News

  • 0

Introspective Systems’ Founder Caryl Johnson Offers a Fresh Approach to Seismic Network Analytics at American Geophysical Union Fall Meeting

Introspective Systems’ xGraph collaborative AI technology uses an executable graph framework and an architecture of streaming connected microservices to provide accurate and timely information about seismic activity to the USGS

PORTLAND, ME ­– December 14, 2017 – Introspective Systems, LLC’s founder and Chief Innovation Officer Caryl Johnson has addressed the 2017 American Geophysical Union’s Fall meeting about the company’s work with the US Geological Survey (USGS). The USGS uses Introspective Systems’ xGraph collaborative AI platform for monitoring and collecting data on earthquakes. xQuake, built upon the xGraph platform, is a data storage and analytics system that offers an innovative approach to seismic network analytics by combing complexity theory with an adaptive architecture of streaming connected microservices to gather and curate diverse seismic data.

The USGS gathers seismic data via a worldwide network as part of its mission to improve real-time detection. The sensors on this network generate many megabytes of data per second, all of which is streamed over satellite links to the USGS National Earthquake Information Center. The USGS team uses xQuake to combine and analyze this massive influx of data in real time to determine if it is necessary to alert first responders and foreign governments.

“Seismic networks have expanded during recent decades, and our need for accurate and timely information has increased dramatically.  xGraph was designed to meet these needs,” says Caryl Johnson, founder and Chief Innovation Officer, Introspective Systems. “The USGS has been a leader in adopting new technologies such as xQuake, built on our xGraph system, to deliver the complex data management and analytics required to accurately track and manage seismic activity.”

The USGS project demonstrates some of the unique capabilities of the xGraph platform. It is designed for challenges that have the following characteristics:

  • Massive Data Loads/High Velocity. Data comes from numerous “systems” worldwide, each of which contains multiple elements: sensors, which are part of global arrays, which are distributed in instrument sites, all of which capture multiple levels of data for each phase or aspect of a geological event. xGraph is designed to manage diverse and disparate data from a variety of systems that operate on their own as well as part of a network.
  • Fractal Structure. The volume of sensor data is so large and streams at such velocity that it requires xGraph’s exceptionally robust platform for collection and analysis.
  • Real-Time Analytics. Minutes count in notifying first responders and populations about earthquake threats, so processing massive amounts of data from various sources must be instantaneous. Because xGraph processes data on the edge of the cloud, it does not lose time or suffer from low latency experienced by traditional centralized big data platforms.
  • Collaborative AI. Seismic patterns change, requiring networks to adapt while running. One example is distinguishing the recent uptick in earthquakes related to oil production and hydro-fracturing from other events. xGraph uses collaborative intelligence to analyze seismic stations in real time to compensate for events and evolve to learn new patterns over time.

Introspective Systems has released the xQuake system under an open source license to encourage and enable the seismic community in further development of its capabilities.

The American Geophysical Union meeting is this week in New Orleans, LA. For more information visit its website.

About Introspective Systems, LLC

Introspective Systems is the developer of xGraph, a collaborative AI platform that helps developers solve complex modern computing problems. It is designed to tackle critical issues defined by massive amounts of streaming data, interconnected systems, and the need for instantaneous decision-making such as managing robot swarms, self-driving vehicles, and complex energy grids. xGraph combines edge computing with distributed micro-services to speed processing time while enabling autonomous and collaborative decision-making. Current applications include managing an IoT-based electrical micro-grid, locating earthquakes worldwide, and playing the game StarCraft II to advance collaborative AI. Learn more at www.IntrospectiveSystems.com


  • 0
Unknown-1

Portland Press Herald | Portland high-tech company lands entrepreneurial thinker Gooding

 

From the Portland Press Herald, December 7, 2017: “The former head of the Maine Center for Entrepreneurial Development has joined a Portland company that develops artificial intelligence. Don Gooding has been named chairman of the board for Introspective Systems, a small company that works on a myriad of complicated, data-based problems such as realigning the power grid for the U.S. Department of Energy.

Gooding, who was an angel, or seed, investor when the company started, has increased his stake. His new duties include guiding Introspective Systems through funding rounds, identifying market opportunities and investors, and providing leadership counsel to ensure and accelerate growth, according to a news release from the company.”


  • 0
btfitz_20140312_0307_M

Don Gooding Named as Chairman of the Board of Collaborative AI Startup Introspective Systems

Gooding’s expertise as a seasoned business growth mentor, technology market research insider, and angel investor will propel launch of company’s xGraph platform for complex digital problems

PORTLAND, ME ­– December 7, 2017 – Introspective Systems, LLC announces that it has named Don Gooding as Chairman of the Board. As an early angel investor, Gooding steps into this new role to accelerate the launch of Introspective Systems’ xGraph collaborative AI platform. He brings a wide range of expertise, including in-depth knowledge of the technology landscape, business acceleration, and top-tier venture capital. xGraph offers a new approach to complex digital problems characterized by massive volumes of streaming data such as from the Internet of Things (IoT) that require a combination of autonomy and collaboration provided by artificial intelligence (AI).

Gooding participated in the first angel investment round of Introspective Systems and has continued to increase his stake to become the company’s largest outside investor. His role as Chairman of the Board will include guiding Introspective Systems through funding rounds, identifying market opportunities and investors, and providing leadership counsel to ensure and accelerate growth.

“Early on, I saw the possibilities of Introspective Systems’ technology and helped guide the company through its ‘Valley of Death’ when its innovation was ahead of the market,” says Don Gooding, Chairman of the Board, Introspective Systems LLC. “xGraph has emerged as a powerful platform for solving complex digital problems and I am excited to be a part of helping Introspective Systems accelerate to catch today’s huge waves of ‘AI Everywhere’ and ‘Edge Computing’”.

“Don’s experience, connections, know-how, and support are exactly what we need to take Introspective Systems to the next level,” says Kay Aikin, CEO, Introspective Systems. “We could not ask for a better guide and champion than Don, and we are thrilled that he has formally joined our team.”

Gooding is a pioneer in Maine’s modern entrepreneurial movement. Before founding Four Colors of Money for Entrepreneurs, a consultancy to help young and growing businesses understand and choose the right financing options, he served as Director of the Maine Center for Entrepreneurial Development (now called the Maine Center for Entrepreneurs). During his five-year tenure, Gooding expanded its Top Gun accelerator program from one to four locations throughout the state. He hosted the first two seasons of Greenlight Maine, a TV show where entrepreneurs compete for a $100,000 prize. Gooding has been a member of the Maine Angels investment group since 2005, including serving as Vice Chair.

Early in his career, Gooding ran telecommunications market research for The Yankee Group; he still maintains close relationships with colleagues who became founders and leaders at firms including Forrester Research and IDC. Gooding was the first Research Partner in venture capital at Accel Partners, where he focused on finding and evaluating telecom and network startups. In that role, he was a board observer for UUNET, the breakout ISP in the early days of the Internet.

In addition to guiding startups and tracking technology, Gooding’s passion is music. He launched A-Cappella.com to sell sheet music and recordings and to start a competition for college and high school singers; it was the “real life” organization popularized in the Pitch Perfect movies. Gooding and his wife are honored by the annual Gooding Cups provided to champion student a cappella groups.

Introspective Systems’ xGraph platform is designed for building solutions to today’s most complex digital challenges, including those with the following characteristics:

  • Massive Data Loads/High Velocity. Data volume and speed of the Internet of Things and other massively distributed systems require xGraph’s exceptionally robust platform for rapid analysis and decision-making.
  • Fractal Nature. xGraph is designed to manage diverse and disparate data from a variety of systems that operate on their own as well as part of a network.
  • Real-Time Analytics. xGraph can process data on the edge of the cloud, which reduces latency to deliver the fast decision-making required by autonomous applications such as drones and self-driving cars.
  • Collaborative AI. xGraph uses collaborative intelligence to analyze data in real time, direct responses and actions, and evolve to learn new patterns over time.

About Introspective Systems, LLC

Introspective Systems is the developer of xGraph, a collaborative AI platform that helps developers solve complex computing problems better, cheaper, and faster. It is designed to tackle critical issues defined by massive amounts of streaming data, interconnected systems, and the need for instantaneous decision-making such as managing robot swarms, self-driving vehicles, and complex energy grids. xGraph combines edge computing to speed processing time with distributed micro-services to enable autonomous and collaborative decision-making. Current applications include managing the IoT-based electrical grid of the near future, locating earthquakes worldwide, and creating a SaaS “laboratory” to advance collaborative AI using the game StarCraft II. Learn more at www.IntrospectiveSystems.com

 

 

 

 


  • 0
post-xGraph-medium

Introspective Systems’ xGraph Collaborative AI Platform at Forrester NewTech Forum

The intersection of AI and IoT technology is one of the topics at Forrester’s New Tech Forum 2017 set for December 6, and Introspective Systems will be at the center. Its xGraph platform is one of the 18 startups to be featured at this annual event that looks into the year ahead to discuss key trends, concepts, and technologies that will shape the digital enterprise.

“With the growth in complexity of digital problems, we believe that collaborative AI with computing on the edge is the way of the future,” says Kay Aikin, CEO of Introspective Systems. “The time is right for our xGraph platform, which is designed for the complex challenges posed by AI and the IoT.”

xGraph is an executable graph framework for intelligent and collaborative edge computing that solves big problems: those that have massive decision spaces, tons of data, are highly distributed, dynamically reconfigure, and need instantaneous decision-making. It’s developed by Portland, ME startup Introspective Systems, led by Kay Aikin and Caryl Johnson.

Many analysts believe that the greatest problem with big analytics and especially the Internet of Things has been the application layer. This application layer is the process that specific devices communicate and accomplish specific tasks. Many industry players have each created their own application layer, including Apple, Google, and Amazon. Others, such as Microsoft with its Azure IoT Suite, have built platform layers that combine multiple technologies with design patterns including hubs and gateways, analytics, storage, web interfaces and data visualization. This creates an environment where there are multiple standards, proprietary layers, and multiple technologies for developers.

Introspective Systems deals with this challenge by using xGraph, which is a distributed and adaptive code and data network in solving complex, real-time, highly dynamic systems-of-systems problems.

The company will share insight on its development, applications, and technology at the Forrester event.


  • 0

From Deep Blue to xGraph: Gaming Out AI Challenges

posterIt’s been 20 years since Deep Blue beat Garry Kasparov in chess. As impressive a win as that was for artificial intelligence and computing, the game has changed. Chess players have about five million options during the course of an average game; StarCraft has roughly 10300 choices per move. Our full stack programmer, Amanda Castonguay, is using our xGraph platform solution to find a way to beat StarCraft.

This is more than a game – it’s a representation of how computing problems have become exponentially more complicated over time. They have massive decision spaces with an ever-growing number of “moves” or options. Any one of those choices can kick off a chain of unintended consequences, meaning complex problems can and will morph and grow even more complex. They have a high volume of streaming data – more than two zettabytes per year are flooding data centers at super high velocity.

Compounding this is the need for speed: self-driving cars can’t wait for data to go from every vehicle on the street to the cloud and back to find out whether or not they’re too close to the car (or kid) in front of them. Latency is one of the issues we’re trying to solve with XGraph, and Amanda’s work is part of this.
xGraph is our new technology platform. It’s an executable graph framework for intelligent and collaborative edge computing that solves big problems: those that have massive decision spaces, tons of data, are highly distributed, dynamically reconfigure, and need instantaneous decision-making.

Amanda recently presented her StarCraft research at the 2017 IEEE MIT Undergraduate Research Technology Conference (URTC). This official IEEE conference brings together undergraduates around the world to present, discuss, and develop solutions to advance technology for humanity. We’re excited that she had the opportunity to share her findings. If you’d like to follow along, sign up for our newsletter to learn more about what we’re doing to beat StarCraft, manage drone swarms, fix our energy grid, discover earthquakes, and more.


  • 0

Transactive Energy for Microgrids

Tags : 

Our CEO Kay Aikin wrote a guest blog for Microgrid Knowledge the premier industry newsletter focused on Microgrid technologies and policy.

MRRA Microgrid

MRRA Microgrid

Toward Market-based Microgrid Control Systems

The largest machine on earth is often said to be our electrical grid. By the end of 2016 there were some 7600 power plants greater than 1 MW with many times more, smaller resources and an even larger number of control systems. The grid is truly a complex machine that is made up of systems collected into ever larger systems.  In the controls world this is called a systems of systems. This is the ultimate in complexity and is true on all large systems like the electric grid and ecological systems.

The advantage of microgrids is they help us tame the complexity the grid by limiting the number of possible interactions within the grid. Smaller more predictable systems are less prone to unintended consequences as can happen in the electrical grid like the Northeast blackout of 2003. However, even with microgrids and the emergence of energy management systems and smart devices the control networks of a microgrid will become even more complex. But this complexity can be used to build better systems as Ecologist Eric Berlow says the more you “embrace complexity the better chance you have finding simple answers”. A great Ted talk illustrating this concept can be seen at: https://www.ted.com/talks/eric_berlow_how_complexity_leads_to_simplicity

It has been shown that market-based systems can be amazingly stable in complex environments because of the many naturally balancing feedback loops within the system. Using this complexity, to lead to simplicity. In the electrical engineering field, the GridWise Alliance has called this idea of market-based system “Transactive Energy”. Much of the research in this area has been done by the Pacific Northwest National Laboratory and one of the successful trials was the Olympia Peninsula Project (OPP) consisting of a field demonstration of price signal-based control of distributed energy resources. The demonstration showed that market-based control was able to manage distribution constraints and reduce peak loads. This was followed with the Pacific Northwest Demonstration Project (PNDP) ending in 2015.

You can consider there to be three main methods for implementing Transactive Energy control systems applicable to microgrids including:

  • Centralized (top down)
  • Centralized (auction-based)
  • Distributed (edge-based)

The difference between the Olympia Peninsula and Pacific Northwest Project was that OPP was a double auction very similar to the current ISO systems and PNDP was a top down model where demand response assets and distributed energy resources were optimally dispatched by individual “transactive” nodes using two-way communications. Both both were generally more centralized paradigms.

The third transactive approach is a fully distributed edge control method that relies on pricing signals reflecting the prediction of future conditions creating a different price at many different scales.  Lower-level devices (or entire systems) respond to those pricing signals from higher levels. This method is currently being researched at Maine’s Brunswick Landing Microgrid Project (add link) for the Department of Energy.

While the first two transactive energy approaches have shown promise in that they have been able to balance energy demand, lowering peak demand and managing grid congestion they rely on large two-way communication networks that are particularly vulnerable to cyber assaults.

This cyber vulnerability should be a concern for the microgrid community because for a wide spread deployment of a system of microgrids this communication vulnerability is of particular concern to today’s infrastructure experts.

The edge-based system being researched at Brunswick Landing has pricing signals (potential of 10 or more different grid scales) are continuously re-calculated, only travel in a downward direction and are acted upon only by edge devices have promise, using the “power of complexity to lead to simplicity” Since the scope of influence of a single node is typically only one or two degrees of separation as described by Eric Berlow, this limits the computing power required to calculate the system state and provides for enhanced capabilities using advanced artificial intelligence techniques and limits security risks with limited communications routes. An effective transactive edge-based energy system can provide increased resilience, versatility, reliability and flexibility when used not only in microgrids but the greater electrical grid.

Kay Aikin is CEO of Introspective Systems, a complex systems architecture and engineering company in Portland Maine. Introspective Systems is the project lead at the Brunswick Landing Microgrid Project researching edge-based transactive energy networks.


  • 0

E2Tech: Electrical Grid Futures

Tags : 

Kay Aikin, Chief Executive Officer of Introspective Systems presented at the Environmental and Energy Technology Councils discussion on Electrical Grid Futures-Innovation in the Energy Sector. Kay described how the companies work at Brunswick Landing and how it is an ideal location for energy technology innovation. Brunswick Landing has its own power grid, including an anerobic digester generation facility that is providing electricity to the business park’s tenants. Introspective Systems is collaborating with Brunswick Landing, and Franklin Electric’s INCON division in Saco, Maine on an innovative microgrid solution to generate green, reliable power and mange it using real-time pricing mechanisms and artificial intelligent control algorithms. Further discussions included the role of enhanced software, traditional utilities, manufacturers, and renewable energy will have in the electricity sector of the future. And, how will innovation in the electricity sector help inform Maine’s energy, environmental, and economic Roadmap to the Future?

Brunswick Landing Electrical Grid

Brunswick Landing Microgrid System Map

Kay’s presentation can be accessed here.

The other presentations at the event  can be accessed at the following link:

 http://www.e2tech.org/event-2589330


  • 0

Department of Energy Phase II SBIR: Fractal Grid Framework for an Evolving Grid Architecture

Press Brief:

On Tuesday, June 20th, 2017 Introspective Systems LLC. was awarded a Phase II Small Business Innovation Research project “Fractal Graph Framework for an Evolving Grid Architecture” for grid$986,802. Using Brunswick Landings Renewable Energy Center 4 MW Microgrid as a test-bed the project further develops the grid architecture and control algorithms for a new grid management technology based on edge intelligence.

The objective of this phase is to confirm the dynamic stability of managing a microgrid (a small self contained energy system with demand, supply and control) using highly granular real-time pricing signals balancing demand and supply while keeping system engineering requirements within acceptable tolerances. The technology developed ­­­uses market-based economic principals and artificial intelligence within Introspective Systems patented xGraph micro-services edge computer architecture that was specially modified for the grid application.

The AI algorithms in gateways distributed throughout the Brunswick Landing microgrid develop a well-defined set of buy/sell pricing signals and transmit those signals to edge devices in the level below. The edge devices use locally and temporally learned intelligent algorithms to control their own operations deciding on when it is most advantageous to consume or supply power to the greater grid. The system control is thus distributed to the very edge of the electrical grid rather than current control schemes that rely on top-down approaches that are computationally inefficient and have potential cyber-security risks. The overall control concept is called “grid edge intelligence”.


  • 0

Earthquake software presented

AGU15-logoIntrospective Systems earthquake location software was presented at the American Geophysical Union (AGU) convention in San Francisco between December 14-18 and made quite a buzz with participants. Sandia National labs, Berkeley University and Air Force Tactical Applications Command (AFTAC) have spent 7 years and millions of dollars in attempts to creat a real-time bayesian approach to locating earthquakes. They each spent more than 2 hours each studying the research and the real-time earthquake display being presented.

About the Software

Legacy processing approaches to seismic networks are based upon algorithms developed in thGlobee last century. Introspective Systems has developed new approaches that bring seismic Networks into the 21st century using all of the capabilities of modern computers for unprecedented analysis and monitoring of seismic events. Introspective Systems network capabilities are based upon xGraph a highly dynamic executable graph framework that distributes the analytics across millions of processes or processors.

The job of an earthquake data associator is to gather an ensemble of data types such as phase picks, beams, or other data from social media into another ensemble of collections identified as discrete events, each representing the occurrence of an earthquake, quarry blast, or nuclear detonation. Legacy approaches to the automation of this process have traditionally been procedural in nature, and more recently combinatorial as in some of the applications of Markov chain Monte Carlo (MCMC) in this problem domain. A similar approach, as a dense, global network of association nodes, has been in use at the International Data Center (IDC) of the Comprehensive Nuclear-Test-Ban-Treaty Organization since its inception in the late 1990s.

Introspective Systems Seismic Networks implementation is radically different from those in use by other seismic observatories. The xGraph framework provides a real-time, in tempo approach that streamlines acquisition and analysis without batching of picks or other data. The algorithm is neither procedural nor is it combinatorial as with MCMC based approaches, for example. Instead it is draws from complexity theory where a catalog is created as a guided, self-organized criticality, designed for a cloud-based environment. Figuratively speaking with respect to the common “sand pile” metaphor, the grains of sand are the swarm of arriving picks and other data, and the emergent sand pile is an earthquake catalog where the binding energy is represented as the Bayesian affinity between picks and quakes.


  • 0

xGraph mentioned in Microsoft podcast

Introspective Systems and xGraph was mentioned in a national podcast Dotnetrocks which focuses on Microsoft .net products. The show focused on dotnetrocksMicrosoft Executive Vice President Jason Zander of the Azure Big Data Platform. The host Carl Franklin understands the benefits of xGraph in the Internet of Things business space that is diametrically opposed to Microsoft’s approach. There were some very interesting discussions about the relative value of keeping the intelligence in the cloud as Microsoft does or  embed intelligence at the edge like xGraph does.

http://dotnetrocks.com/?show=1213

Our section starts at 2:45 and goes to 5:00

Show synopsis: How can Azure change your business? Carl and Richard talk to Jason Zander, one of the original developers of .NET and now a corporate vice president, about the power of Azure to affect change in your business. Jason talks about the landscape of Azure today, and how the engineers are able to push out a feature almost every week – 500 new features in the past year! You may not need to move that fast, but it’s nice to know if you build against Azure, that’s what is possible. The conversation ranges over the role of IoT, security, web sites and more – there’s a lot of things that can be built in Azure, and the opportunities are massive!


Jason Zander is the Corporate Vice President of the Microsoft Azure Team in the Cloud & Enterprise group at Microsoft Corporation. Jason’s team responsibilities include the architecture and development of core Azure technologies (such as compute, storage, networking, public cloud data centers and engineering systems). Jason also leads development of the Azure IoT effort at Microsoft.

Before joining the Azure team, Jason was CVP of engineering for the Visual Studio team and has worked on numerous products at Microsoft including the first several releases of the CLR and .NET Framework, SourceSafe, and ODBC. Before joining Microsoft in 1992, Jason worked at IBM on Distributed SQL and SQL/400 at the Rochester lab. Jason holds a Bachelor of Science in Computer Science from MSU.