Tag

Machine Learning

High school student predicts seizures using Falkonry LRS

By | Industry Presentations | No Comments

At Oracle OpenWorld 2017, Falkonry CEO, Nikunj Mehta and Cupertino High School student, Ayush Jain, took the stage at the Innovation Studio to show how easily Falkonry LRS machine learning system can be used to find invisible patterns in data.

While Falkonry is typically used to solve hard operational problems in industrial environments like semiconductor fabs, automotive plants and chemical manufacturing, it can analyze any multivariate time series data and discover the invisible patterns that can provide reliable, early warnings of problems.

In this example, Ayush studied EEG data using Falkonry LRS automated machine learning to see if an early warning to an epileptic seizure could be detected. While Ayush was new to EEG data and to Falkonry, he had a very ambitious goal. Within 2 weeks, he had to create the prediction system for seizures. See the results from Ayush’s project in the video below.

This is a powerful example of how predictive analytics and automated machine learning can be used for positive change. It also shows the simplicity and ease-of-use of the Falkonry LRS system. Great work Ayush! We know you will go on to solving more hard problems and leave your mark on society, and are really glad to be a part of your journey.

If you’re interested in learning more, check out our Product page to watch a short demo of Falkonry LRS ready-to-use machine learning system.

OSIsoft Intern Creates System to Predict Energy Pricing

By | Industry Presentations | No Comments

OSIsoft customer support intern, Miwa Teranishi, leverages the power of OSIsoft PI System and Falkonry LRS ready-to-use machine learning system to create an energy price prediction system to understand electricity pricing in Kyushu, Japan. With her system, energy consumers could identify the best time to consume or avoid electricity use during a day. Let’s take a look at her project in this video:

As Miwa highlighted, Falkonry LRS “makes creating models quite simple with no programming and no knowledge of data science algorithms required.” Further, the integration between OSIsoft and Falkonry allows customers to quickly unlock value from existing data assets and deliver insights from their PI System.

Thanks Miwa for a very interesting use of PI System and Falkonry LRS. This is a creative example of how automated machine learning could be applied to everyday challenges, like the daily, fluctuating price of electricity.

In addition to PI System, Falkonry LRS machine learning is easily integrated into many popular solution platforms such as Splunk and Microsoft Azure IoT, accelerating the time to value for customers.

If you’re interested in learning more, check out our Product page to watch a short demo of Falkonry LRS ready-to-use machine learning system.

Ciner creates early-warning system for mining operations with Falkonry LRS

By | Industry Presentations

At the OSIsoft User Conference in San Francisco, joint customer, Ciner Resources (pronounced “jin-ner”) spoke about the digital transformation they were experiencing by employing OSIsoft’s PI system along with Falkonry’s condition recognition software.

The three Ciner speakers included their CIO, SMART plant lead and their process engineer.  They spoke about the challenges, triumphs and future goals of digitally transforming their business.  Ciner’s mine and processing plant in Wyoming has a ball mill, called a vertimill, that was experiencing unexpected downtime due to fluctuations in ore grade. Each time the vertimill clogged due to “unknown” low quality ore grade, it cost the company $30,000/hour of production potentially leading to a loss of $720,000 per day.

Ciner’s engineers and practitioners worked with OSIsoft and Falkonry to look into historical data to devise a solution. Together they employed the PI System and Falkonry’s machine learning software and within a few weeks, were able to recognize the patterns that led to blockage conditions. From there, an early-warning system was created to detect low-quality ore grade upstream of the vertimill in order to mitigate downtime events. Where the raw material took two hours from intake to reach the vertimill, Falkonry discovered 10 hour early warnings which surprised the process engineers.

In the video below, Wyatt Keller, Process Engineer at Ciner, shares his experience using OSIsoft and Falkonry LRS ready-to-use machine learning system.

If you’re interested in learning more about our success with Ciner, check out our customer video page to watch additional videos.

Falkonry is a Critical Component in the Next Wave of Technology

By | IT/OT Management

Falkonry discovers, recognizes, and predicts operating and performance conditions from time series data, using machine learning/pattern recognition. From its inception, the goal was to help industry become more productive using advanced signal processing techniques.

 

Falkonry logo

Falkonry builds on key technologies that are becoming prevalent in industry: #IoT/#IIOT (the Internet of Things); #ML (Machine Learning) and #AI (Artificial Intelligence).  Let’s look at each of these.

#IoT: Markets and Markets values the Internet of Things at $157.05 Billion in 2016, growing to $661.74 Billion by 2021, at a Compound Annual Growth Rate (CAGR) of 33.3% over the period. Growth is facilitated by a number of factors: the decreasing cost and proliferation of devices/sensors; the accessibility of low cost services, such as the #cloud and #big data processing, and the influx of new application providers, bringing choice into the market for users.  They highlight data management software is gaining high market traction with the help of cloud and predictive analytics to manage data generated from machineries. Falkonry, at its core, uses predictive analytics and pattern recognition to identify operations improvement opportunities, such as OEE (Overall Equipment Effectiveness) tuning.

#ML: Ironpaper.com defines machine learning as an area of computer science where machines learn patterns and recognize material without being explicitly programmed by a human. The cognitive computing market size is expected to surpass $12 billion by 2022, according to Grand View Research.  Falkrony uses Machine Learning to analyze data set features and provide condition analysis on a custom data set. It allows the automation of large-scale condition detection that previously required a massive human effort and cost to implement. Typical applications are Industrial and Transportation operations and maintenance, IT system anomaly detection, and others.

#AI: In a separate study, Markets and Markets forecasts the artificial intelligence (AI) market to be worth $16.06 Billion by 2022, growing at a CAGR of 62.9% from 2016 to 2022. Artificial intelligence is a consolidation of state-of-the-art technologies which are used to develop products which work similar to human intelligence.  AI is a broad category and includes hardware, software and services providers. Falkonry AI provides a cognitive/AI service that can automate the interpretation of telemetry and sensor data from industrial activity, IT processes, and high-end consumer assets. The Falkonry AI block allows you to analyze your massive realtime streams of data, and lets you build models or predict future action based on those data streams.

So, with Falkonry, the key technologies that are shaping the industrial market–#IoT, #ML, and #AI, are offered in one easy to use service. It allows companies that generate massive amounts of data, such as those in the process industries (oil and gas; metals, mining and minerals; pulp and paper) to detect, analyze, and understand the areas within a process that can be improved; where bottlenecks and choke points reside; and where in the process flow tuning can be applied.

For more information on Falkonry, please visit www.falkonry.com.

Falkonry Webinar on Control Global

By | IT/OT Management

Webinar: Using Patterns in Real-Time Data for Operational Excellence

Hosted by Control Global, Presented by Falkonry

May 17, 2017 at 2pm ET

Register Now
Manufacturing and utilities companies are challenged with increasing the performance of their assets amid an aging infrastructure, retiring workforce, and rapidly changing technology.  Falkonry has uncovered a new way of working with existing automation and control architectures to provide predictive insights using time-based pattern recognition, fed from real-time data.

Falkonry works with your staff, leveraging their resident knowledge, to provide context and meaning to the patterns. It ‘institutionalizes’ learning, using both artificial intelligence and machine learning, enabling continuous improvement for operations. This webinar will discuss how users are better able to predict downtime, maximize uptime, yield and quality using pattern recognition technology.

After viewing this webinar, visitors will take away:

  • What is pattern recognition?
  • How can pattern recognition help me in my every day operations?
  • How long will it take to see benefits?
  • How easy is it to start?
  • How have others used pattern recognition in their operations?

Meet our Speaker:

Nikunj-Mehta-square-e1487891568951.png

Nikunj Mehta
CEO & Founder
Falkonry

Nikunj Mehta
started Falkonry in 2012 under the belief that operational excellence comes from thorough and up-to-the-minute understanding of system conditions. Nikunj gained experience with operations and analytics challenges at the world’s largest and most sophisticated industrial organizations at C3 Energy as its Architect and VP of Customer Success.

Nikunj has a Ph.D. and M.S. in Computer Science from USC, and a B.E. in Computer Engineering from University of Mumbai. He has contributed to standards at both W3C and IETF.

Sponsored by:
Logo.jpg

Register Now!

Solution fit and flexibility: why Falkonry product teams invest time in redesigning their UI

By | IT/OT Management

Solution fit and flexibility: why Falkonry product teams invest time in redesigning their UI

A series on product improvements at Falkonry – Part 1

By Jeegar Shah – Sr. Director Products,  Falkonry

 

When designing a complex product that has all the bells and whistles under the hood, it is imperative that the user experience be all the more simplistic, streamlined and intuitive.

As any product evolves to encompass new ideas and optimizations, there is a tendency to display the prowess of your algorithmic horsepower to your users by giving them YAF (Yet Another Feature). This is all dandy for developers that have an exploding cache of Github repos and code checkins as they try their one upmanship on their colleagues with cumulative pull requests. But this is often a nightmare for customers who are asymptotically approaching their steady state of product usage and now have to deal with YAF – when they never really asked for one.

At Falkonry, our overarching objective is solution fit and flexibility. We believe that the promise of bringing OT (Operations Technology) efficiency lies in the ability to empower subject matter experts (SMEs) with the tools to make timely and effective decisions. A YAF may exactly help you achieve the opposite. Falkonry is committed to making AI-based pattern recognition an easy-to-incorporate component of any operations-oriented solution. With the addition of every algorithmic richness and IP, we enforce a mindset of looking back at our footsteps to remind us of the silhouette of an army of SME’s that will be following the same steps and pausing every so often to ask each other the question “Did we just throw a YAF?”

Read More

Visual Goodies in Falkonry’s Learning

By | IT/OT Management

The learning capabilities of Falkonry can be pretty addictive because they are so visually gratifying. Now to round out that Learning & Monitor interface, we added a few nifty features.

Improved Condition Summary

Increased visual capabilities

 

You can now quickly find out which elements are in a given condition(s) in a given period of time, so that you can navigate directly to those events. A dog-eared pull-out is available on all summary blocks so that you can see which conditions occur in that block of time and which events are in that condition, and for how much of the time. You can select the events you want to investigate further so that you can quickly assess the findings of a condition in a period of time.

This new feature will enable you to evaluate how conditions arise across your data set and what patterns are discovered by Falkonry in that data. When dealing with large numbers of things this feature will improve your ability to inspect Falkonry’s findings without losing context of the overall problem.

Improved Condition Buttons

This works in combination with the condition selections, which we have further simplified by reducing all unlabeled conditions to a circular buttons and user-named conditions to rounded corner rectangular buttons. Additionally, you only see buttons for conditions you are currently looking at. You can always bring in additional condition timelines from the selection panel by choosing old classifications or new verification, etc.

There are other usability improvements that provide you the information you are looking for, removing delays such as:

  • Increased the amount of vertical screen space available for viewing by removing the fixed navigation bar
  • Reduced the lag between creating a pipeline and being able to access the pipeline for learning
  • Added the ability to inspect your data just before starting the learning process
  • Auto-updating of the histories of flow segments and learning in the Configuration view

Analytics Performance Improvement

On top of all these visual improvements, the speed of the analytics was improved so that your model revisions complete faster and with fewer and more understandable errors.

These improvements are iterative steps we’ve taken to increase performance and interoperability. We’ll continue to work on increasing connections to a variety of data sources, providing greater manageability of pipelines, as well as improving the learning and monitoring interfaces. All of this is to improve reliability, performance and intelligence.

This post was originally written March 6, 2016

How Can Operations AI Help with Banking Risk Management?

By | IT/OT Management

At first blush, you might think the kind of Artificial Intelligence applications that are deployed in industrial contexts would have nothing to do with financial services. Whether we’re talking manufacturing, process control, automotive, or any other industrial setting, what could they possibly have in common with banking? The answer: a surprisingly large amount. And there are lessons learned from industrial applications that could help to shape the development of FinTech solutions, particularly those focused on risk management.

Traditional Banking Can Benefit from AI

 

This blog post results from a conversation between Nikunj Mehta, Founder and CEO of Falkonry, and Graham Seel, a 30 year banking veteran and Principal with BankTech Consulting. We think you’ll be surprised at the parallels and the promise.

Industrial Control and Banking Operations

What is similar between an engineer controlling an industrial facility and a bank operations manager controlling payment processing? Both deal with operational risks that require constant scrutiny and immediate action at the earliest sign of trouble.

But aside from this intuition that complex operations are the same everywhere, what really is alike between industrial operations and financial operations? We thought this blog post would highlight the similarities with an eye to techniques that can cross over from one to the other. The similarities include:

  1. Data-Informed Operations
  2. Alarm Fatigue
  3. Trends And Human Processing
  4. Expert instead of Learning Systems
  5. Regulatory Oversight
  6. Trade Secrets

Financial Institutions are primarily risk managers and manage a variety of financial risks — market, credit, operational, currency, liquidity, and others. All industrial companies are managing a variety of industrial risks — safety, quality, efficiency, asset, inventory, and others.

Here is our take on the similarities between financial and industrial operations that suggests that data-informed technologies being applied to one apply to the other and vice versa.

1. Data-Informed Operations

Data-informed operations is the basis for day to day operations. No matter how sophisticated the data collection and processing systems are, a trained human is ultimately responsible for making critical decisions. For industrial activities, there are closed-loop control systems that will automatically carry out safety and basic efficiency measures. Similarly, bank operations managers can carry out exception processing and error recovery based upon system-generated communications. Still, any analytics applications that are guiding their decisions will submit their findings to a human to effect any changes to operations. This is why all operations are run by substantial-sized teams who are challenged to keep improving their effectiveness through improved process and technology. Nothing at this point suggests that will change completely this decade.

2. Alarm Fatigue

One of the biggest hurdles to improving effectiveness is false alarms in the analytics technologies used. Not being sensitive enough means that the risk exposure is higher. However, being too sensitive also means burying the operations team under a lot of busy work chasing false alarms.

Most operations teams use expert systems, primarily based on rules, to monitor operations and generate alerts. These expert systems are set up by, who else, experts in the domain! Rules, thresholds, specific KPIs, and the tuning of each of these takes up a lot of the expert’s time in order to maintain the balance between risk exposure and team efficiency– and that has its own consequences. In Anti-Money Laundering (AML) and anti-fraud applications, for example, traditional rules-based models must be submitted to periodic re-calibration and detailed testing, at significant cost.

Rules-based systems pose a major operational dilemma. Depending on the cost of missing a true exception, models may be tuned very conservatively. In the case of AML monitoring, bank fines have been so high that the probability of missing a money-laundering transaction must be reduced as near to zero as possible. This results in large quantities of false positives (legitimate transactions flagged for operator review). Not only does this significantly increase operational cost, but they also create “alarm fatigue,” in which operators expect false alarms to such an extent that they miss a true positive and allow an improper transaction.

Institutions in both areas are looking for innovative means of reducing alarm fatigue while, at the same time, improving effectiveness.

3. Trends and Human Processing

Both when dealing with sensor and transaction-oriented data, there is an important aspect of time that affects how decisions are made. In general, humans are good at interpreting simple time trends and looking at slopes and levels. Some of these trends can be encoded in expert systems but any complex trend cannot. This is mainly due to the limits of human ability to describe complex time trends. Also, where different pieces of information don’t arrive at the same time or rate, incorporating the trends in such information into any expert system tends to be hard.

This problem is exacerbated in financial services applications where trends are formed (and change) over periods of days, weeks or even years. Individual operators cannot expect to recognize long-term trends in customer behavior without computer assistance.

The result of such difficulties is that current operations systems are not programmed to recognize complex trends. This implies that issues are discovered later than when their first signs start to appear. Also, it increases the amount of effort people have to put into confirming alarms by interpreting patterns.

4. Learning instead of Expert Systems

The next issue is that expert systems do not change by themselves, as they have to be programmed by experts. Learning occurs in the minds of experts who then apply the lessons of their learning into new versions of the rule base used by the expert system. In today’s fast changing landscape, this means operational systems cannot be evolved rapidly enough. In fraud detection, for example, monitoring rules bases need to be modified sufficiently frequently to keep up with new criminal approaches to fraud, in addition to incorporating experts’ learning about signs of potential fraud.

It should ideally be possible to use a thumbs-up/thumbs-down or a photo-tagging type of approach to improve the quality of results continuously. That requires that operations management use a learning system, more akin to the AI we use now with voice assistants, for example.

Some work is being done today in financial services monitoring “after the fact” to incorporate AI machine learning techniques. This includes providing feedback to the model as to real and false positives, allowing automated tuning of the models that identify suspect transactions. However, this is primarily still a batch process, working on feeds of transactional data from core banking systems. Wire transfers undergo real-time monitoring (particularly for sanctions scanning) but their volumes are relatively low. As other payments move more toward real-time and irreversible mechanisms across the world, the importance of real-time monitoring will greatly increase.

Experience with high-volume real-time industrial monitoring may be transferable to financial services, particularly as AI technologies are deployed and operationally proven.

5. Regulatory Oversight

A critical requirement of systems in both domains is that regulators exercise oversight over methods used in operations. Therefore, it is necessary to keep the methods explainable and, potentially, easily provable. Techniques much more complex than high school algebra can hardly be defended.

This makes black box methods, especially those that include proprietary algorithms, harder to deploy in a scalable manner. Also, it means that vendors should be willing to disclose the algorithms used and not rely on secret sauces. Finally, it requires math to be simple enough to explain to various stakeholders.

6. Trade Secrets

The last, but not the least important, challenge is to maintain complete control over the analytics models in order to derive a competitive advantage through superior operations, as well as protect against unwanted legal and criminal behavior from third parties.

Many vendors and techniques are operated as SaaS under full control of the vendors. Moreover, many vendors create data derivatives from their customer data and use some of those derivatives as an offering to other customers. While operational methods should be open to scrutiny and best practices can be shared across an industry, models should be under the control of institutions.

Falkonry, through its AI for Live Operations, provides a learning system for automatically processing trends in the data to create an accurate and timely warning system for undesirable conditions– without requiring any trade secrets about the operation to be divulged to Falkonry or anyone else. Its methods can be explained to regulators and it equips existing operations teams to operate informed by the operations data in a more effective and efficient manner.

While the linked video is specifically focused on the Internet of Things (IoT) and industrial applications, it is easy to see where there are analogous opportunities within Financial Services. A number of FinTech companies are looking at IoT for opportunities around, for example, customer experience. Banks have also traditionally looked at industrial process management methodologies to improve their own operational procedures (e.g. Six Sigma).

In summary, there is an opportunity to use industrial process exception management techniques to address some of the most challenging issues in bank operations today, the highest value among other potential areas being fraud detection and Anti-Money Laundering.

Graham Seel is an expert in commercial banking, and provides strategic insight and internal business cases to banks. He works as a fractional Customer Success Executive to Fintech firms, facilitating their partnership with banks.

 

This post was originally written June 06, 2016.

 

 

Falkonry Service Improves the AI Experience: More Connections, Better Deployment Options

By | IT/OT Management

Falkonry Service was introduced a few months ago to improve the Falkonry solution fit and flexibility. As we’ve interacted with customers over the last several months, we’ve gained a greater understanding of the many different types of solutions people are trying to build with embedded pattern recognition capabilities provided by Falkonry. This release is a reaction to those needs and includes:

  • Improved data consumption capabilities
  • Expanded connection options
  • Simpler private deployment options

Falkonry Service Architecture

 

You can watch videos about Falkonry on our Website to learn more.

Improved data consumption capabilities

To better address current needs and in anticipation of future needs, we added a new core architectural element to Falkonry called Event Buffer.  Event Buffers separate the responsibility for managing data inbound to Falkonry from Pipelines that process that data.  One obvious benefit from the addition of Event Buffers is that one data source can supply data to multiple pipelines. This capability can be used to simply allow reuse of previously loaded data or to support more complex real-time simultaneous pattern recognition scenarios. Each pipeline, for example can make independent choices on how to interpret and process the same data stream.  An additional capability associated with Event Buffers is the ability to chain Pipeline executions to each other – i.e. to route output from one Pipeline to an Event Buffer that feeds other Pipelines.

Event Buffers also support and provide a focus for a growing set of capabilities related to data consumption.  The new release, for example, allows users to supply data to Falkonry in ‘Data Historian style’ format – sets of points in the form of a <timestamp, tag, value> triple format.  This augments the tabular structure supported previously.  Likewise, JSON (line delimited) support was added to complement CSV.

Expanded Connection Options

The new release also makes it easier to connect to Falkonry and to embed it in your solutions.  New capabilities include:

  • MQTT connectivity
  • Webhooks connectivity
  • Updated REST API
  • Expanded set of client libraries
  • Updated Splunk plugin

The Falkonry UI makes it easy for an Event Buffer to subscribe to a MQTT topic or for a Pipeline’s outflow  to be published to a MQTT topic or a Webhook URL. Additional subscription and publication options will be added in the future.

Client libraries are now available for Javascript and Python and others (e.g. C# and Java) are under development.

The Falkonry Splunk App that makes it easy to bi-directionally connect Splunk to a Falkonry Service instance was updated and streamlined, and plug-ins for other data platforms are under development.

Simpler Private Deployment Options

While the Falkonry Sandbox provides a useful and effective path for experimentation and early solution development, private deployments are the primary way Falkonry gets delivered.  Private deployments can be either:

  • Virtual Private Cloud (VPC) Deployments: Falkonry deployed in customer specific VPCs on public provider infrastructure like Oracle, Microsoft Azure, Google Compute Engine (GCE), Amazon Web Services (AWS)
  • Private Deployments: Falkonry deployed on customer controlled compute and storage infrastructure.

To make installation much simpler, Falkonry now offers a downloadable option for Private Deployments.

Falkonry AI Augments Operational Decision-making

As extreme growth in the generation of live operational data continues, the need to build solutions that recognize patterns in this data grows in parallel. Falkonry is committed to making AI-based pattern recognition an easy to incorporate component of any operations-oriented solution, and this release furthers us down that path.

For more information, please visit www.falkonry.com