Automation offers substantive benefits as companies look for ways to manage evolving workforces and workplace expectations. More than half of U.S. businesses now plan to increase their automation investment to help increase their agility and improve their ability to handle changing conditions quickly, according to Robotics and Automation News.

Businesses also need to be able to solve problems at scale, something that organizations are increasingly turning to machine learning to do. By creating algorithms that learn over time, its possible for companies to streamline decision-making with data-driven predictions. But creating the models can be complex and time-consuming, putting an added strain on businesses that may be low on resources.

Automated machine learning combines these two technologies to tap the best of both worlds, allowing companies to gain actionable insights while reducing total complexity. Once implemented, AutoML can help businesses gather and analyze data, respond to it quickly and better manage resources.

WATCH: Find out how organizations can empower digital transformation and secure remote work.

AutoML goes a step further than classic machine learning, says Earnest Collins, managing member of Regulatory Compliance and Examination Consultants and a member of the ISACA Emerging Technologies Advisory Group.

AutoML goes beyond creating machine learning architecture models, says Collins. It can automate many aspects of machine learning workflow, which include data preprocessing, feature engineering, model selection, architecture search and model deployment.

AutoML deployments can also be categorized by the format of training data used. Collins points to examples such as independent, identically distributed (IID) tabular data, raw text or image data, and notes that some AutoML solutions can handle multiple data types and algorithms.

There is no single algorithm that performs best on all data sets, he says.

Leveraging AutoML solutions offers multiple benefits that go beyond traditional machine learning or automation. The first is speed, according to Collins.

AutoML allows data scientists to build a machine learning model with a high degree of automation more quickly and conduct hyperparameter search over different types of algorithms, which can otherwise be time-consuming and repetitive, he says. By automating key processes from raw data set capture to eventual analysis and learningteams can reduce the amount of time required to create functional models.

Another benefit is scalability. While machine learning models cant compete with the in-depth nature of human cognition, evolving technology makes it possible to create effective analogs of specific human learning processes. Introducing automation, meanwhile, helps apply this process at scale in turn, enabling data scientists, engineers and DevOps teams to focus on business problems instead of iterative tasks, Collins says.

A third major benefit is simplicity, according to Collins. AutoML is a tool that assists in automating the process of applying machine learning to real-world problems, he says.

By reducing the complexity that comes with building, testing and deploying entirely new ML frameworks, AutoML streamlines the processes required to solve line-of-business challenges.

For machine learning solutions to deliver business value, ML models must be optimized based on current conditions and desired outputs. Doing so requires the use of hyperparameters, which Collins defines as adjustable parameters that govern the training of ML models.

Optimal ML model performance depends on the hyperparameter configuration value selection; this can be a time-consuming, manual process, which is where AutoML can come into play, Collins adds.

By using AutoML platforms to automate key hyperparameter selection and balancing including learning rate, batch size and drop rate its possible to reduce the amount of time and effort required to get ML algorithms up and running.

While AutoML isnt new, evolution across machine learning and artificial intelligence markets is now driving a second generation of automated machine learning platforms, according to RTInsights. The first wave of AutoML focused on building and validating models, but the second iterations include key features such as data preparation and feature engineering to accelerate data science efforts.

But this market remains both fragmented and complex, according to Forbes, because of a lack of established standards and expectations in the data science and machine learning (DSML) industry. Businesses can go with an established provider, such as Microsoft Azure Databricks, or they can opt for more up-and-coming solutions such as Google Cloud AutoML.

There are more tools around the corner. According to Synced, Google researchers are now developing AutoML-Zero, which is capable of searching for applicable ML algorithms within a defined space to reduce the need to create them from scratch. The search giant is also applying its AutoML to unique use cases; for example, the companys new Fabricius tool which leverages Googles AutoML vision toolset is designed to decode ancient Egyptian hieroglyphics.

Technological advancements combined with shifting staff priorities are somewhat driving robotic replacements. According to Time, companies are replacing humans wherever possible to reduce risk and improve operational output. But that wont necessarily apply to data scientists as AutoML rises, according to Collins.

The skills of professional, well-trained data scientists will be essential to interpreting data and making recommendations for how information should be used, he says. AutoML will be a key tool for improving their productivity, and the citizen data scientist, with no training in the field, would not be able to do machine learning without AutoML.

In other words, while AutoML platforms provide business benefits, recognizing the full extent of automated advantages will always require human expertise.

See the original post here:
What is AutoML and Why Should Your Business Consider It - BizTech Magazine

Machine learning and artificial intelligence are helping automate an ever-increasing array of tasks, with ever-increasing accuracy. They are supported by the growing volume of data used to feed them, and the growing sophistication in algorithms.

The flip side of more complex algorithms, however, is less interpretability. In many cases, the ability to retrace and explain outcomes reached by machine learning models (ML) is crucial, as:

"Trust models based on responsible authorities are being replaced by algorithmic trust models to ensure privacy and security of data, source of assets and identity of individuals and things. Algorithmic trust helps to ensure that organizations will not be exposed to the risk and costs of losing the trust of their customers, employees and partners. Emerging technologies tied to algorithmic trust include secure access service edge, differential privacy, authenticated provenance, bring your own identity, responsible AI and explainable AI."

The above quote is taken from Gartner's newly released 2020 Hype Cycle for Emerging Technologies. In it, explainable AI is placed at the peak of inflated expectations. In other words, we have reached peak hype for explainable AI. To put that into perspective, a recap may be useful.

As experts such as Gary Marcus point out, AI is probably not what you think it is. Many people today conflate AI with machine learning. While machine learning has made strides in recent years, it's not the only type of AI we have. Rule-based, symbolic AI has been around for years, and it has always been explainable.

Incidentally, that kind of AI, in the form of "Ontologies and Graphs" is also included in the same Gartner Hype Cycle, albeit in a different phase -- the trough of disillusionment. Incidentally, again, that's conflating.Ontologies are part of AI, while graphs, not necessarily.

That said: If you are interested in getting a better understanding of the state of the art in explainable AI machine learning, reading Christoph Molnar's book is a good place to start. Molnar is a data scientist and Ph.D. candidate in interpretable machine learning. Molnar has written the bookInterpretable Machine Learning: A Guide for Making Black Box Models Explainable, in which he elaborates on the issue and examines methods for achieving explainability.

Gartner's Hype Cycle for Emerging Technologies, 2020. Explainable AI, meaning interpretable machine learning, is at the peak of inflated expectations. Ontologies, a part of symbolic AI which is explainable, is in the trough of disillusionment

Recently, Molnar and a group of researchers attempted to addresses ML practitioners by raising awareness of pitfalls and pointing out solutions for correct model interpretation, as well as ML researchers by discussing open issues for further research. Their work was published as a research paper, titledPitfalls to Avoid when Interpreting Machine Learning Models, by the ICML 2020 Workshop XXAI: Extending Explainable AI Beyond Deep Models and Classifiers.

Similar to Molnar's book, the paper is thorough. Admittedly, however, it's also more involved. Yet, Molnar has striven to make it more approachable by means of visualization, using what he dubs "poorly drawn comics" to highlight each pitfall. As with Molnar's book on interpretable machine learning, we summarize findings here, while encouraging readers to dive in for themselves.

The paper mainly focuses on the pitfalls of global interpretation techniques when the full functional relationship underlying the data is to be analyzed. Discussion of "local" interpretation methods, where individual predictions are to be explained, is out of scope. For a reference on global vs. local interpretations, you can refer to Molnar's book as previously covered on ZDNet.

Authors note that ML models usually contain non-linear effects and higher-order interactions. As interpretations are based on simplifying assumptions, the associated conclusions are only valid if we have checked that the assumptions underlying our simplifications are not substantially violated.

In classical statistics this process is called "model diagnostics," and the research claims that a similar process is necessary for interpretable ML (IML) based techniques. The research identifies and describes pitfalls to avoid when interpreting ML models, reviews (partial) solutions for practitioners, and discusses open issues that require further research.

Under- or overfitting models will result in misleading interpretations regarding true feature effects and importance scores, as the model does not match the underlying data generating process well. Evaluation of training data should not be used for ML models due to the danger of overfitting. We have to resort to out-of-sample validation such as cross-validation procedures.

Formally, IML methods are designed to interpret the model instead of drawing inferences about the data generating process. In practice, however, the latter is the goal of the analysis, not the former. If a model approximates the data generating process well enough, its interpretation should reveal insights into the underlying process. Interpretations can only be as good as their underlying models. It is crucial to properly evaluate models using training and test splits -- ideally using a resampling scheme.

Flexible models should be part of the model selection process so that the true data-generating function is more likely to be discovered. This is important, as the Bayes error for most practical situations is unknown, and we cannot make absolute statements about whether a model already fits the data optimally.

Using opaque, complex ML models when an interpretable model would have been sufficient (i.e., having similar performance) is considered a common mistake. Starting with simple, interpretable models and gradually increasing complexity in a controlled, step-wise manner, where predictive performance is carefully measured and compared is recommended.

Measures of model complexity allow us to quantify the trade-off between complexity and performance and to automatically optimize for multiple objectives beyond performance. Some steps toward quantifying model complexity have been made. However, further research is required as there is no single perfect definition of interpretability but rather multiple, depending on the context.

This pitfall is further analyzed in three sub-categories: Interpretation with extrapolation, confusing correlation with dependence, and misunderstanding conditional interpretation.

Interpretation with Extrapolation refers to producing artificial data points that are used for model predictions with perturbations. These are aggregated to produce global interpretations. But if features are dependent, perturbation approaches produce unrealistic data points. In addition, even if features are independent, using an equidistant grid can produce unrealistic values for the feature of interest. Both issues can result in misleading interpretations.

Before applying interpretation methods, practitioners should check for dependencies between features in the data (e.g., via descriptive statistics or measures of dependence). When it is unavoidable to include dependent features in the model, which is usually the case in ML scenarios, additional information regarding the strength and shape of the dependence structure should be provided.

Confusing correlation with dependence is a typical error. The Pearson correlation coefficient (PCC) is a measure used to track dependency among ML features. But features with PCC close to zero can still be dependent and cause misleading model interpretations. While independence between two features implies that the PCC is zero, the converse is generally false.

Any type of dependence between features can have a strong impact on the interpretation of the results of IML methods. Thus, knowledge about (possibly non-linear) dependencies between features is crucial. Low-dimensional data can be visualized to detect dependence. For high-dimensional data, several other measures of dependence in addition to PCC can be used.

Misunderstanding conditional interpretation. Conditional variants to estimate feature effects and importance scores require a different interpretation. While conditional variants for feature effects avoid model extrapolations, these methods answer a different question. Interpretation methods that perturb features independently of others also yield an unconditional interpretation.

Conditional variants do not replace values independently of other features, but in such a way that they conform to the conditional distribution. This changes the interpretation as the effects of all dependent features become entangled. The safest option would be to remove dependent features, but this is usually infeasible in practice.

When features are highly dependent and conditional effects and importance scores are used, the practitioner has to be aware of the distinct interpretation. Currently, no approach allows us to simultaneously avoid model extrapolations and to allow a conditional interpretation of effects and importance scores for dependent features.

Global interpretation methods can produce misleading interpretations when features interact. Many interpretation methods cannot separate interactions from main effects. Most methods that identify and visualize interactions are not able to identify higher-order interactions and interactions of dependent features.

There are some methods to deal with this, but further research is still warranted. Furthermore, solutions lack in automatic detection and ranking of all interactions of a model as well as specifying the type of modeled interaction.

Due to the variance in the estimation process, interpretations of ML models can become misleading. When sampling techniques are used to approximate expected values, estimates vary, depending on the data used for the estimation. Furthermore, the obtained ML model is also a random variable, as it is generated on randomly sampled data and the inducing algorithm might contain stochastic components as well.

Hence, themodel variance has to be taken into account. The true effect of a feature may be flat, but purely by chance, especially on smaller data, an effect might algorithmically be detected. This effect could cancel out once averaged over multiple model fits. The researchers note the uncertainty in feature effect methods has not been studied in detail.

It's a steep fall to the peak of inflated expectations to the trough of disillusionment. Getting things done for interpretable machine learning takes expertise and concerted effort.

Simultaneously testing the importance of multiple features will result in false-positive interpretations if the multiple comparisons problem (MCP) is ignored. MCP is well known in significance tests for linear models and similarly exists in testing for feature importance in ML.

For example, when simultaneously testing the importance of 50 features, even if all features are unimportant, the probability of observing that at least one feature is significantly important is 0.923. Multiple comparisons will even be more problematic, the higher dimensional a dataset is. Since MCP is well known in statistics, the authors refer practitioners to existing overviews and discussions of alternative adjustment methods.

Practitioners are often interested in causal insights into the underlying data-generating mechanisms, which IML methods, in general, do not provide. Common causal questions include the identification of causes and effects, predicting the effects of interventions, and answering counterfactual questions. In the search for answers, researchers can be tempted to interpret the result of IML methods from a causal perspective.

However, a causal interpretation of predictive models is often not possible. Standard supervised ML models are not designed to model causal relationships but to merely exploit associations. A model may, therefore, rely on the causes and effects of the target variable as well as on variables that help to reconstruct unobserved influences.

Consequently, the question of whether a variable is relevant to a predictive model does not directly indicate whether a variable is a cause, an effect, or does not stand in any causal relation to the target variable.

As the researchers note, the challenge of causal discovery and inference remains an open key issue in the field of machine learning. Careful research is required to make explicit under which assumptions what insight about the underlying data generating mechanism can be gained by interpreting a machine learning model

Molnar et. al. offer an involved review of the pitfalls of global model-agnostic interpretation techniques for ML. Although as they note their list is far from complete, they cover common ones that pose a particularly high risk.

They aim to encourage a more cautious approach when interpreting ML models in practice, to point practitioners to already (partially) available solutions, and to stimulate further research.

Contrasting this highly involved and detailed groundwork to high-level hype and trends on explainable AI may be instructive.

Read more:
Explainable AI: From the peak of inflated expectations to the pitfalls of interpreting machine learning models - ZDNet

In a recent development, Amazon announced that it will make online courses by its Machine Learning University available to the public. The classes were previously only available to Amazon employees.

The company believes that machine learning has the potential to transform businesses in all industries, but theres a major limitation: demand for individuals with ML expertise far outweighs supply. Thats a challenge for Amazon, and for companies big and small across the globe.

The Machine Learning University (MLU) was founded with an aim to meet this demand in 2016. It helped ML practitioners sharpen their skills and keep them abreast with the latest developments in the field. The classes are taught by Amazon ML experts.

The tech giant now plans to make these classes available to the ML community across the globe. It will include nine more in-depth courses before the year ends. As the blog post notes, by the beginning of 2021, all MLU classes will be available via on-demand video, along with associated coding materials. It will cover topics such as natural language processing, computer vision and tabular data while addressing various business problems.

By going public with the classes, we are contributing to the scientific community on the topic of machine learning, and making machine learning more democratic, said Brent Werness, AWS research scientist and MLUs academic director.

This initiative to bring our courseware online represents a step toward lowering barriers for software developers, students and other builders who want to get started with practical machine learning, he added.

Instead of a three-class sequence that takes upwards of 18 or 20 weeks to complete, in the accelerated classes we can engage students with machine learning right up front, shared Ben Starsky, MLU program manager.

The company said that similar to other open-source initiatives, MLUs courseware will evolve to improve over time based on input from the builder community. It also looking to rebuild its curriculum to further integrate dive into deep learning into class sessions.

The company wants to include as many important things as possible while offering flexibility in the way people can take these classes.

comments

Srishti currently works as Associate Editor at Analytics India Magazine. When not covering the analytics news, editing and writing articles, she could be found reading or capturing thoughts into pictures.

The rest is here:
Amazon's Machine Learning University To Make Its Online Courses Available To The Public - Analytics India Magazine

About the position

This is a researcher training position aimed at providing promising researcher recruits the opportunity of academic development in the form of a doctoral degree.

Bringing intelligence into Internet-of-Things systems is mostly constrained by the availability of energy. Devices need to be wireless and small in size to be economically feasible and need to replenish their energy buffers using energy harvesting. In addition, devices need to work autonomously, because it is unfeasible to operate them manually or change batteriesthere's simply too many of them. To make the best of the energy available, IoT devices should plan wisely how they spend their energy, that means, which tasks they should perform and when. This requires the development of policies. Due to the different situations, the various devices may find themselves in, it will also vary from device to device which policies are best, which suggests the use of machine learning for the autonomous, individual development of energy policies for IoT devices.

One special focus in this project is the modeling of the power supply of the IoT devices, that means, the submodule that combines energy harvesting and energy buffering. Both are processes that are highly stochastic and probabilistic and vary over time and with the age of the device yet have major impact on a devices ability to perform well. In addition, due to the constraints, the approach itself must be computationally feasible and not itself consume too much energy. Combining machine learning for power supplies with the application goals of the IoT device is therefore a research challenge.

You will reportto the Head of Department.

Duties of the position

Within this project, we will design and validate machine-learning approaches to model power supplies to know more about their current and future state, and energy budget policies that allow IoT devices to perform better and autonomously. The project is cross-disciplinary involving electronic design, software and statistical techniques and machine learning. Depending on the skills of the candidate, different aspect may be emphasized, for instance focusing on statistical modelling of relevant effects, transfer learning and model identification, and explainability of machine learning models. Experience with electronics may be beneficial but are not strictly required.

The research will be carried out in an interdisciplinary environment of several research groups, and under guidance of three supervisors,

The research environments include

Required selection criteria

The PhD-position's main objective is to qualify for work in research positions. The qualification requirement is that you have completed a masters degree or second degree (equivalent to 120 credits) with a strong academic background in computer science, statistical machine learning, applied mathematics, communication- and information technology, electrical engineering, electronic engineering, or an equivalent education with a grade of B or better in terms ofNTNUs grading scale. If you do not have letter grades from previous studies, you must have an equally good academic foundation. If you are unable to meet these criteria you may be considered only if you can document that you are particularly suitable for education leading to a PhD degree.

The appointment is to be made in accordance with the regulations in force concerningState Employees and Civil Servants and national guidelines for appointment as PhD, post doctor and research assistant.

Preferred selection criteria

Personal characteristics

In the evaluation of which candidate is best qualified, emphasis will be placed on education, experience and personal suitability, in terms of the qualification requirements specified in the advertisement.

We offer

Salary and condition

PhD candidate:

PhD candidates are remunerated in code 1017, and are normally remunerated at gross from NOK 479 600 per annum, depending on qualifications and seniority. From the salary, 2% is deducted as a contribution to the Norwegian Public Service Pension Fund.

The period of employment is 4 years including 25% of teaching assistance. Students at NTNU can also apply for this position as part of an integrated PhD program (https://www.ntnu.edu/iik/integrated-phd).

Appointment to a PhD position requires that you are admitted to the PhD programme in Information Security and Communication Technologywithin three months of employment, and that you participate in an organized PhD programme during the employment period.

The engagement is to be made in accordance with the regulations in force concerning State Employees and Civil Servants, and the acts relating to Control of the Export of Strategic Goods, Services and Technology. Candidates who by assessment of the application and attachment are seen to conflict with the criteria in the latter law will be prohibited from recruitment to NTNU. After the appointment you must assume that there may be changes in the area of work.

It is a prerequisite you can be present at and accessible to the institution daily.

About the application

The application and supporting documentation to be used as the basis for the assessment must be in English.

Publications and other scientific work must follow the application. Please note that applications are only evaluated based on the information available on the application deadline. You should ensure that your application shows clearly how your skills and experience meet the criteria which are set out above.

The application must contain:

Joint works will be considered. If it is difficult to identify your contribution to joint works, you must attach a brief description of your participation.

NTNU is committed to following evaluation criteria for research quality according toThe San Francisco Declaration on Research Assessment - DORA.

General information

Working at NTNU

A good work environment is characterized by diversity. We encourage qualified candidates to apply, regardless of their gender, functional capacity or cultural background.

The city of Trondheimis a modern European city with a rich cultural scene. Trondheim is the innovation capital of Norway with a population of 200,000. The Norwegian welfare state, including healthcare, schools, kindergartens and overall equality, is probably the best of its kind in the world. Professional subsidized day-care for children is easily available. Furthermore, Trondheim offers great opportunities for education (including international schools) and possibilities to enjoy nature, culture and family life and has low crime rates and clean air quality.

As an employeeatNTNU, you must at all times adhere to the changes that the development in the subject entails and the organizational changes that are adopted.

Information Act (Offentleglova), your name, age, position and municipality may be made public even if you have requested not to have your name entered on the list of applicants.

Questions about the position can be directed to Frank Alexander Kraemer, via kraemer@ntnu.no

Please submit your application electronically via jobbnorge.no with your CV, diplomas and certificates. Applications submitted elsewhere will not be considered. Diploma Supplement is required to attach for European Master Diplomas outside Norway.

Chinese applicants are required to provide confirmation of Master Diploma fromChina Credentials Verification (CHSI).

Pakistani applicants are required to provide information of Master Diploma from Higher Education Commission (HEC) https://hec.gov.pk/english/pages/home.aspx

Applicants with degrees from Cameroon, Canada, Ethiopia, Eritrea, Ghana, Nigeria, Philippines, Sudan, Uganda and USA have to send their education documents as paper copy directly from the university college/university, in addition to enclose a copy with the application.

Application deadline: 13.09.2020

NTNU - knowledge for a better world

The Norwegian University of Science and Technology (NTNU) creates knowledge for a better world and solutions that can change everyday life.

Department of Information Security and Communication Technology

Research is vital to the security of our society. We teach and conduct research in cyber security, information security, communications networks and networked services. Our areas of expertise include biometrics, cyber defence, cryptography, digital forensics, security in e-health and welfare technology, intelligent transportation systems and malware. The Department of Information Security and Communication Technology is one of seven departments in theFaculty of Information Technology and Electrical Engineering

Deadline13th September 2020EmployerNTNU - Norwegian University of Science and TechnologyMunicipalityTrondheimScopeFulltimeDurationTemporaryPlace of serviceNTNU Campus Glshaugen

Visit link:
PhD Research Fellowship in Machine Learning for Cognitive Power Management job with NORWEGIAN UNIVERSITY OF SCIENCE & TECHNOLOGY - NTNU | 219138 -...

Swami Sivasubramanian, Vice-President, Amazon Machine Learning, AWS (Amazon Web Services), who leads a global AI/ML team, has built more than 30 AWS services, authored around 40 referred scientific papers and been awarded over 200 patents. He was also one of the primary authors for a paper titled, Dynamo: Amazons Highly Available Key-value Store, along with AWS CTO and VP, Werner Vogels, which received the ACM Hall of Fame award. In a conversation with BusinessLine from Seattle, Swami said people always assume AI and ML are futuristic technologies, but the fact is AI and ML are already here and it is happening all around us.excerpts:

Bengaluru, August 12

The popular use cases for AI/ML are predominantly in logistics, customer experience and e-commerce. What AI/ML use cases are likely to emerge in the post-Covid-19 environment?

We dont have to wait for post-Covid-19, were seeing this right now. Artificial Intelligence (AI) and Machine Learning (ML) are playing a key role in better understanding and addressing the Covid-19 crisis. In the fight against Covid-19, organisations have been quick to apply their machine learning expertise in several areas, including, scaling customer communications, understanding how Covid-19 spreads, and speeding up research and treatment. Were seeing adoption of AI/ML across all industries, verticals and sizes of business. We expect this to not only continue, but accelerate in the future.

Of AWSs 175+ services portfolio, how many are AI/ML services?

We dont break out that number, but what I can tell you is AWS offers the broadest and deepest set of machine learning services and supporting cloud infrastructure putting machine learning in the hands of every developer, data scientist and expert practitioner.

Then why has AWS not featured in Gartners Data Science and ML Platforms Magic Quadrant?

Gartner's inclusion criteria explicitly excluded providers who focus primarily on developers. However, the Cloud AI Developer Services Magic Quadrant does cite us as a leader. Also, the recently released Gartner Solution Scorecard, which evaluated our capabilities in the Data Science and Machine Learning space, scored Amazon SageMaker higher than offerings from the other major providers.

Where is India positioned on the AI/ML adoption curve compared to developed economies?

I think, India is in a really good place. I remember visiting some of our customers and start-ups in India, there is so much innovation happening in India. I happen to believe that transformation comes because at a ground level, developers start adopting technologies and this is one of those things where I think India, especially at a ground level when it comes to the start-up ecosystem, have been jumping into in a big way to adopt machine learning technology.

For example, machine learning is embedded in every aspect of what Freshworks, a B2B unicorn in India, is doing. In fact, they build something like 33,000 models and they are iterating and theyre trying to build ML models, again using some of our technologies like Amazon SageMaker. Theyve cut down from eight weeks to less than one week. redBus, which Im a big fan of as I travel back and forth between Chennai to Bengaluru, is also using some of our ML technologies and their productivity has increased. One of the key things we need to be cognizant of is that machine learning technology is not going to get mainstream adoption if people are just using it for extremely leading-edge use cases. It should be used in everyday use cases. I think even in India now, it is starting to get into mainstream use cases in a big and meaningful way. For instance, Dish TV uses AWS Elemental, our video processing service to process video content and then they feed it into Amazon Rekognition to flag inappropriate content. There are start-ups like CreditVidya, who are building an ML platform on AWS to analyze behavioural data of customers and make better recommendations.

The greater the adoption of AI/ML, the more job losses are likely as organisations fire people to induct skilled talent. Please comment.

One thing is for sure, there is change coming and technology is driving it. Im very optimistic about the future. I remember the days where there used to be manual switching of telephones, but then we moved to automated switching. Its not like those jobs went away. All these people re-educated themselves and they are actually doing more interesting, more challenging jobs. Lifelong education is going to be critical. In Amazon, my team, for instance, runs Machine Learning University. We train our own engineers and Amazon Associates on various opportunities and expose them to leading-edge technology such as machine learning. Now, we are actually making this available for free as part of the AWS Training and Certification programs. In November 2018 we made it free, and within the first 48 hours of us making this free, we had more than one lakh people registered to learn. So, there is a huge appetite for it. In 2012, we decided, every organisation within Amazon had to have a machine learning strategy, even when machine learning was not even actually considered cool. So Jeff and the leadership team said, machine learning is going to be such a pivotal thing for every line of business irrespective of whether they run cloud computing or supply chain or financial technology data, and we required every business group in their yearly planning, to include how they were going to leverage machine learning in their business. And no, we do not plan to was not considered an acceptable answer.

What AI/ML tools do AWS offer, and for whom?

The vast majority of ML being done in the cloud today is on AWS. With an extensive portfolio of services at all three layers of the technology stack, more customers reference using AWS for machine learning than any other provider. AWS released more than 250 machine learning features and capabilities in 2019, with tens of thousands of customers using the services, spurred by the broad adoption of Amazon SageMaker since AWS re:Invent 2017. Our customers include, American Heart Association, Cathay Pacific, Dow Jones, Expedia.com, Formula 1, GE Healthcare, UKs National Health Service, NASA JPL, Slack, Tinder, Twilio, United Nations, the World Bank, Ryanair, and Samsung, among others.

Our AI/ML services are meant for: Advanced developers and scientists who are comfortable building, tuning, training, deploying, and managing models themselves, AWS offers P2 and P3 instances at the bottom of the stack which provide up to six times better performance than any other GPU instances available in the cloud today together with AWSs deep learning AMI (Amazon Machine Image) that embeds all the major frameworks. And, unlike other providers who try to funnel everybody into using only one framework, AWS supports all the major frameworks because different frameworks are better for different types of workloads.

At the middle layer of the stack, organisations that want to use machine learning in an expansive way can leverage Amazon SageMaker, a fully managed service that removes the heavy lifting, complexity, and guesswork from each step of the ML process, empowering everyday developers and scientists to successfully use ML. SageMaker is a sea-level change for everyday developers being able to access and build machine learning models. Its kind of incredible, in just a few months, how many thousands of developers started building machine learning models on top of AWS with SageMaker.

At the top layer of the stack, AWS provides solutions, such as Amazon Rekognition for deep-learning-based video and image analysis, Amazon Polly for translating text to speech, Amazon Lex for building conversations, Amazon Transcribe for converting speech to text, Amazon Translate for translating text between languages, and Amazon Comprehend for understanding relationships and finding insights within text. Along with this broad range of services and devices, customers are working alongside Amazons expert data scientists in the Amazon Machine Learning Solutions Lab to implement real-life use cases. We have a pretty giant investment in all layers of the machine learning stack and we believe that most companies, over time, will use multiple layers of that stack and have applications that are infused with ML.

Why would customers opt for AWSs AI/ML services versus competitor offerings from Microsoft, Google?

At Amazon, we always approach everything we do by focusing on our customers. We have thousands of engineers at Amazon committed to ML and deep learning, and its a big part of our heritage. Within AWS, weve been focused on bringing that knowledge and capability to our customers by putting ML into the hands of every developer and data scientist. But we do take a different approach to ML than others may we know that the only constant within the history of ML is change. Thats why we will always provide a great solution for all the frameworks and choices that people want to make by providing all of the major solutions so that developers have the right tool for the right job. And our customers are responding! Today, the vast majority of ML and deep learning in the cloud is running on AWS, with meaningfully more customer references for machine learning than any other provider. In fact, 85 per cent of TensorFlow being run in the cloud, is run on AWS.

See original here:
Machine learning is pivotal to every line of business, every organisation must have an ML strategy - BusinessLine