Category Archives: Machine Learning

Understanding failures and attacks can help us build safer AI applications.

Modern machine learning (ML) has become an important tool in a very short time. We're using ML models across our organisations, either rolling our own in R and Python, using tools like TensorFlow to learn and explore our data, or building on cloud- and container-hosted services like Azure's Cognitive Services. It's a technology that helps predict maintenance schedules, spots fraud and damaged parts, and parses our speech, responding in a flexible way.

SEE:Prescriptive analytics: An insider's guide (free PDF)(TechRepublic)

The models that drive our ML applications are incredibly complex, training neural networks on large data sets. But there's a big problem: they're hard to explain or understand. Why does a model parse a red blob with white text as a stop sign and not a soft drink advert? It's that complexity which hides the underlying risks that are baked into our models, and the possible attacks that can severely disrupt the business processes and services we're building using those very models.

It's easy to imagine an attack on a self-driving car that could make it ignore stop signs, simply by changing a few details on the sign, or a facial recognition system that would detect a pixelated bandanna as Brad Pitt. These adversarial attacks take advantage of the ML models, guiding them to respond in a way that's not how they're intended to operate, distorting the input data by changing the physical inputs.

Microsoft is thinking a lot about how to protect machine learning systems. They're key to its future -- from tools being built into Office, to its Azure cloud-scale services, and managing its own and your networks, even delivering security services through ML-powered tools like Azure Sentinel. With so much investment riding on its machine-learning services, it's no wonder that many of Microsoft's presentations at the RSA security conference focused on understanding the security issues with ML and on how to protect machine-learning systems.

Attacks on machine-learning systems need access to the models used, so you need to keep your models private. That goes for small models that might be helping run your production lines as much as the massive models that drive the likes of Google, Bing and Facebook. If I get access to your model, I can work out how to affect it, either looking for the right data to feed it that will poison the results, or finding a way past the model to get the results I want.

Much of this work has been published in a paper in conjunction with the Berkman Klein Center, on failure modes in machine learning. As the paper points out, a lot of work has been done in finding ways to attack machine learning, but not much on how to defend it. We need to build a credible set of defences around machine learning's neural networks, in much the same way as we protect our physical and virtual network infrastructures.

Attacks on ML systems are failures of the underlying models. They are responding in unexpected, and possibly detrimental ways. We need to understand what the failure modes of machine-learning systems are, and then understand how we can respond to those failures. The paper talks about two failure modes: intentional failures, where an attacker deliberately subverts a system, and unintentional failures, where there's an unsafe element in the ML model being used that appears correct but delivers bad outcomes.

By understanding the failure modes we can build threat models and apply them to our ML-based applications and services, and then respond to those threats and defend our new applications.

The paper suggests 11 different attack classifications, many of which get around our standard defence models. It's possible to compromise a machine-learning system without needing access to the underlying software and hardware, so standard authorisation techniques can't protect ML-based systems and we need to consider alternative approaches.

What are these attacks? The first, perturbation attacks, modify queries to change the response to one the attackers desire. That's matched by poisoning attacks, which achieve the same result by contaminating the training data. Machine-learning models often include important intellectual property, and some attacks like model inversion aim to extract that data. Similarly, a membership inference attack will try to determine whether specific data was in the initial training set. Closely related is the concept of model stealing, using queries to extract the model.

SEE:5G: What it means for IoT(free PDF)

Other attacks include reprogramming the system around the ML model, so that either results or inputs are changed. Closely related are adversarial attacks that change physical objects, adding duct tape to signs to confuse navigation or using specially printed bandanas to disrupt facial-recognition systems. Some attacks depend on the provider: a malicious provider can extract training data from customer systems. They can add backdoors to systems, or compromise models as they're downloaded.

While many of these attacks are new and targeted specifically at machine-learning systems, they are still computer systems and applications, and are vulnerable to existing exploits and techniques, allowing attackers to use familiar approaches to disrupt ML applications.

It's a long list of attack types, but understanding what's possible allows us to think about the threats our applications face. More importantly they provide an opportunity to think about defences and how we protect machine-learning systems: building better, more secure training sets, locking down ML platforms, and controlling access to inputs and outputs, working with trusted applications and services.

Attacks are not the only risk: we must be aware of unintended failures -- problems that come from the algorithms we use or from how we've designed and tested our ML systems. We need to understand how reinforcement learning systems behave, how systems respond in different environments, if there are natural adversarial effects, or how changing inputs can change results.

If we're to defend machine-learning applications, we need to ensure that they have been tested as fully as possible, in as many conditions as possible. The apocryphal stories of early machine-learning systems that identified trees instead of tanks, because all the training images were of tanks under trees, are a sign that these aren't new problems, and that we need to be careful about how we train, test, and deploy machine learning. We can only defend against intentional attacks if we know that we've protected ourselves and our systems from mistakes we've made. The old adage "test, test, and test again" is key to building secure and safe machine learning -- even when we're using pre-built models and service APIs.

Be your company's Microsoft insider by reading these Windows and Office tips, tricks, and cheat sheets. Delivered Mondays and Wednesdays

Here is the original post:
Microsoft: This is how to protect your machine-learning applications - TechRepublic

Advances in computer power, machine-learning and predictive algorithms are creating paradigm shifts in many industries. For example, when analgorithm outperformed six radiologistsin reading mammograms and accurately diagnosing breast cancer, this raised questions around the role of machine-learning in medicine and whether it will replace, or enhance, the work being done by doctors.

Similarly, when Googles AI software AlphaGo beat the worlds top Go master in what is described as humankinds most complicated board game, The New York Timesdeclared it isnt looking good for humanity when an algorithm can outperform a human in a highly complex task.

Both these examples point to narrow uses of artificial intelligence, specific types of machine-learning that are hugely effective. The medical example illustrates supervised learning, where a computer is programmed to solve a particular problem by looking for patterns. It is given labelled data sets, in this case X-rays with the diagnosis of presence or absence of breast cancer. When given a new X-ray, the computer applies an algorithm based on what it has learnt from all the previous X-rays to make a diagnosis. Unsupervised learning is a sort of self-optimisation where a computer has a set of rules, such as how to play Go, and through playing millions of games learns how to apply these rules and improve.

What is machine-learning?

Machine-learning is a phenomenal tool. To fully harness its potential it is essential to understand what machine-learning is (and isnt) and to demystify some of the hype and the fear around what it can and cant be used for. We have anthropomorphised computers; we speak about them in terms of intelligence and learning. But in essence, a machine computes it does not learn. Its algorithms are designed to mimic learning. In essence, these algorithms minimise the errors of a complicated function that maps inputs to outcomes and we interpret that as solving a problem, but the machine doesnt know what problem it is solving or that it is playing a game. The intelligence rests with the humans who design the algorithms and configure them for specific tasks.

Now, more than ever, we need intelligent and well-educated people who can apply these techniques in the correct context and interpret the results. When an algorithm fails, the consequences can be catastrophic. An obvious example is a fatal accident caused by aself-driving car. We need to build in fault tolerance. Data integrity is also an important issue what we put in is going to affect what we get out. Education is critical in making sure we get these elements right. And, of course, there are broader ethical issues to consider surrounding data collection, such as what data can be used, where it is sourced, and whether different data sets can be combined.

Machine-learning is particularly valuable in the financial sector. Many applications are already in use in banking, insurance and asset management. Financial institutions use pattern recognition successfully for fraud detection. It is also valuable for looking at trends in data sets and finding patterns that humans may not be able to identify directly, for example in profiling people who apply for credit. There are even robo-advisory applications for individual asset allocation. In financial modelling, machine-learning can be applied to pricing, calibration and hedging.

For example, valuing derivatives contracts depends on many complex factors and variables such as interest rates, exchange rates, equity values all of which fluctuate all the time. Financial mathematicians use models for this, but they are complicated and not easy to solve in a closed form. We may be able to build and apply a model to one contract, but banks have hundreds of contracts, and risk management and regulatory frameworks need to be updated all the time. Machine-learning, specifically deep learning and neural nets, provides a powerful shortcut. We can use classical numerical methods to produce financial models and then use them as labelled data sets as in the X-ray example. An algorithm can take this input to generate the output for multiple contracts.

Industries and organisations that are pulling ahead are figuring out where to replace standard methods and complex, time-consuming computations with machine-learning. They are also using it for more complex modelling approaches, adding further variables that cannot usually be factored into standard methodologies. The most obvious benefit is that it is faster machines can compute millions of times faster than humans. These techniques also have the potential to be far more accurate and allow us to make better-informed decisions.

But the human element is critical. The accuracy of potentially life-changing outcomes will depend on how we identify where we use these techniques, how we build the algorithms, how we choose and manage data and, finally, in how we interpret and act upon the results.

Prof McWalter is an applied mathematician who lectures computational finance at UCTs African Institute of Financial Markets and Risk Management. Prof Kienitz lectures at the University of Wuppertal and is an adjunct associate professor at UCT. His research interests include numerical methods in finance and machine-learning applied to financial problems and derivative instruments.

Read this article:
Machine-learning is a boon, but it still needs a human hand - Business Day

The world has seen an unprecedented challenge and is battling this invisible enemy with all their might. The Novel coronavirus spread has left the global economies holding on to strands, businesses impacted and most people locked down. But while the physical world has come to a drastic halt or slow-down, the digital world is blooming. And in addition to understanding the possibilities of home workspaces, companies are finally understanding the scope of Machine Learning and Artificial Intelligence. A trend that was already gardening all the attention in recent years, ML & AI have particularly taken the centre-stage as more and more brands realise the possibilities of these tools. According to a research report released in February, demand for data engineers was up 50% and demand for data scientists was up 32% in 2019 compared to the prior year. Not only is machine learning being used by researchers to tackle this global pandemic, but it is also being seen as an essential tool in building a world post-COVID.

This pandemic is being fought on the basis of numbers and data. This is the key reason that has driven peoples interest in Machine Learning. It helps us in collecting, analysing and understanding a vast quantity of data. Combined with the power of Artificial Intelligence, Machine Learning has the power to help with an early understanding of problems and quick resolutions. In recent times, ML & AI are being used by doctors and medical personnel to track the virus, identify potential patients and even analyse the possible cure available. Even in the current economic crisis, jobs in data science and machine learning have been least affected. All these factors indicate that machine learning and artificial intelligence are here to stay. And this is the key reason that data science is an area you can particularly focus on, in this lockdown.

The capabilities of Machine Learning and Data Sciences One of the key reasons that a number of people have been able to shift to working from home without much hassle has to be the use of ML & AI by businesses. This shift has also motivated many businesses, both small-scale and large-scale, to re-evaluate their functioning. With companies already announcing plans to look at a more robust working mechanism, which involves less office space and more detailed and structured online working systems, the focus on Machine Learning is bound to increase considerably.

The Current PossibilitiesThe world of data science has been coming out stronger during this lockdown and the interest and importance given to the subject are on the rise. AI-powered mechanics and operations have already made it easier to manage various spaces with lower risks and this trend of turning to AI is bound to increase in the coming years. This is the reason that being educated in this field can improve your skills in this segment. If you are someone who has always been intrigued by data sciences and machine learning or are already working in this field and are looking for ways to accelerate your career, there are various courses that you can turn to. With the increased free time that staying at home has facilitated us with, beginning an additional degree to pad up your resume and also learn some cutting-edge concepts while gaining access to industry experts.

Start learning more about Machine Learning & AIIf you are wondering where to begin this journey of learning, a leading online education service provider, upGrad, has curated programs that would suit you! From Data Sciences to in-depth learnings in AI, there are multiple programs on their website that covers various domains. The PG Diploma in Machine Learning and AI, in particular, has a brilliant curriculum that will help you progress in the field of Machine Learning and Artificial Intelligence. A carefully crafted program from IIIT Bangalore which offers 450+ hours of learning with more than 10 practical hands-on capstone projects, this program has been designed to help people get a deeper understanding of the real-life problems in the field.

Understanding the PG Diploma in Machine Learning & AIThis 1-year program at upGrad has been articulated especially for working professionals who are looking for a career push. The curriculum consists of 30+ Case Studies and Assignments and 25+ Industry Mentorship Sessions, which help you to understand everything you need to know about this field. This program has the perfect balance between the practical exposure required to instil better management and problem-solving skills as well as the theoretical knowledge that will sharpen your skills in this category. The program also gives learners an IIIT Bangalore Alumni Status and Job Placement Assistance with Top Firms on successful completion.

Read the original here:
Rise in the demand for Machine Learning & AI skills in the post-COVID world - Times of India

In keeping with its tradition of high excellence and achievements, Greenwood High International School's student Tarun Paparaju of Class 12 has achieved the 'Grand Master' level in kernels, the highest accreditation in Kaggle, holding a rank of 13 out of 118,024 Kagglers worldwide. Kaggle is the world's largest online community for Data Science and Artificial Intelligence.

There are only 20 Kernel Grandmasters out of the three million users on Kaggle worldwide, and Tarun, aged 17 years, is honored to be one of the 20 Kernel Grandmasters now. Users of Kaggle are placed at different levels based on the quality and accuracy of their solutions to real-world artificial intelligence problems. The five levels in ascending order are Novice, Contributor, Expert, Master, and Grandmaster.

Kaggle hosts several data science competitions and contestants are challenged to find solutions to these real-world challenges. Kernels are a medium through which Kagglers share their code and insights on how to solve the problem.

These kernels include in-depth data analysis, visualisation, and machine learning, usually written in Python or R programming language. Other Kagglers can up vote a kernel if they believe it provides useful insights or solves the problem. 'Kernels Grandmaster' title at Kaggle requires 15 kernels qualified with gold medals.

Tarun's passion for Calculus, Mathematical modeling, and Data science from a very young age got him interested in participating and contributing to the Kaggle community.

He loves solving real-world Data Science problems, especially in the areas based on Deep learning: Natural language processing, Signal processing. Tarun is an open-source contributor to Keras, a Deep learning framework.

He has proposed and added Capsule NN layer support to Keras framework. He writes blogs about his adventures and learnings in data science.

Now, he closely works with the Kaggle community and aspires to be a scholar in the area of Natural language processing. Additionally, he loves playing cricket and football. Sports is a large part of his life outside Data science and academics.

Read:UGC releases new academic calendar: Here are top 10 important UGC updates

Read: MPhil, PhD students to get extension of 6 months, viva-voce through video conferencing: UGC

Read: WBBSE Madhyamik Result 2020: WB Class 10 result date to be fixed after Covid-19 lockdown ends

Visit link:
This 17-year-old boy created a machine learning model to suggest potential drugs for Covid-19 - India Today

ORBIS RESEARCH has recently announced Global Machine Learning in Medicine Market report with all the critical analysis on current state of industry, demand for product, environment for investment and existing competition. Global Machine Learning in Medicine Market report is a focused study on various market affecting factors and comprehensive survey of industry covering major aspects like product types, various applications, top regions, growth analysis, market potential, challenges for investor, opportunity assessments, major drivers and key players

This report is directed to arm report readers with conclusive judgment on the potential of mentioned factors that propel relentless growth in Global Machine Learning in Medicine Market. The report on Machine Learning in Medicine Market makes concrete headway in identifying and deciphering each of the market dimensions to evaluate logical derivatives which have the potential to set the growth course in Global Machine Learning in Medicine Market. Besides presenting notable insights on Machine Learning in Medicine Market factors comprising above determinants, the report further in its subsequent sections of this detailed research report on Machine Learning in Medicine Market states information on regional segmentation, as well as thoughtful perspectives on specific understanding comprising region specific developments as well as leading market players objectives to trigger maximum revenue generation and profits. This high-end research comprehension on Machine Learning in Medicine Market renders major impetus on detailed growth.

This study covers following key players:Monday.com

Request a sample of this report @ https://www.orbisresearch.com/contacts/request-sample/4328851

The report is directed to arm report readers with conclusive judgment on the potential of mentioned factors that propel relentless growth in Global Machine Learning in Medicine Market. A thorough run down on essential elements such as drivers, threats, challenges, opportunities are discussed at length in this elaborate report on Machine Learning in Medicine Market and eventually analyzed to document logical conclusions. This Machine Learning in Medicine Market also harps on competitive landscape, accurately identifying opportunities as well as threats and challenges. This report specifically unearths notable conclusions and elaborates on innumerable factors and growth triggering decisions that make this Machine Learning in Medicine Market a highly remunerative one.

This meticulous research based analytical review on Machine Learning in Medicine Market is a high end expert handbook portraying crucial market relevant information and developments, encompassing a holistic record of growth promoting triggers encapsulating trends, factors, dynamics, challenges, and threats as well as barrier analysis that accurately direct and influence profit trajectory of Machine Learning in Medicine Market. This high-end research comprehension on Machine Learning in Medicine Market renders major impetus on detailed growth facets, in terms of product section, payment and transaction platforms, further incorporating service portfolio, applications, as well as a specific compilation on technological interventions that facilitate ideal growth potential in Global Machine Learning in Medicine Market.

Access Complete Report @ https://www.orbisresearch.com/reports/index/global-machine-learning-in-medicine-market-professional-survey-2019-by-manufacturers-regions-countries-types-and-applications-forecast-to-2024

Market segment by Type, the product can be split intoOn-premisesSoftware-as-a-Service (SaaS)Cloud Based

Market segment by Application, split intoAerospaceAutomotive industryBiotech and pharmaceuticalChemical industryConsumer productsAerospace

This high-end research comprehension on Machine Learning in Medicine Market renders major impetus on detailed growth facets, in terms of product section, payment and transaction platforms, further incorporating service portfolio, applications, as well as a specific compilation on technological interventions that facilitate ideal growth potential in Global Machine Learning in Medicine Market.

The report also incorporates ample understanding on numerous analytical practices such as SWOT and PESTEL analysis to source optimum profit resources in Machine Learning in Medicine Market. Other vital factors related to the Machine Learning in Medicine Market such as scope, growth potential, profitability, and structural break-down have been distinctively documented in this keyword report to leverage holistic market growth.

For Enquiry before buying report @ https://www.orbisresearch.com/contacts/enquiry-before-buying/4328851

Some Key TOC Points:1 Industry Overview of Machine Learning in Medicin2 Industry Chain Analysis of Machine Learning in Medicine3 Manufacturing Technology of Machine Learning in Medicine4 Major Manufacturers Analysis of Machine Learning in MedicineContinued

About Us:Orbis Research (orbisresearch.com) is a single point aid for all your market research requirements. We have vast database of reports from the leading publishers and authors across the globe. We specialize in delivering customized reports as per the requirements of our clients. We have complete information about our publishers and hence are sure about the accuracy of the industries and verticals of their specialization. This helps our clients to map their needs and we produce the perfect required market research study for our clients.

Contact Us:Hector CostelloSenior Manager Client Engagements4144N Central Expressway,Suite 600, Dallas,Texas 75204, U.S.A.Phone No.: USA: +1 (972)-362-8199 | IND: +91 895 659 5155

See more here:
Machine Learning in Medicine Market 2020-2024 Review and Outlook - Latest Herald

Monitors display a video showing facial recognition software in use at the headquarters of the artificial intelligence company Megvii, in Beijing, May 10, 2018. Beijing is putting billions of dollars behind facial recognition and other technologies to track and control its citizens.

Gilles Sabri | The New York Times

The world is facing its biggest health crisis in decades but one of the world's most promising technologies artificial intelligence (AI) isn't playing the major role some may have hoped for.

Renowned AI labs at the likes of DeepMind, OpenAI, Facebook AI Research, and Microsoft have remained relatively quiet as the coronavirus has spread around the world.

"It's fascinating how quiet it is," said Neil Lawrence, the former director of machine learning at Amazon Cambridge.

"This (pandemic) is showing what bulls--t most AI hype is. It's great and it will be useful one day but it's not surprising in a pandemic that we fall back on tried and tested techniques."

Those techniques include good, old-fashioned statistical techniques and mathematical models. The latter is used to create epidemiological models, which predict how a disease will spread through a population. Right now, these are far more useful than fields of AI like reinforcement learning and natural-language processing.

Of course, there are a few useful AI projects happening here and there.

In March, DeepMind announced that it hadused a machine-learning technique called "free modelling" to detail the structures of six proteins associated with SARS-CoV-2, the coronavirus that causes the Covid-19 disease.Elsewhere, Israeli start-up Aidoc is using AI imaging to flag abnormalities in the lungs and a U.K. start-up founded by Viagra co-inventor David Brown is using AI to look for Covid-19 drug treatments.

Verena Rieser, a computer science professor at Heriot-Watt University, pointed out that autonomous robots can be used to help disinfect hospitals and AI tutors can support parents with the burden of home schooling. She also said "AI companions" can help with self isolation, especially for the elderly.

"At the periphery you can imagine it doing some stuff with CCTV," said Lawrence, adding that cameras could be used to collect data on what percentage of people are wearing masks.

Separately, a facial recognition system built by U.K. firm SCC has also been adapted to spot coronavirus sufferers instead of terrorists.In Oxford, England, Exscientia is screening more than 15,000 drugs to see how effective they are as coronavirus treatments. The work is being done in partnership withDiamond Light Source, the U.K.'s national "synchotron."

But AI's role in this pandemic is likely to be more nuanced than some may have anticipated. AI isn't about to get us out of the woods any time soon.

"It's kind of indicating how hyped AI was," said Lawrence, who is now a professor of machine learning at the University of Cambridge. "The maturity of techniques is equivalent to the noughties internet."

AI researchers rely on vast amounts of nicely labeled data to train their algorithms, but right now there isn't enough reliable coronavirus data to do that.

"AI learns from large amounts of data which has been manually labeled a time consuming and expensive task," said Catherine Breslin, a machine learning consultant who used to work on Amazon Alexa.

"It also takes a lot of time to build, test and deploy AI in the real world. When the world changes, as it has done, the challenges with AI are going to be collecting enough data to learn from, and being able to build and deploy the technology quickly enough to have an impact."

Breslin agrees that AI technologies have a role to play. "However, they won't be a silver bullet," she said, adding that while they might not directly bring an end to the virus, they can make people's lives easier and more fun while they're in lockdown.

The AI community is thinking long and hard about how it can make itself more useful.

Last week, Facebook AI announced a number of partnerships with academics across the U.S.

Meanwhile, DeepMind's polymath leader Demis Hassabis is helping the Royal Society, the world's oldest independent scientific academy, on a new multidisciplinary project called DELVE (Data Evaluation and Learning for Viral Epidemics). Lawrence is also contributing.

See original here:
A.I. can't solve this: The coronavirus could be highlighting just how overhyped the industry is - CNBC