Visit to Bell Labs Shannon Conference April 2016
Title: The Bell Labs Shannon Conference on the Future of the Information Age.
April 28 Day:
April 28 Evening:
April 29 Day:
My Participation and Preparation
It was fortuitous to be invited to this event by Marcus Weldon, the President of Bell Labs. There is a pleasing synchronicity, the 100th anniversary of Claude Shannon's birth is the 50th anniversary of the start of my career at Bell Labs; and I received my Master's degree in Information Science (aka Computer Science) 20 years after Shannon's seminal 1948 paper founding Information Theory. I feel that I am a kind of surrogate for thousands of retired Bell Labs employees who cherish their fond memories of improving the world while working at Bell Labs.
For conference preparation, I bought a copy of Shannon's 1948 paper The Mathematical Theory of Communication. I also bought a copy of Bell Labs book (fall 2015), The Future X Network A Bell Labs Perspective. Scanning these books helped familiarize me with terminology. It was fun to read portions of Shannon's 1937 Master thesis (free pdf copy). I learned basic Boolean algebra in the summer of 1961 as a high school junior at a 6-week NSF advanced math camp.
The other key preparation was to listen (on Audible.com) for the second time to the magisterial history of Bell Labs, The Idea Factory Bell Labs and the Great Age of American Innovation by Jon Gertner. This book focused on six giants of Bell Labs, and one of them is Claude Shannon. It is the best currently available biography of Claude Shannon.
Another recommended book is the wondrous The Information by the acclaimed author James Gleick, who explains the relationship of scientific ideas effortlessly. Claude Shannon is one of the major figures in this delightful book. Gleick has a very good explanation of Shannon's Information Theory and the concept of entropy. He explains the deep impact Information Theory had other sciences like genetics.
Other memories triggered when I noticed the Murray Hill auditorium was named for R. W. Hamming. One of the first things I learned in my career in 1966 working on No. 1 ESS was the Hamming code used in the Program Store memory. There were 44 bits in each word: 37 information bits and 7 extended Hamming bits. These 7 bits provided single error correction and double error detection. This is my first memory of the power of information theory. While writing this report I found a lecture Hamming gave in 1995 on Information Theory. Of course, he knew Shannon.
To get in the spirit for the Shannon conference, I made a road trip to visit major sites associated with some of the top disruptive inventors in American history. A 2800 mile road trip from my home in the suburbs of Chicago to Murray Hill and back made a nice loop to see the homes, labs, graves and basic geography of Claude Shannon, Thomas Edison, Henry Ford, the Wright Brothers and Benjamin Franklin. I created a separate trip report for that "Inventor's Loop".
In preparation for that trip, I discovered that Claude's elder sister, Catherine, lived in Downers Grove, IL. This is where I live! She was a organist for the First Presbyterian Church. This is where I vote just 2 blocks from my home! She was a Math Professor at North Central College in Naperville, the city where I spent my entire career working at Bell Labs.
I visited Claude Shannon's home town of Gaylord, Michigan on the way to the conference. There they called him the Dad of Digital. Here was the headline in the local newspaper last year on the 99th anniversary of his birth.
|Conference Tents near Hamming Auditorium|
To get a feel for the importance of this conference, watch this 5 minute video on The Shannon Limit.
Presentations were held in the Hamming Auditorium in the main building of Bell Labs and demos were held in other parts of the building. Morning coffee, lunch, dinner and the Human Digital Orchestra were held in tents adjoining the Auditorium which gave a festive air.
For a visual feel for the conference and its benevolent atmosphere, watch this eight minute YouTube Video.
In addition to industry luminaries and noted researchers giving presentations, one of the technical talks was from a Nobel prize winner, Christopher Sims.
The invitation-only conference was smoothly organized with the well known Bell Labs collegial atmosphere. It was an intellectual treat. During breaks, I had some brief conversations with Morris Tanenbaum, the inventor of the silicon transistor and former executive VP of AT&T and with Bob Metcalfe, the inventor of Ethernet.
Bell Labs employees volunteered to be a kind of host/point of contact for guests. I enjoyed having lunches with my volunteer host, Shreyas Shah, a recent PhD hire, and learn about his project. It was fascinating to think that he is at the same stage of his career that Dr. Tanenbaum was when he was assigned to work on the silicon transistor.
Welcome from the Bell Labs President
Dr. Marcus Weldon opened the conference. He is fluent and charming with a microphone in his hand and the same in one on one discussions. As well as explaining the purposes and objectives of the conference and its historical background, he shared his vision for Bell Labs.
Shannon Visionary Award Winners
All five were well deserving of a Shannon Visionary award. It is interesting to note they were born in separate decades suggesting the range of the candidates considered. All five have been widely recognized and have academic and business experience.
1. Irwin Jacobs is the co-founder of Qualcomm and one of the pioneers and proponents of CDMA. Professor Irwin theme was if you believe your product or technical approach is superior then stick to your guns and you should win in the end game. He knew that CDMA with its inherent higher traffic capacity would be the long-term winner. He persisted in his vision for over 10 years as standards groups, countries and other players in the mobile space gradually converted to his vision. Now over 3 billion mobile users are on CDMA. See his presentation.
2. Bob Metcalfe, Internet Pioneer, was co-inventor of Ethernet while at Xerox PARC and founded 3Com. He also formulated Metcalfe’s Law. Currently he is a Professor at the Innovation Center in the Cockrell School of Engineering at the University of Texas.
(Dealers of Lightening: Xerox PARC and the Dawn of the Computer Age" is an informative history.)
Bob's talk expanded on Jacobs’ theme, of the need for tenacity and perseverance in the face of standards bodies and large established companies with their own agendas. I’m still laughing at Bob’s quip about de jure, de facto & de ibmo.
Bob explained other needs for startups to succeed. One model for success is to have inventors, adult supervision and angel investors. Adult supervision is often the missing piece. See his presentation.
I had a brief chat with Bob before the presentations the next day. I mentioned the Parrish Computer Science Scholarship at UT, and Bob responded with a Hook ‘em Horns and a big smile. Two weeks after I graduated from UT in 1966, I started at Bell Labs in Holmdel.
3. Eric Schmidt is the Executive Chairman of Alphabet, the owner of Google. He is the classic “adult supervision” mentioned in Metcalfe’s talk. One theme Eric mentioned that was not discussed by other speakers was the need for low regulation or no regulation. Eric pointed out that invention could not occur without political liberty.
See his presentation at the conference.
Early in his career Eric worked at both Bell Labs and Xerox PARC.
Clearly Eric is one of the powerful players in the Information age. His comments all seemed reasonable. I just bought a copy of his 2014 book How Google Works.
See her presentation at the conference.As background here is a 2014 video that explains her point of view on Calm Technology. What she is advocating seems so logical and some of her solutions seem so obvious (after you see them) and nice you wonder why this isn’t part of standard design for everything in the information age. Calm technology is part of the general theme of saving time or even “creating” time for users, one of the topics in Bell Labs Network X vision.
5. Henry Markram is a professor of neuroscience at the Swiss Federal Institute for Technology. In 2005 he founded the Blue Brain project to develop a radically new approach to neuroscience - an algorithmic and digital reconstruction and simulation of the brain on supercomputers.
Here is Henry's 2009 TED talk that explains the brain simulation project. Henry's talk at this conference was an updated and extended version of his TED talk with amazing graphics of the simulations of different parts of the brain "fighting" each other. Frankly his talk was difficult for me to grasp because I do not have a background in neuroscience. So how to evaluate or even understand a simulation of the human brain at a biological level without knowing how the human brain operates in real life? I was deeply impressed by the scope of this project.
Shannon Amphitheater & IEEE Milestone Dedication
|Cropped View of Granite Stone in the Concrete Patio|
Because of the threat of rain, the dedication schedule was pulled up a few hours. This was facilitated by the app we had downloaded for the conference which had a lot of useful info like the revised schedule.
Marcus Weldon and Alon Orlitsky unveiled the IEEE Milestone. The men in green polo shirts unleashed the confetti. Like most attendees I had a glass of champagne for the toast yet my guest Steve Cuppy, a former Bell Labs employee, had got coffee before the dedication was rescheduled. We all celebrated with whatever beverage we had. We enjoyed the brief unicycle & juggling displays, a perfect tribute to Shannon. Then we waved to the drone taking a group photo.
Discussion of Claude Shannon's Life & Work
There was a presentation by Sergio Verdú of Shannon's life and work. It was followed by a "fireside chat" panel moderated by Leonard Kleinrock. Panel members were Elwyn Berlekamp, Bob Gallager, Yakov Sinai and Sergio Verdú. These men knew Shannon personally and are contributors to the field of Information Theory.
It was a unique pleasure to hear their first hand stories about Shannon and their answers to questions from the audience.
Watch the entire 1 hour 9 minute panel's fireside chat.
Also I found this 30 minute YouTube video produced by the University of California. It summarizes the importance of Shannon's work and gives a sense of the man himself. Some of the speakers in this conference like Elwyn Berlekamp, Irwin Jacobs and Elon Orlitsky appear in this video. It has interviews with other researchers and various video clips.
Another treat I found is a one hour video of the talk that Bob Gallager gave last year at MIT's seminar on Boole and Shannon. Bob, a long time MIT professor, used the method that he learned from Claude Shannon's 1952 talk on Creative Thinking to explain how Shannon arrived at Information Theory. There are some interesting points on George Boole in his talk. The theme is how to do research.
The stated goal is: This competition sponsored by Nokia Bell Labs aims to promote Information Theory as a discipline and more importantly as a toolset for emerging areas in information and communication technologies.
These were all serious proposals which were presented to the audience like a standard technical talk. They were judged by a panel selected from conference speakers sitting on the front row. The winners were announced at dinner.
This is one more way that Bell Labs remains linked to universities and emerging talent and creative ideas.
By design because of time limitations, it was not possible for a given person to see all of the research demos. Also one could not linger on a given demo because of the need to continue to the next demo. And the demos period coincided with the showcase viewing opportunity. This conference was focused on Information Theory. The demos gave some clues to the Bell Labs Future X Network perspective..
Here is the list of the demos:
My favorite demo was a demonstration of contextual searching where the user could resort the search space after seeing graphically how "far apart" and how key words in a search were linked together. The demo room, called the "Anomaly Room", had a floor to ceiling glass wall with a rear projector. It felt as if you were inside the search space. I thought it would be great to be able to rent such a room by the hour to experience the power of this augmented intelligence tool. Perhaps a virtual reality version of the search space would allow one to "touch" or "move" words and do a "physical search".
Other demos showed cooperating robots holding different sides of a tray and stopping a rolling marble in the center of the tray. That demo and the one showing the remote control of driverless cars showed the need for very low latency response times. When the latency times were varied, dramatic differences in real world control resulted. Real time control requires edge clouds to achieve low latency. Milliseconds matter!
Near the showcase area a few minutes after looking at the world's first transistor (It was made of germanium), I saw Morris Tanenbaum. Fortunately, I remembered a few details about his invention of the silicon transistor in 1954 from the book The Idea Factory, recommended above. Dr. Tanenbaum shared the story and we had a delightful conversation with him. He is a gentleman with vast knowledge and experience. He looks great at 87.
Shannon Effect and Human Digital Orchestra
After cocktails and before the dinner, we were treated to a unique performance of the Human Digital Orchestra entitled The Shannon Effect. This was a collaboration between Bell Labs and the local Stevens Institute of Technology.
The audience was excited sitting on folding chairs inside the performance tent with its white walls a kind of screen. We didn’t know what to expect.
In front of us were an orchestra, a piano player, a conductor, a small stage, and a control pit of a dozen or so people with earphones sitting in front of large control screens, like an orchestra pit of DJs.
The big picture explanation is: imagine a standard performance with another layer of computer analysis & projection artistically blended in. You could see, for example, a pixilated color projection of the sounds generated by the piano player subtly time shifted. It heightened your enjoyment of the music.
The third layer of the blend was having the audience activate the Shannon Effect app on their cell phones. We would move them as directed by the conductor and the control computer projected points of lights on the tent walls.
The only way to experience this performance is to see an eleven minute YouTube video.
My way of explaining the name: imagine a blend of three performances: Orchestra (the basic regular performance), Digital (a sensory transformative view of the regular performance produced by computers), and Human (a kind of audience participation performance as filtered by computers). This is not virtual reality, but augmented reality.
Of course, with my Bell Labs background I also thought about some additional ideas for audience participation during the performance. I also had some flash backs to Henry Markram’s brain simulation talk.
Claude Shannon, who was also a musician, would have loved the Human Digital Orchestra performance. It was science, technology, and art with a touch of whimsy.
As expected dinner was well organized with guests assigned to designated tables. I got the last seat at my table. I felt I had won some kind of lottery. I was seated next to Steve Capus, Executive Producer of CBS Evening News and his wife, Sophia Faskianos, a former NBC news producer. The three of us had a pleasant, wide-ranging discussion that flowed from topic to topic. There was little discussion of the conference because they had just arrived.
Shannon's work and its Legacy
The second day of the conference was more technical. Marcus delighted the researchers with his fun introductory statement to the effect that now the real work of the conference begins. He also continued to participate as on the first day with occasional, but insightful technical questions in the Q&A sessions.
The initial talk was a survey of Shannon's work and its Legacy by Vincent Poor & Michelle Effros. They alternated between them section by section. It was impressive to see the depth and scope of Shannon's achievements as applied to specific areas. It was a very enjoyable talk that put some meat on the bones. (See YouTube video.)
Technical talks on various aspects of Information Theory
I did not take notes, but did enjoy the big picture aspect of these technical talks. My takeaways were basic. Information theory has been applied to many more areas than I realized. The number of people in academia working on information theory was larger than I realized.
One general theme at the conference and this session was how well we have approached the Shannon Limit.
I did find a tutorial lecture by Michelle Effros on the topic Information Theory for Large Networks that gives a flavor of these talks technical level. The increasing complexity of networks is a hot area of Information Theory.
Muriel Medard gave a presentation entitled The Coding Matrix Reloaded. (See the YouTube Video.) It discussed Information Theory applied to approaches for survivability of data in networks. Some of the material was related to Muriel's talk in 2015 on Replacing Replication at a Science and Information conference.
Olgica Milenkovic's presentation was entitled Compressive Sensing - Theory and Practice. (See the YouTube Video.) This presentation was about the use of Information Theory in DNA analysis. And it reminds us that Shannon in his PhD thesis created an Algebra for Theoretical Genetics. Her 2015 talk on Coding for DNA-based Storage Systems is useful background.
Christopher Sims talked about rational inattention in economic theory. Its the idea that it is rational to limit the time spent making a decision. As background here is short article on that concept and how Information Theory can be used to link it to economics. This will give a flavor of the basic issue in Sims' talk. See the YouTube Video.
David Tse gave a talk on The Unreasonable Effectiveness of Information Theory. (See the YouTube Video.) As background you might enjoy his talk It's Easier to Approximate. Also his talk Information Theory: From Communication to DNA Sampling is excellent background material.
Another speaker Emmanuel Abbe was the winner of of the 2014 Bell Labs prize. His talk was Fundamental limits in Mining Social Data. (See the YouTube Video.) As background you may enjoy his 2015 talk on Fundamental Limit of Community Detection in Stochastic Block Models available on YouTube. His work is a key, for example, on how to distribute data to make 5G mobile networks function properly in real time.
Panel on the Future of Information Theory
There was a discussion of the tradeoffs of keeping researchers working in Information Theory departments vs. using Information Theory techniques in various other disciplines. Yes, everyone agreed that Information Theory had a future. The moderator, Thomas Marzetta, kept a tight rein on the time and the panel and the conference ended exactly on time.
See the lively discussion with Rudi Urbanke, Emre Telatar, Gerhard Kramer, Andrea Goldsmith and Alon Orlitsky.
It was a profound delight to be invited and to attend this conference. It was a unique opportunity to learn about the scope and impact of Information Theory, and to learn more about the mysterious Claude Shannon. It was a thrill to watch and even speak to some of the pioneers of the Information Age.
My other goal was to get a sense of the current state of Bell Labs and its future. I came away reassured and optimistic.
Leadership is the key to success. Leaders who have a clear vision based on a deep understanding of their field, who can communicate their vision, and who have the confidence and patience to listen carefully are virtually unstoppable. Marcus is one of those leaders. He makes it look effortless. He is also a natural talent scout.
On display at the conference were expressions of affection for Bell Labs, for what it stands for and for what it has meant to people on a personal level. Even a summer internship is remembered vividly decades later. This facilitates the flow of ideas and people with academia.
This conference triggered ideas. For example, instead of "Human Factors in Software Design" it seems we are headed to "borderless" relationships with computers. As Bell Labs tackles the big problems in telecommunications looking for disruptive solutions, how can they avoid inventing things that are applicable beyond telecommunications?
There is a fun 35 minute video recap of the two day conference.
The day after the Shannon Conference with my head spinning with new ideas and possibilities, I stopped at the Thomas Edison National Park in West Orange, NJ for a repeat of my 1966 visit. My visit to that Lab and to Edison's home and grave is covered in my Inventor's Loop report. Then I journeyed down memory lane for brief stops at the two places associated with the start of my career - Succassuna, NJ and Holmdel, NJ.
Succassuna, NJ (144 Route 10 West) is the location of the first commercial No 1. Electronic Switching System. This is ground zero of the whole revolution in telecommunications. Before No. 1 ESS became operational on May 30, 1965 every other telephone switch in the world was manual or electromechanical.
From the invention of the Transistor in 1947 and Information Theory in 1948, it took 17 years before the start of commercial ESS telephone service. Here is a brief article to explain the importance of and difficulty to develop No. 1 ESS, the foundational step on the path to cellular telephony, digital switching systems, the Internet, etc.
It was an initial shock to see Verizon on the New Jersey Telephone company building. But New Jersey Bell became part of Bell Atlantic at the time of the 1984 breakup of AT&T. And in 2000, Bell Atlantic became Verizon. Regardless of the corporate rebranding, the No. 1 ESS historical marker is still clearly visible. Note the 2 photos below.
In 1966, I was interviewed at Holmdel. And 3 months later I started my career with Bell Labs there. The building was sold many years ago. The good news is that the building is being repurposed and not torn down; a project called Bell Works. Holmdel was a huge building sheathed in glass when I started; it was enlarged later. It is located on 472 acres.
As I approached the property, I saw the first of 3 modest goals for my visit, the famous Holmdel water tower shaped like a transistor. There is a long drive providing the time to admire the enormous building designed by Eero Saarinen.
Thru skill and luck, I was able to enter the building for a brief visit. While the outside is being preserved, the inside is under partial re-construction on the cross aisles. However, the sunken lobby is still there intact. It created a dramatic effect in an enormous space with a 6 story ceiling by sinking the lobby a few feet. I was able to walk across the building on the first floor and in that general area only.
I had memories of the summer of 1966 when some of the windows were removed to allow trees to be placed inside on the first floor and to have large hanging plants on impressive circular planters suspended from the ceiling.
Today there are still Bell Labs historical photos on the walls in the lobby area. And the transformation is in progress to the building's new future. Ironically the first tenant is a start-up company. The designer of the Bell Works project, Paola Zamudio, discusses her four lessons in this article.
This video from 2018 shows the historic Bell Labs building is now a Metroburb, the town center of Holmdel.
I left Holmdel with a smile on my face and headed to Wilmington, DE to see the origin of the DuPont fortune in America. A hundred and fifty years later, DuPont was the source of purified silicon that Bell Labs and Morris Tanenbaum used to make the first silicon transistor.