Most existing blockchain protocols fail to meet several key enterprise requirements, including confidentiality, acceptable transaction throughput and latency, computational efficiency (e.g. energy costs for proof -of-work consensus), and effective governance. Coco Framework, an open-source system, enables any blockchain ledger that integrates with it to meet these needs. It does so by leveraging Trusted Execution Environments (TEEs) such as Intel’s Software Guard Extensions (SGX) to make a blockchain network that can use efficient consensus protocols and implement access control rules to protect blockchain state from being accessed by unauthorized parties, even ones that operate the nodes on which it runs. Mark will describe Coco Framework’s algorithms, as well as Microsoft’s broader research and development efforts in TEEs.

Faculty Host: Majd Sakr

In our jobs and our studies we work with ideas all day long - but is there a better way to think about ideas? This entertaining and provocative talk will teach you timeless patterns and useful insights from the history of technology that can help you be more creative and productive in your own work. From the surprising origins of big ideas we take for granted, to developing creative confidence and getting better feedback from others, this lecture will change how you think about creative thinking.

Scott Berkun is a bestselling author and popular speaker on creativity, philosophy, culture, business and many other subjects. He’s the author of six books, including:  The Myths of Innovation,  Confessions of a Public Speaker, and The Year Without Pants. His work has appeared, or been mentioned, in The New York Times, The Washington Post, Forbes, The Wall Street Journal, The Economist, The Guardian, Wired magazine, USA Today, Fast Company, National Public Radio, The Huffington Post and other media.

Born and raised in Queens, NYC, he studied philosophy, computer science and design at Carnegie Mellon University, was a manager at Microsoft (’94-’03) and (’10-’12), taught creativity at the University of Washington, was a co-host of CNBC’s The Business of Innovation TV show, is named on 5 U.S. patents, blogs for Harvard Business and BusinessWeek, and has appeared as an expert on various subjects on CNN, CNBC, NPR and MSNBC. He’s also the MC and speaker coach for Ignite Seattle, a finalist in the Amtrak 2014 writer’s residency program and the director of the short film We Make Seattle.

His latest business book, The Dance of The Possible, launched on March 2017.

Booksigning: Copies of Scott's books will be available at the lecture for signature and great enjoyment.

Jeff Holden is Uber's Chief Product Officer (since early 2014), where he focuses on future disruption threats and relevant paradigm shifts. Jeff founded the Advanced Technology Group that developed Uber’s self-driving technology and is currently building Uber’s AI Lab. 

Prior to Uber, he was the Senior Vice President of Product Management at Groupon (2011-2014), and before Groupon, Jeff was the co-founder and CEO of Pelago (2006-2011), whose flagship product, Whrrl, was designed to help people escape their social rut and discover new things in their cities.  Previous to this, Jeff was at (1997-2006). For the latter half of his tenure there, he was SVP of Consumer Applications, Worldwide. He was responsible for the worldwide consumer website experience, including personalization, ordering, community, search, automated merchandising and online traffic (Associates, SEM/SEO). One of his proudest accomplishments during this period was leading the development of Amazon Prime. Earlier at Amazon, Jeff led the development of Amazon's supply chain technology, evolving it from a script and a modem to the architecture still largely in place today.

Prior to Amazon, he worked at D. E. Shaw & Co., L.P. in New York from 1992 to 1997, where he joined fresh out of school as a software engineer working with Jeff Bezos to develop Shaw's third-market strategy. Jeff attended the University of Illinois at Urbana-Champaign from 1986-1991, where he earned a BS and MS in Computer Science.

4:10 pm - Distinguished [ˈdōˌnət]s

 Vice President and Head Scientist for Amazon Alexa returns to campus to talk about Alexa, machine learning, and the challenges in conversational AI.

Joined for post-talk presentations and discussion by:

A new technology that reaches the market most likely finds its roots in a fundamental discovery. Basic researchers have the potential to have direct societal impact when they expand their efforts beyond research and actively engage in turning their ideas into useful technologies. This requires building relationships outside the research world and helping to bridge the gap between the lab and industry. When researchers accept this role, the outcome can be extraordinary: people’s lives may be transformed, new jobs created, and wealth generated for the inventors and their institution. So what’s the problem? Let’s do it then!

Fred Farina is chief innovation and corporate partnerships officer at the California Institute of Technology. His responsibilities include commercializing inventions made at Caltech and the Jet Propulsion Laboratory (JPL/NASA) through the creation of startup ventures and partnerships with established companies. His office is responsible for evaluating inventions, supervising patent prosecution and portfolio management, negotiating licensing deals with industry, assisting Caltech/JPL entrepreneurs with the creation of startup companies and establishing research collaborations with industry. Previously, Farina was at a law firm prosecuting patent applications on various technologies before the U.S. and European patent offices. Earlier, he was a research engineer in the GPS field at JPL and the University of Miami. He has a degree in electrical engineering from the Institut National des Sciences Appliquees in Lyon, France, and a master's from Caltech. He is a registered patent agent with the U.S. Patent and Trademark Office.

This lecture is hosted in collaboration with the Swartz Center for Entrepreneurship at Carnegie Mellon University.

3:00 pm: Distinguished Donuts

Microsoft has been deploying FPGAs in every Azure server over the last several years, creating a cloud that can be reconfigured to optimize a diverse set of applications and functions.  The performance, cost, power usage, and flexibility of FPGAs sit between the generality of CPUs and the fixed-function of ASICs, making them the ideal vehicle for relatively static functionality not suitable for the permanence of ASICs. Microsoft is already using FPGAs for Bing search ranking, deep neural network (DNN) training, and software defined networking (SDN) acceleration.  Using FPGAs, for example, Azure accelerated networking reduces the latency between virtual machines by up to 10x while making CPUs available for other tasks. Mark will describe the FPGAs capabilities, explain how these applications take advantage of them, show live demos of the performance that FPGAs provide, and discuss possible future uses.

Mark Russinovich is Chief Technology Officer for Microsoft Azure, Microsoft’s global enterprise-grade cloud platform. A widely recognized expert in distributed systems and operating systems, Mark earned a Ph.D. in computer engineering from Carnegie Mellon University. He later co-founded Winternals Software, joining Microsoft in 2006 when the company was acquired. Today he remains the primary author of the Sysinternals tools and website, which include dozens of popular Windows administration and diagnostic utilities. Mark is a popular speaker at industry conferences such as IPExpo, Microsoft Ignite and Build, and RSA Conference. He has also authored several nonfiction and fiction books, including the Microsoft Press Windows Internals book series, as well as fictional cyber security thrillers Zero Day, Trojan Horse and Rogue Code.

Faculty Host: Majd Sakr

Can passwords be both HUMANLY USABLE and SECURE? Indeed, what can a human compute in his/her head, with no paper, pencil or computing device? To address these questions, we propose a rigorous model of human computation and apply it to the everyday task of generating and remembering passwords. Our main finding is that there exist password schemas that are (a) precisely defined (b) humanly usable and (c) secure to a well-defined extent, in that individual passwords are hard to guess, and knowing several passwords does not enable an adversary to infer others. Our Human Usability Measure (HUM) has other applications, both useful (optimal user-interface design) and fun (coin-flipping over skype).

This is joint work with (and largely by) MNLBLM.

Santosh Vempala is a Distinguished Professor of Computer Science in the College of Computing at the Georgia Institute of Technology. He joined the College of Computing in the fall of 2006 as a professor of Computer Science. Hespearheaded the Algorithms and Randomness Center and ThinkTank at Georgia Tech, and served as its first director from 2006 until 2011. His research interests include algorithms, randomness, geometry and computing-for-good (C4G). He graduated from CMU in 1997, advised by Avrim Blum, and then taught at at MIT until 2006 except for a year as a Miller fellow at UC Berkeley.  Vempala is also a Sloan, Guggenheim, ACM and generally excitable Fellow, especially when a phenomenon that appears complex from one perspective, turns out to be simple from another. In recent years, he has been trying to understand, with little success, how the brain works and how to model its computational abilities.

Faculty Host: Manuel Blum

Distinguished Donuts: 4:15 pm

Matt Humphrey enrolled at CMU at the young age of 13 and received his BS in Computer Science (‘07) and an MBA (‘08). During business school he and colleagues developed a peer-to-peer video delivery company, Eivod, which became the first Project Olympus PROBE and was eventually funded by Y Combinator.

Currently, Matt is the co-founder and CEO of LendingHome, the leader in mortgage marketplace lending. LendingHome has made over $500M in home loans across the country and raised $100M+ in venture capital. Prior to LendingHome, Matt had founded HomeRun, an e-commerce platform company which was acquired for north of $100M just over a year after launch. Matt  has also co-founded Envival, Cloudant and Bumba Labs. He has been an active angel investor and advisor since 2012, with over 75 investments in notable technology companies around the world.

A native of Pittsburgh, Matt loves driving race cars, traveling the world and watching his beloved Pittsburgh Steelers.

Is the mind computable? Can we build sentient machines? What are the implications for humanity if we can?

The human race may be singular, unique across all of time and space. It may be just one of multitudes. Most likely, however, it is an extremely rare thing, an exquisitely precious consequence of the unfolding of the laws of the universe. Still, one truth that we can assert with confidence is that we are. We are self-aware; we know that we know we exist. There is within humanity a drive to recreate itself, to be as a god to things we create in our own image. From the Golem of Jewish mythology, to Leonardo’s robot, to the contemporary Kenshiro robot, we project our hopes and our fears into cunning mechanism that mirror us. While these anthropomorphic robots are interesting (and perhaps a bit creepy), there is a less visible revolution taking place in cognitive computing, whose advances are not only helping us better understand the operation of the human brain, they are leading us to create the illusion of sentience.

In this lecture, Grady explores the development of intelligent computers as projections of what we both dream and what we fear. We examine what it means to be intelligent, and take a journey through past and future approaches to building sentient software-intensive systems. Some such as Minsky believe the mind to be computable; others such as Penrose do not. In the end, we are compelled to consider the question of what it means to be human: producing even the illusion of the mind raises profound questions as to their personhood and our relationship to these machines.

Grady is recognized internationally for his innovative work in software architecture, software engineering, and collaborative development environments. He has devoted his life's work to improving the art and the science of software development. Grady served as Chief Scientist of Rational Software Corporation since its founding in 1981 and through its acquisition by IBM in 2003. Grady has been deeply involved in IBM's cognitive systems strategy, for Watson and systems that extend and transcend Watson. Grady led the IBM Global Technology Outlook topic on cognitive systems, and now he continues to work with the key architects of the Watson Group and the sister organization in IBM Research to further advance the science and the practice of cognitive systems.

He now is part of the IBM Almaden Research Laboratory serving as Chief Scientist for Software Engineering, where he continues his work on the Handbook of Software Architecture and also leads several projects in software engineering that are beyond the constraints of immediate product horizons. Grady continues to engage with customers working on real problems and maintains deep relationships with academia and other research organizations around the world, including serving as strategic advisor to Dr. Pat O'Sullivan's applied innovation team at th IBM Dublin Software Laboratory. Grady is one of the original authors of the Unified Modeling Language (UML) and was also one of the original developers of several of Rational's products. 

Grady is the author of six best-selling books, including the UML Users Guide and the seminal Object-Oriented Analysis and Design with Applications. He writes a regular column on architecture for IEEE Software. Grady has published several hundred articles on software engineering, including papers published in the early '80s that originated the term and practice of object-oriented design (OOD), plus papers published in the early 2000's that originated the term and practice of collaborative development environments (CDE).

He is an IBM Fellow, an ACM Fellow, an IEEE Fellow, a World Technology Network Fellow, a Software Development Forum Visionary, and a recipient of Dr. Dobb's Excellence in Programming award plus three Jolt Awards.

Hosts: Mary Shaw, Nancy Mead

Our CMDragons team of autonomous soccer robots is the 2015 champion of the RoboCup Small Size League. The team won all of its six games, scoring a total of 48 goals and conceding 0. This unprecedented dominant performance is certainly the result of many algorithms, including a very effective defense, but we strongly credit our new coordinated, aggressive, and continuous offense.

In the talk, we will present our coordination algorithm for real-time planning and execution at multiple behavioral levels: (i) a multi-robot play representation to map the state of the game to an appropriate set of tactical roles, (ii) the optimal robot-to-role assignment, (iii) coordinated evaluation to fully instantiate each role, and (iv) the probabilistic optimal action selection of individual robots. We demonstrate the effectiveness of the CMDragons-2015 team through results and statistics of the team performance at RoboCup 2015, and through controlled experiments using a physics-based simulator and an automated referee. The talk will be illustrated with the algorithm narration of the highlights of the real games. Finally, the talk will include a comprehensive historic overview of our research and team performance in robot soccer since its beginning in 1997.

Highlights from the entire tournament, with every goal for which we have footage, can be found at:
Detailed highlights from the final game (score 5-0) can be found at:

The CMDragons research team is composed of Manuela Veloso, Joydeep Biswas, PhD RI, 2014, PhD students Juan Pablo Mendoza (RI), Richard Wang (CS), Danny Zhu (CS), Philip Cooksey (RI), and Master student Steven Klee (CS). The 10-year old robots were created by former CMU researcher Mike Licitra and remarkably maintained by Biswas and Wang.


Subscribe to SCS DLS