Monday, September 28, 2020

Hiring: Two Ph.D., one Post-Doc position at TU Wien/Austria

Starting September 1, 2020 I'm a full professor for ubiquitous computing at TU Wien. I'm curious to continue my quest of investigating the boundaries between human and machine.
After some wonderful years with the Artificial & Human Intelligence research group at Siemens Corporate Technology in Berkeley I have moved on and I'm happy to be back in academia. I'm grateful for the experience from industry of making things real that matter. I was fascinated by the drive, passion and focus of my collaborators at Siemens.
Back in academia, I'm free again to investigate technologies with research questions in mind.
I'm looking for allies who are willing to explore
  • how can we build digital companions [1] supporting humans in their tasks?
  • which user-interfaces can we build in order to establish a seamless interaction between human and machine? [2]
  • which modalities are most suited to proactively support users in their tasks? [3]
I'm in the process of setting up a new research lab in the heart of downtown Vienna:
  • hardware lab: digital/physical tools in a workshop to create research prototypes
  • creation lab: workspace for students to code, build, create, experiment
  • user interaction lab: evaluation of prototypes, user studies, media-interaction
If you want to join the team please do have a look at my two open PhD and one post-doc position. I'm looking forward to your applications!

[1] Mareike Kritzler, Jack Hodges, Dan Yu, Kimberly Garcia, Hemant Shukla, and Florian Michahelles. 2019. Digital Companion for Industry. In Companion Proceedings of The 2019 World Wide Web Conference (WWW '19). Association for Computing Machinery, New York, NY, USA, 663–667. 
[2] Tian, R., Saran, V., Kritzler, M., Michahelles, F. and Paulos, E., 2019, October. Turn-by-wire: Computationally mediated physical fabrication. In Proceedings of the 32nd Annual ACM Symposium on User Interface Software and Technology (pp. 713-725).
[3] Funk, M., Marky, K., Mizutani, I., Kritzler, M., Mayer, S. and Michahelles, F., 2019, May. LookUnlock: Using Spatial-Targets for User-Authentication on HMDs. In CHI Conference on Human Factors in Computing Systems, Late Breaking Work.

Thursday, April 2, 2015

"Information is not knowledge. Knowledge is not wisdom. Wisdom is not truth."


Just recently I had the opportunity to attend the event Iot:: Empowering the Enterprise. The event was opened by a welcome note by Padmasree Warrior, CTO of CISCO, who predicted IoT as the next revolution of technology wave.
Then Wim Elfrink stepped onto the stage and started off that his kids would only have two statuses in life, asleep or alive. He questioned IoT being a hype and emphasized the fact that IoT could only be built in a joined effort. Furthermore he presented evidence that the selling of sensors would be exploding and sensors are being deployed everywhere. Only 1% is of our world today would be connected so far. He mentioned it would be less about the data more about how to transfer data to knowledge and wisdom. (Which reminded me after the famous Frank Zappa quote;) Surprising to me, Wim mentioned the public sector as the largest opportunity for sensors, he was calling for new governmental regulations to drive new opportunities. Finally, opened the stage for startup called relayr.io that provides connectivity to sensors bundled with an analytics platform.
Next on the agenda was panel discussion on the IoT eco-system. Debjit Mukerji from Siemens presented the notion of Web of Systems as an approach for moving from sensor-cloud analytics systems to decentralized, local information processing systems. Thus, IoT systems would not solely depend on connectivity to a cloud but could operate and reason locally in real time closer to the phenomenon. Debjit mentioned the importance of enterprises for startups to learn about industrial business cases, receive real-world feedback, access to testing resources and customers.
Amit Chaturvedy from Cisco Investments said the IoT would a global phenomenon wherof silicon valley startups would be mostly focusing on consumer applications, Germany would be leading in industrial developments, Asia about smart cities...different geographies would have different drivers.

From the startup pitches I recall pubnub allowing to connect devices in a secure manner, helium building networks on top of commercial wireless networks, edyn for monitoring plants (I can't help but hasn't Koubachi done than some time ago?), skyspecs making drones saving for industrial use, Ayla offering a network layer for remote monitoring, and placemeter quantifying the real world in real time using computer vision and video sensors.

VR Ferose from SAP sent out the message that connecting would be only the first step, the real value would be to learn from the data and most importantly provide enhanced customer experience. As an example he joked about settling down in the Bay area could be based on three apps google express, amazon prime and uber. IoT instead could go far beyond this experience provide new insights into predicitive maintenance and connected insurance. VR underlined his arguments by a remote control car demo where the premium rate has been adjusted to the driver's driving behavior. Personally, I was happy to see how earlier research [1] finally found its way into practice.
The last keynote on the program was delivered by Helmuth Ludwig who described integration of data during industrial manufacturing could be used the optimize performance and time to market. He mentioned the example of Omneo allowing to analyze events at real time. As you can't do physical prototypes for testing mars rover, simulation software can be used to eliminate prototypes, save resources, and increase efficiency. The future would rely on tools helping companies to produce in new ways, more efficiently and faster. For that reason, Siemens has opened its research lab in Berkeley to build bridges from the knowledge pool of universities and startups to the corporate world. As an example of this bridge he presented the video The Missing Link.

Overall, I enjoyed very much being at this event feeling the great passion for Internet of Things both in the corporate and the startup world. While I found a lot of belief that IoT will happen (e.g. presented by cisco), I still see more thoughts to be spend on innovative ideas about how IoT can really provide value across domains. Connecting millions, billions, trillions - whatever the large number should be - is just the first step. Just having access to more data does not really provide value in itself. 

How can we leverage these investments across completely different domains? How can a temperature sensors describe itself as a temperature sensor and team up with an air pressure close by to provide a more comprehensive picture and simultaneously calibrate itself using the data of another temperature sensor close by? Why would the owner of the temperature sensor allow access to his sensor, what could be an appropriate business model look like? How can we keep track of this connected layer, upgrade it and maintain it over time? How much can we achieve in the cloud, what do have to do locally, how can we transition effectively between cloud and edge?

To make IoT a reality in the enterprise there has to be more than collecting data, as Frank Zappa already recognized: "Information is not knowledge. Knowledge is not wisdom. Wisdom is not truth..." ;)

[1] Johannes Paefgen, Flavius Kehr, Yudan Zhai, Florian Michahelles: Driving Behavior Analysis with Smartphones: Insights from a Controlled Field Study, Proceedings of the 10th ACM SIGMOBILE Conference on Mobile and Ubiquitous Multimedia (MUM'12), Ulm, Germany, December 2012 [PDF].

Thursday, March 5, 2015

Summaries and Impressions: Internet of Things 2014 Conference, MIT, Oct 6-8, 2014

www.iot2014.orgIoT2014 is the continuation of IoT2008 (Zurich), IoT2010 (Tokyo), IoT2012 (Wuxi). This conference focuses on both academia and industry, it applies a rigorous review process to research papers (according to IEEE guidelines), features focused workshops on IoT subtopics as well as demo and poster session presenting early ideas.





The audience comprises a rather unique set of researchers, consultants, end-users, funding agency representatives and entrepreneurs.

Siemens’ Web of things research team has been the this year’s gold sponsor. This yielded an honorary mention of Siemens in the opening of the conference, a keynote speech (presented by K. Bettenhausen), sponsor booth and appearance on print material. Additionally, team members of WoT were visible beyond by organizing the web of things workshop and various talks.


The keynote of Kurt Bettenhausen (Siemens) highlighted the economic importance of manufacturing industry across the world. He expressed the need for innovation in order to keep up with the pace of innovation coming from China. Thus, in Germany the vision of Industrie 4.0 has been created which suggests products carrying all the data throughout the production process describing the steps for the product necessary to get produced. The driver for this vision are individual customer needs, dynamic design of engineer, new services, increased efficiency, and the importance of work-life balance of workers. After the lost battle in internet technologies, software and mobile services, it's the European industries' chance to make Industrie 4.0 a reality. This should allow companies to gain more flexibility in manufacturing control, reduction of energy and shorter time to market. In addition to the efforts and opportunities of big data analytics, there is much more knowledge available locally during the production process already which should be leveraged: why re-learning constraints from data analytics as they were already know during the design process. Questions from the audience focused on the performance of the Industrie 4.0 project progress in Germany and explicit examples of I4.0. The keynote speaker mentioned successful collaboration between research, industry and government but was not willing to disclose specific examples of I4.0.
David Clark (MIT) explained why the Internet has been so successful despite and because of the many design flaws and compromises. He started out with reflecting on the initial goals of the Internet back in 1975: development of a generic, global reaching, layered approach following network. In retrospect he admitted various flaws. Configuring wifi routers in today's home can be challenging as the design of the IP protocol would not care about configuration. Also the systems would neither know nor care what the user was doing and could not never tell if there was going something wrong. Furthermore, the Internet was never designed for mobility. Instead, the IP protocol was designed as a stupid network. As key learnings David derived,

If you build an ugly system, everybody says how ugly it is, if you build a well-defined system, nobody will ever look at it.
Thus, the consequences for today are that connectivity is not a cost issue anymore but configuration is. Initially the Internet has been designed to route packets, the rest was left up to the application, today the Internet is more about routing money, package routing is just a side-effect;)
An IoT panel discussion has been lead by Alex Ilic from the Auto-ID Labs. The keypoints raised by the panelists Sanjay Sarma (MIT), Elgar Fleisch (ETH Zurich/HSG) and Scott Jenson (Google) were interesting and diverse. The Internet of Things should develop successfully in specific verticals first and then only derive generic principles for cross domain platforms. This would counter the "European" approach of EU-funded projects which have spend time and money on developing bullet proof and shiny solutions which have a hard time of adoption in any domain. There was, however, no consensus which approach would succeed, at least the Internet has followed build-it/fix-it approach;). History of industry has shown that whenever a dominant design [1] has evolved, the industry became successful. For the IoT we are still missing this dominant design, who could lead this effort, IT players, or verticals? Finally, healthcare was mentioned to bear the greatest potential for IoT, as here it is not just about money but about a huge societal problem. Healthcare is so expensive today and differences between developed and developing world are tremendous. IoT would have the potential to link information and devices, provide access, empower people, drive costs, and challenge expensive equipment, e.g. MRI's, with simpler sensors embedded in connected billions of phones.
Shouman Datta (MIT) built upon the argument of the previous panel discussion that IoT should not be just about money but even more about societal challenges. The grand challenge would be to build a platform that creates an ecosystem for other to follow. The vision of the IoT would be rather old anyway. Shouman mentioned projects like 'pay per pee', monitoring at the toilet, odd projects like 'print your face' and others. The ultimate potential for IoT in healthcare will be to connect a plethora of dumb disconnected device of today, increase patient safety and increase building a more comprehensive understanding what's going in the human body for longer periods of time.
Overall, the IoT conference has been a very exciting event. It has provided the unique mixing researchers and practitioners which is hard to find. I'm very much looking forward to the 2015 edition in Korea: www.iot2015.org
[1] Mastering the Dynamics of Innovation, Utterback (1994) Harvard Business School Press ISBN 0-87584-342-5 Library call numbers HD58.8.U87 1994 658.4'06—dc20, p 24.

Wednesday, June 26, 2013

IoT has arrived in industry

Eight years from now I started with the Auto-ID Labs at ETH Zurich. Our mission statement has been "architecting the Internet of Things". The underlying principle was to build physical presences of physical objects, i.e. digital counterparts that would be linked to physical things through RFID, barcodes, or any other kinds of identifiers [1].
For specifying the architecture we investigated the application using RFID against anti-counterfeiting. We developed various mechanisms to verify the authenticity of a product based on its identifier, history of trace and physical fingerprints [2]. Simultaneously, we recognized the tremendous potential of not only linking supply-chain partners to the Internet of Things but to also develop new channels to consumers and to also captcollect crowd-sourced data from them [3]. While NFC technology has been not ready for this for a while we started with mobile barcode reader software first [4].
In addition to the Auto-ID Labs' research agenda of how to make things, products and objects part of the Internet, other communities, such as Web-of-Things researchers, have started exploring how to reuse existing web-standards, e.g. REST, or to expand to new standards, e.g. CoaP, in order to let addressing and communications with the Internet reaching out to more resource-constrained devices and sensors [5].
Meanwhile, governments (e.g. Germany and China) and industry have recognized the Internet of Things as a new paradigm for upgrading their sectors and products from mere physical objects to networked items coupled with services. Several terminologies, such as Industrial Internet [6], Internet of Things and Services, Social web of things, Internet of everything [7], describe attempts of occupying space in the emerging Internet of Things through proprietary and still mainly closed developments.
The success of the Internet was mainly based on the absence of interest of industrial partners in the early days, such that a few academics could set out the communication principles of TCP/IP which was designed to be open and flexible on future developments on the top in terms of new applications and from the bottom regarding new communication technologies and protocols.
The development of the Internet of Things, instead, is challenged various powerful players who aim at building their own platforms separat from others. Here, I believe in the evolution of developments as we could observe from the MNO's: today, all walled gardens have disappeared and mobile users can access the Internet regardless of their provider.
The power of the Internet of Things lies in the integration of information and sensor data of various sources, the re-use of the same sensor data for various applications, and the creativity of 3rd party developers as the app-store models have proven successful.
The Internet of Things has arrived in industry, I'm very excited about how the various players will manage their transformation from platform owner, hardware developers with long development cycles, and contracted partner networks to agile software developers that release beta-products and stay in a continuous dialogue with their customers.

At this point I'm very happy to hand-over the lead of the Auto-ID labs to Alexander Ilic. Alexander has a strong background in retail applications, investigating offerings of new services to retail clients via digital receipts (see cosibon). Simultaneously, I'm very happy for the opportunity to drive forward the development  of the Internet of Things in industry!

[1] D. Uckelmann, M. Harrison, & F. Michahelles(Eds.), Architecting the Internet of Things. Berlin, Germany: Springer. ISBN 978-3-642-19156-5
[2] Lehtonen, M.; Michahelles, F.; Fleisch, E. : Trust and Security in RFID-based Product Authentication Systems. IEEE Systems Journal, Special Issue on RFID Technology: Opportunities and Challenges, First Quarter of 2008.
[3] S. Karpischek, F. Michahelles, E. Fleisch, my2cents – enabling research on consumer-product interaction, to appear in Special Issue on "Smartphone Applications and Services for Pervasive Computing", Journal of Personal and Ubiquitous Computing - Springer, 2011, ISSN 1617-490,Doi: 10.1007/s00779-011-0426.
[4] F. von Reischach, S. Karpischek, R. Adelmann, F. Michahelles: Evaluation of 1D barcode scanning on mobile phones, Internet of Things 2010 Conference (IoT2010), Tokyo, Japan, November - December 2010




[5] Vlad Trifa, Dominique Guinard, Simon Mayer: Leveraging the Web for a Distributed Location-aware Infrastructure for the Real World.In: Erik Wilde, Cesare Pautasso (Eds.): REST: From Research to Practice. Springer, ISBN 978-1-4419-8302-2, pp. 381-400, New York, 2011
Peter C. Evans, Marco Annunziata: Industrial Internet, GE, Nov 26, 2012. [pdf]
[7] Dave Evans: The Internet of Everything, Whitepaper, Cisco, 2012.

Tuesday, May 7, 2013

2nd CfP: Workshop on Pervasive Technologies in Retail Environments at UbiComp 2013

PeTRE - Workshop on Pervasive Technologies in Retail Environments

organized by
Markus Löchtefeld, Petteri Nurmi,Florian Michahelles, Carsten Magerkurth, Patrik Floréen and Antonio Krüger
Zürich, Switzerland, September 8th, 2013

 

About

The workshop on Pervasive Technologies in Retail Environments (PeTRE) is the continuation of MIRE 2011 (http://dfki.de/mire held at MobileHCI 2011) and provides an established forum for researchers from academy and industry exploring how pervasive technologies can be embedded into retail environments to create new shopping experiences and services. The goals of the workshop are to
  • discuss the integration of pervasive technologies into retailing;
  • construct a roadmap for future integration of pervasive technologies in retailing
  • strengthen and extend the community of this rapidly developing field.
We invite both case studies discussing real-world deployments as well as original research contributions in one of the following topic areas:
  • mobile applications for retail contexts,
  • location sensing and customer flow analysis
  • pervasive public displays
  • personalization and user/customer modeling for the retail domain
  • human product interaction
  • infrastructure standards
  • integration and interaction with robots and drones
  • infrastructure combining eCommerce with brick and mortar stores
  • social media in the retail context,
  • business models and social impact.

 

Workshop format and submissions

The workshop on Pervasive Technologies in Retail Environments (PeTRE) is a full-day workshop with an extended thematic scope and aims to bring together researchers and practitioners from academia and industry with multidisciplinary insights being explicitly encouraged. We ask for papers (two-pages position-statement-paper or four-pages research-paper) that address one or more of the research questions mentioned above, or that describe findings that relate to these research questions based on systems the authors have built.
We highly encourage bringing a demonstration of your prototype if available and presenting it at the Workshop!
Papers should be formatted according to the Ubicomp 2013 Workshop Format. (http://www.ubicomp.org/ubicomp2013/calls/templates.php please remove the copyright notice). Papers should be submitted using the electronic submission system available on the website (http://www.pervasive-retail.com). At least one author of an accepted paper is required to register for the workshop.
Each paper will be reviewed by at least two members of an international program committee. All accepted papers will be made available online on the workshop website as well as in the ACM Digital Library.

 

Important dates

May 20, 2013:          Submission Deadline
June 3, 2013:          Author Notification
June 20, 2013:         Submission of camera-ready version
September 8, 2013:     Workshop in Zurich, Switzerland

 

Website and more information

http://www.pervasive-retail.com
If you have any additional questions please contact us via email:
markus.loechtefeld@dfki.de

 

 Organizers

Markus Löchtefeld, German Research Center for Artificial Intelligence (DFKI), Germany
Petteri Nurmi, Helsinki Institute for Information Technology (HIIT), Finland
Florian Michahelles, Auto-ID Labs, ETH Zurich, Switzerland
Carsten Magerkurth, SAP (Switzerland) AG, Switzerland
Patrik Floréen, Helsinki Institute for Information Technology (HIIT), Finland
Antonio Krüger, German Research Center for Artificial Intelligence (DFKI), Germany

Tuesday, April 23, 2013

Streetspotr - the mechanical turk in the wild

As part of my lecture Businees Aspects of the Internet of Things I could welcome Werner Hoier and Dorothea Utzt from Streetspotr.com as guest lecturers (slides).
They shared some interesting insights about their start-up which implements the idea of match-making between location-based micro-task issuing and fullfilling (as described earlier by [1]). Having reached out to 180k users who are only waiting for earning some fractions of a EURO by solving tasks like "take a picture of the menu in restaurant", "collect the GPS location of a parking lot", or "check the availability of an ATM". Without marketing but mostly relying on public media and blogs who featured their story they could build up their user base.
Interesting enough they are no longer looking for more users but rather for more job issuers. Also, so far they refrained from allowing to share user-to-user services. Their service nicely extends traditional services of market survey agencies. They reported about their most successful user who already earned a 600 EUR on their platform.

I'm really curious how this micro-task business evolves. I could well see opportunities for people waiting or being stuck somewhere and either earn money or giving something back by solving small little tasks in their downtime. Furthermore, finding a proper alignment between situation of the user and available tasks and corresponding rewards should be key to success, see also [2].

[1] Florian Alt, Alireza Sahami Shirazi, Albrecht Schmidt, Urs Kramer, and Zahid Nawaz. 2010. Location-based crowdsourcing: extending crowdsourcing to the real world. In Proceedings of the 6th Nordic Conference on Human-Computer Interaction: Extending Boundaries (NordiCHI '10). link
[2] Navkar Samdaria, Akhil Mathur, and Ravin Balakrishnan. 2012. Paying in kind for crowdsourced work in developing regions. In Proceedings of the 10th international conference on Pervasive Computing (Pervasive'12), Judy Kay, Paul Lukowicz, Hideyuki Tokuda, Patrick Olivier, and Antonio Krüger (Eds.). link

Thad Starner - the personalized Augmented Human



I was lucky to visit the closing keynote of Thad Starner at the 3rd Augmented Human Conference held in Stuttgart in March 2013.

As part of Thad's talks he pulled out his favorite 2-second rule which says: the longer accessing a device exceeds 2s, the more its actually usage would decrease exponentially. Thus, he made a claim that wrist watch interface always sitting on one's wrist ready to use should be more successful than mobile phones which have to pulled out of the pocket. He showed a compelling example of implementing a pie-chart interface using the 12 positions of an analog watch to "dial into apps" [1]. 

Then he showed the promo video on google glasses, which uses voice control to record video and display related information.
He ended his talk with teaching muscle memory applied to the example of how to learn to play piano (well, or at least to play some one-handed melodies;). Finally, he showed some working examples of brain-computer interfaces which generally perform very slow in allowing users to convey control information at only about 1bit per second. However, he could show how to pick up American sign language gestures from reactions of the motor cortex. His final vision would be to allow patients without movements to communicate with their environment.

Overall, I was intrigued by Thad's lecture style: instead of selling his research he puts himself into the position of a critical spectator of his own research who gets excited and convinced by his own results. Indeed, a smart way of presenting.

[1] Daniel Ashbrook, Kent Lyons, and Thad Starner. 2008. An investigation into round touchscreen wristwatch interaction. In Proceedings of the 10th international conference on Human computer interaction with mobile devices and services (MobileHCI '08).