can get you into trouble is if you are reliant on it as an oracle of everything," says Merritt Maxim, researcher director for security at analyst firmForrester .

"If the inputs are bad and it's passing things through it says are okay, but it's actually passing real vulnerabilities through because the model hasn't been properly tuned or adjusted -- that's the worst case because you think you're fully protected because you have AI".

Maxim notes that AI-based cybersecurity has "a lot of benefits", but isn't a complete replacement for human security staff; and like any other software on the network, you can't just install it and forget about it -- it needs to be regularly evaluated.

"You can't assume that AI and machine learning are going to solve all the problems," he says.

Indeed, there's the potential that AI and machine learning could create additional problems, because while the tools help to defend against hackers, it's highly likely that cyber criminals themselves are going to use the same techniques in an effort to make attacks more effective.

A report by Europol has warned that artificial intelligence is one of the emerging technologies that could make cyberattacks more dangerous and more difficult to spot than ever before. It's even possible that cyber criminals have already started using these techniques to help conduct hacking campaigns and malware attacks.

"A lot of it is, at the moment, theoretical, but that's not to say that it hasn't happened. It's quite likely that it's been used, but we just haven't seen it or haven't recognised it," says Philipp Amann, head of strategy for Europol's European Cybercrime Centre (EC3).

It's possible that by using machine learning, cyber criminals could develop self-learning automated malware, ransomware, social engineering or phishing attacks. Currently, they might not access to the deep wells of technology that cybersecurity companies have, but there's code around that can provide cyber criminals with access to these resources.

"The tools are out there -- some of them are open source. They're freely available to everyone and the quality is increasing, so it's likely to assume that this will be part of a criminal's repertoire if it isn't already," Amann says.

While it may be unclear if hackers have used machine learning to help develop or distribute malware, there is evidence of AI-based tools being used to conduct cybercrime; last year it was reported that criminals used AI generated audio to impersonate a CEO's voice and trick employees into transferring over 220,000 ($243,000) to them.

This 'deepfake voice attack' added a new layer to business email compromise scams, in which attackers claim to be the boss and request an urgent transfer of funds. This time, however, the attackers used AI to mimic the voice of the CEO and request the transfer of funds.

The nature of a CEO's job means their voice is often in the public domain, so criminals can find and exploit voice recordings and -- and this case isn't the only example.

"Other industry partners have confirmed other cases that haven't been reported; so that's where we know an AI-based system has been used as part of a social-engineering attack," says Amann.

AI-based deepfake technology has already caused concern when it comes to spreading disinformation or even abuse of individuals via fake videos, leading tocalls for deepfake regulation.

SEE: California takes on deepfakes in porn and politics

But the problem here is that while most people would see this as a warning not to meddle in this field, cyber criminals will simply look to take advantage of any technology in order to make money from malicious hacking. For example, machine learning could be employed to send out phishing emails automatically and learn what sort of language works in the campaigns, what generates clicks and how attacks against different targets should be crafted.

Like any machine-learning algorithm, success would come from learning over time, meaning that it's possible that phishing attacks could be driven in the same way security vendors attempt to defend against them.

"There's a whole cyber criminal element here for financial gain, which can leverage AI and machine learning effectively," warns Lemos.

However, if AI-based cybersecurity tools continue to develop and improve, and are applied correctly alongside human security teams, rather than instead of them, this could help businesses stay secure against increasingly smart and potent cyberattacks.

"One thing we can be certain of is the offices of tomorrow aren't going to like those of past. AI is how technology responds to our ever-changing world. It updates automatically and learns how humans react. We can't second-guess technology, but we can watch it and learn from it and adapt," says Gustafsson.

"You could move into a world where your whole cybersecurity posture is enhanced, with the ultimate vision being you could end up with a self-learning and self-healing network that can learn negative behaviours and stop them happening," she says.

Read this article:
AI is changing everything about cybersecurity, for better and for worse. Here's what you need to know - ZDNet