Lecture 10: Future trends

This lecture, which I hosted at the from the University Carlos III of Madrid, Spain, recaps the whole 2013 series and looks at the future of AI, Robotics and …the Shanghai Lectures themselves :-)

“How the body shapes the way we think – principles and insights ” by Rolf Pfeifer

In yet another great lecture, Rolf summarizes the main ideas and concepts we have reviewed in the Shanghai Lectures. Check here for his thought-stimulating insights:

Guest lecture: “Autonomous Service Robots that Can Ask for Help” by Manuela Veloso

In this guest lecture Prof. Manuela Veloso, very appropriately H. A. Simon Professor at Carnegie Mellon University, gives us a compressed survey of her research. She also shows us  how we may work around current limitations of AI and Robotics by simply letting robots ask for our help! What does she mean by this? Watch her lecture here:

Manuela Veloso is President of AAAI, past president of the RoboCup Federation, and a fellow of IEEE, AAAI and AAAS.

That’s all, folks, for 2013! Stay tuned, the best is yet to come!

About the ShanghAI Lectures

While in the classical approach “intelligence” was essentially viewed as information processing taking place in the brain, the more recent insight that interaction with the environment is of central importance is gaining acceptance. This has led to the metaphor of embodiment, i.e., that intelligence is always a property of an entire organism — an idea that has far-reaching implications and often leads to surprising insights, but which has not so far been widely exploited in industry practice.

The ShanghAI Lectures project aims to:

• Build a sustainable community of students and researchers in the area of Embodied Intelligence
• Make education and knowledge on cutting-edge scientific topics accessible to everyone
• Explore novel methods of knowledge transfer
• Overcome the complexity of a multi-cultural and interdisciplinary learning context
• Bring global teaching to a new level

These lectures about Natural and Artificial Intelligence have been held via videoconference at the University Carlos III of Madrid in Spain, the University of Zurich in Switzerland, Scuola Superiore Sant’Anna of Pisa, Italy, Humboldt University Berlin in Germany, University of Plymouth and University of Salford in the UK, and 10 other universities around the globe. Students from the participating universities are still working together on the exercises, using Webots by Cyberbotics, and Ludobots by the University of Vermont.

The lectures have also been streamed to allow remote participation to anybody.

The ShanghAI Lectures differ from ‘conventional’ MOOCs as they exploit telecommunication technology to build a global, distributed lecture hall that allows rich interaction rather than simply implementing the good old fashioned TV broadcasting model on a different medium. They also differ from other AI courses as they propose a new paradigm approach to embodied cognition (a.k.a. AI and Robotics). It is a kind of Copernican revolution with respect to GOFAI and its robotics application — and thus a research program for the coming decades.

This year I coordinated the lectures, with help from Prof. Rolf Pfeifer and Dr. Nathan Labhart at the University of Zurich . Rolf Pfeifer and I provided the context (introduction, moderation, and conclusion). As always, there were 2-3  invited guest lectures each week.

The lay notion of how human communication works is, basically, that we exchange packages of information that are encoded and then decoded in our heads using language. But it just takes a little observation to see that this “computer metaphor” just doesn’t apply to the human way of communicating. We can’t reduce communication to a transfer of abstract information, and the same happens to be true of human-machine communication if we expect it to be of any use.

Think of turn-taking during a conversation. Speakers and listeners switch roles very quickly, often faster than typical reaction times. Silences between turns last less than 300 milliseconds, leading us to believe that a lot more is going on than simply waiting for the information to arrive in order to decide the next step. In fact, a lot of anticipatory planning is happening in both the speaker and listener. There’s a continuous monitoring of the other in search of subtle cues as to when the turn will be yielded. Such cues include changes of prosody, gestures, gaze shifts toward the listener, implicit verbal cues such as filler utterances…. Turn-taking is just one example of embodiment, a critical dimension of communication that most classical models ignore.

In fact it was the the study of human language that led the way toward the revolution of embodiment in the cognitive sciences. Cognition, in this view, unfolds in time while we perceive and act. Cognitive processes are continuously linked to sensory information and continuously coupled to motor systems that impact on the sensed world.

Communication is embodied in the obvious sense that natural communication occurs between people with physical and physiological bodies, and their cognitive abilities originate from processes in their nervous systems. But the embodiment stance to human communication is deeper than that. It extends the scientific approach to human language by adding a new emphasis on the grounding of language in the ongoing perception of the scene.

We begin to understand the emergence of language during development as a consequence of the immersion of the child in an environment shared with speakers. Development helps us explain how gesture and other non-linguistic cues are tightly coupled into language processing in the here and now of human communication. Embodied communication is also essential for researchers building artificial cognitive systems that autonomously generate language to communicate with a human and, conversely, make sense of human language and gesture by tying it to the system’s own scene perception.

Researchers exploring these different approaches to the study of embodied communication will share their ideas and insights during the Fifth EUCogIII Members Conference in Bochum 19-20 March 2014. Peter Dominey from the INSERM Stem Cell and Brain Research Institute in Lyon, France will show, for instance, how the iCub robot acquires “meaning” by establishing a link between physical action and language, and Yiannis Aloimonos from the University of Maryland in the USA will demonstrate how an “action grammar” is able to parse human action into meaningful sequences. Psychologists like Larissa Samuelson and Susan Wagner Cook from the University of Iowa will present the developmental perspective on embodied communication, focusing on how children learn words and how gesture and language co-develop. Srini Narayana will show in theoretical models how meaning and grammar are intimately tied together in the learning process.

Our aim is to advance the interdisciplinary interchanges between artificial and human cognition research fields. Both technical and natural aspects of communication will therefore be represented at the conference. To develop cognitive systems we need to understand the variety of ways that speech and gesture are coordinated, and develop technical aspects like creating automatic semantic descriptions of complex human activity from perceptual data. Knowing more about how humans really communicate and make sense of human communication is important if we want to create machines that are smart like us.

EUCog is the European Network for the Advancement of Artificial Cognitive Systems, Interaction and Robotics.

Lecture 4 – Evolution: Cognition from scratch

If you believe that it’s too difficult to develop the mathematical models required to arrive at a new concept of embodied AI and robotics, this lecture will offer you hope. In her talk, Prof. Verena Hafner, from Humboldt University in Berlin, discusses the concept of ‘body maps’ from an embodied intelligence perspective, and shows how they can emerge from sensorimotor flow, through information distances in the data flow and/or an internal simulation. In this context the ‘body map’ is not a symbolic ‘representation’ of the agent’s body, but rather an emerging structure used by the agent in its interaction with its environment.

“Creating body maps” by Verena Hafner

“Evolution: cognition from scratch” by Rolf Pfeifer

In the second talk, Rolf Pfeifer shows how evolutionary processes at different time scales drive the development of embodied intelligence in nature, and how they can be mimicked (with surprisingly good results) for the development of artificial systems, including … pipes and antennas!

How can we exploit Gene Regulatory Networks for the development of a new paradigm for robots?

Guest Lecture – “How can we test for artificial life? Complexity and computation in Nature” by Hector Zenil

An unlikely yellow ‘computer’?

The skeptics will think that certain concepts such as ‘life’ or ‘self-organization’ are too  elusive and ill-defined to be useful to AI. Hector Zeni, from the Karolinska Institute in Stockolm, Sweden, does cutting-edge research on the applications of Algorithmic Information Theory to the modeling of life and cognition. Witness how AIT might come to the rescue with this lecture, also hosted by Humboldt University Berlin …

About the ShanghAI Lectures

While in the classical approach “intelligence” was essentially viewed as information processing taking place in the brain, the more recent insight that interaction with the environment is of central importance is gaining acceptance. This has led to the metaphor of embodiment, i.e., that intelligence is always a property of an entire organism — an idea that has far-reaching implications and often leads to surprising insights, but which has not so far been widely exploited in industry practice.

The ShanghAI Lectures project aims to:

• Build a sustainable community of students and researchers in the area of Embodied Intelligence
• Make education and knowledge on cutting-edge scientific topics accessible to everyone
• Explore novel methods of knowledge transfer
• Overcome the complexity of a multi-cultural and interdisciplinary learning context
• Bring global teaching to a new level

These lectures about Natural and Artificial Intelligence have been held via videoconference at the University Carlos III of Madrid in Spain, the University of Zurich in Switzerland, Scuola Superiore Sant’Anna of Pisa, Italy, Humboldt University Berlin in Germany, University of Plymouth and University of Salford in the UK, and 10 other universities around the globe. Students from the participating universities are still working together on the exercises, using Webots by Cyberbotics, and Ludobots by the University of Vermont.

The lectures have also been streamed to allow remote participation to anybody.

The ShanghAI Lectures differ from ‘conventional’ MOOCs as they exploit telecommunication technology to build a global, distributed lecture hall that allows rich interaction rather than simply implementing the good old fashioned TV broadcasting model on a different medium. They also differ from other AI courses as they propose a new paradigm approach to embodied cognition (a.k.a. AI and Robotics). It is a kind of Copernican revolution with respect to GOFAI and its robotics application — and thus a research program for the coming decades.

This year I coordinated the lectures, with help from Prof. Rolf Pfeifer and Dr. Nathan Labhart at the University of Zurich . Rolf Pfeifer and I provided the context (introduction, moderation, and conclusion). As always, there were 2-3  invited guest lectures each week.

Lecture 3 – Cognition and embodiment – Language learning in children and robots: A developmental robotics approach 

This week’s lecture discusses an important problem: how ‘symbols’ and language emerge and are ‘grounded’ from the flow of sensory-motor data. In this lecture, Angelo Cangelosi, Director of the Centre for Robotics and Neural Systems at the University of Plymouth, explains the developmental approach to robotics (and psychology) and offers a remarkable example of ‘synthetic methodology’  for the study of children’s intellectual development.  Cangelosi, who’s original background is in psychology, is a leading researcher in developmental robotics.

Do you want to know what ‘gavagai’ means? And why this is an important issue? Watch the video:

Notice the experiments on the emergence of natural number from sensory-motor data flow.  Aren’t they intriguing? Do you remember ‘Where mathematics come from?’ by Lakoff and Nuñez? You may also check Nuñez’ s lectures in the ShanghAI archive!

This year the ShanghAI Lectures have been more ‘distributed’ than they had been in the past. While Rolf and I have developed the main ideas for the embodied cognition approach to AI and Robotics following the approach described in ‘How the Body Shapes the Way We Think’, in our text book we have devoted significant space to describing overlapping perspectives of embodied cognition from different researchers. Some lectures have been ‘hosted’  sites other than Madrid or Zurich. This was the first lecture broadcast by the University of Plymouth.

Guest Lecture – “Does Robotics Really Need Embodiment?” by Vincent Mueller

What we usually refer to as ‘embodiment’ is not the trivial concept that ‘intelligence needs a body’. The epistemological problems raised by the ‘embodiment’ and the ‘synthetic methodology’ approaches are not trivial either. In this talk, Vincent Mueller plays the devil’s advocate:

Vincent Mueller is a philosopher with the Anatolia College Thessaloniki and the University of Oxford. He is the coordinator of euCogIII, the largest and most important European network cognition science researchers. With almost 1000 members, euCogIII spans many disciplines, including AI, neuroscience and philosophy.

Guest Lecture – “Morphological Computation and Soft Robotics” by Helmut Hauser

A central aspect of ’embodied intelligence’ or ’embodied cognition’ is the the fact that part of the ‘control’ is offloaded to the physical dynamics of the agent, in what is called  ‘morphological computation’. In order to exploit morphological computation in the development of artificial cognitive agents, and to better understand its role in natural intelligence, quantitative models are needed. In this lecture  Helmut Hauser proposes a possible model, and also shows how issues related to morphological computation are intertwined with soft robot modeling.

Helmut Hauser is with the AI Lab of the University of Zurich and is one of the few people who dares to look for quantitative models of morphological computation (another one is …myself :-)).

About the ShanghAI Lectures

While in the classical approach “intelligence” was essentially viewed as information processing taking place in the brain, the more recent insight that interaction with the environment is of central importance is gaining acceptance. This has led to the metaphor of embodiment, i.e., that intelligence is always a property of an entire organism — an idea that has far-reaching implications and often leads to surprising insights, but which has not so far been widely exploited in industry practice.

The ShanghAI Lectures project aims to:

• Build a sustainable community of students and researchers in the area of Embodied Intelligence
• Make education and knowledge on cutting-edge scientific topics accessible to everyone
• Explore novel methods of knowledge transfer
• Overcome the complexity of a multi-cultural and interdisciplinary learning context
• Bring global teaching to a new level

These lectures about Natural and Artificial Intelligence have been held via videoconference at the University Carlos III of Madrid in Spain, the University of Zurich in Switzerland, Scuola Superiore Sant’Anna of Pisa, Italy, Humboldt University Berlin in Germany, University of Plymouth and University of Salford in the UK, and 10 other universities around the globe. Students from the participating universities are still working together on the exercises, using Webots by Cyberbotics, and Ludobots by the University of Vermont.

The lectures have also been streamed to allow remote participation to anybody.

The ShanghAI Lectures differ from ‘conventional’ MOOCs as they exploit telecommunication technology to build a global, distributed lecture hall that allows rich interaction rather than simply implementing the good old fashioned TV broadcasting model on a different medium. They also differ from other AI courses as they propose a new paradigm approach to embodied cognition (a.k.a. AI and Robotics). It is a kind of Copernican revolution with respect to GOFAI and its robotics application — and thus a research program for the coming decades.

This year I coordinated the lectures, with help from Prof. Rolf Pfeifer and Dr. Nathan Labhart at the University of Zurich . Rolf Pfeifer and I provided the context (introduction, moderation, and conclusion). As always, there were 2-3  invited guest lectures each week.

This week we publish the second of the

Lecture 2 – Embodied Intelligence

“Evolution embodies information in every part of every organism. … This information doesn’t have to be copied into the brain at all. It doesn’t have to be “represented” in “data structures” in the nervous system. It can be exploited by the nervous system, however, which is designed to rely on, or exploit, the information in the hormonal systems just as it is designed to rely on, or exploit, the information embodied in your limbs and eyes.” – D.Dennett, Kind of Minds, 1996

 

Many people will probably agree with Dennett from a philosophical standpoint, but what are the implications for robotics research and robot development?

There are many: What is the Frame-of-Reference problem? How we can get rid of it in principle? How can we get rid of it in practice?

Unfortunately,  the real world doesn’t come with labels. How can we associate symbols with a continuous stream of sensory stimulation? What is the cognitive role of physical interaction of the agent with the world? Check Rolf’s lecture to inspire your thoughts.

Guest Lecture – “From Bio-Inspired to Institutional-Inspired Collective Robotics” by Pedro Lima

The implications of the embodied intelligence paradigm are wide and deep.

In this talk Pedro Lima, from the Instituto Superior Tecnico in Lisbon, Portugal, shows how  human institutions can be seen as emerging processes in swarms of embodied intelligent agents, a.k.a. ‘robots’. This view has some implication on the economic sciences, too. In Portugal they are sensitive to ‘economics paradigms’ and know that current ones have some flaws :-). On the other hand an ‘institutional-inspired’ approach seems quite beneficial also for network robot development.

Intriguing, right? Watch the full lecture here:

Pedro Lima, is a professor at IST and a leading researcher in robotics and AI, and above all he is a nice guy and a subtle and deep thinker!

About the ShanghAI Lectures

While in the classical approach “intelligence” was essentially viewed as information processing taking place in the brain, the more recent insight that interaction with the environment is of central importance is gaining acceptance. This has led to the metaphor of embodiment, i.e., that intelligence is always a property of an entire organism — an idea that has far-reaching implications and often leads to surprising insights, but which has not so far been widely exploited in industry practice.

The ShanghAI Lectures project aims to:

• Build a sustainable community of students and researchers in the area of Embodied Intelligence
• Make education and knowledge on cutting-edge scientific topics accessible to everyone
• Explore novel methods of knowledge transfer
• Overcome the complexity of a multi-cultural and interdisciplinary learning context
• Bring global teaching to a new level

These lectures about Natural and Artificial Intelligence have been held via videoconference at the University Carlos III of Madrid in Spain, the University of Zurich in Switzerland, Scuola Superiore Sant’Anna of Pisa, Italy, Humboldt University Berlin in Germany, University of Plymouth and University of Salford in the UK, and 10 other universities around the globe. Students from the participating universities are still working together on the exercises, using Webots by Cyberbotics, and Ludobots by the University of Vermont.

The lectures have also been streamed to allow remote participation to anybody.

The ShanghAI Lectures differ from ‘conventional’ MOOCs as they exploit telecommunication technology to build a global, distributed lecture hall that allows rich interaction rather than simply implementing the good old fashioned TV broadcasting model on a different medium. They also differ from other AI courses as they propose a new paradigm approach to embodied cognition (a.k.a. AI and Robotics). It is a kind of Copernican revolution with respect to GOFAI and its robotics application — and thus a research program for the coming decades.

This year I coordinated the lectures, with help from Prof. Rolf Pfeifer and Dr. Nathan Labhart at the University of Zurich . Rolf Pfeifer and I provided the context (introduction, moderation, and conclusion). As always, there were 2-3  invited guest lectures each week.

This week marks the launch of the ShanghAI Lectures 2013 Edition on Robohub — we will release a new lecture from this series every Monday until the series is complete. Please use the comments section below to send us your questions, and we will do our best to respond! You can learn more about the ShanghAI lectures here.

Lecture 1 – Intelligence: Things can be seen differently

If you are not interested in precisely predicting the motion of a handful of ‘anomalous’ stars (nowadays known as planets), then there is no reason to assume that the earth is not sitting at the center of the universe. Ptolemaic astronomy was good astronomy, after all. And if you were an astronomer of the XVI century you may have had good reason to think that ‘solving astronomy’ was simply a matter of adding the ‘right’ epicycles ;-). Many did. Why worry about just five stars that do not model correctly when there are several thousand stars that are perfectly predictable?

AI and Robotics have been very successful in the past fifty years, thanks to Google, IBM Watson, SIRI or Da Vinci, Opportunity, Bosto … oops Google again ;-). Yet something is still missing …

With all the due ‘caveats’, the first lecture of the 2013 ShanghAI series insinuates the doubt that we might be in a similar situation as regards AI and Robotics … that a paradigm change might be necessary in order to be able to build ‘real’ robots that are more similar to those dreamed by Hollywood, and that this new paradigm should be more similar to the massively multi-agent, self organizing, biological processes, exploiting ‘morphological computation’ than to the Good Old Fashioned top down ‘Cartesian’ paradigm, maybe with some machine learning in addition (modern ‘epicycles’?).

Are you intrigued? Do you wildly disagree? Check out the first lecture and learn about all the 2013 participants in the 2013 inaugural lecture:

About the ShanghAI Lectures

While in the classical approach “intelligence” was essentially viewed as information processing taking place in the brain, the more recent insight that interaction with the environment is of central importance is gaining acceptance. This has led to the metaphor of embodiment, i.e., that intelligence is always a property of an entire organism — an idea that has far-reaching implications and often leads to surprising insights, but which has not so far been widely exploited in industry practice.

The ShanghAI Lectures project aims to:

• Build a sustainable community of students and researchers in the area of Embodied Intelligence
• Make education and knowledge on cutting-edge scientific topics accessible to everyone
• Explore novel methods of knowledge transfer
• Overcome the complexity of a multi-cultural and interdisciplinary learning context
• Bring global teaching to a new level

These lectures about Natural and Artificial Intelligence have been held via videoconference at the University Carlos III of Madrid in Spain, the University of Zurich in Switzerland, Scuola Superiore Sant’Anna of Pisa, Italy, Humboldt University Berlin in Germany, University of Plymouth and University of Salford in the UK, and 10 other universities around the globe. Students from the participating universities are still working together on the exercises, using Webots by Cyberbotics, and Ludobots by the University of Vermont.

The lectures have also been streamed to allow remote participation to anybody.

The ShanghAI Lectures differ from ‘conventional’ MOOCs as they exploit telecommunication technology to build a global, distributed lecture hall that allows rich interaction rather than simply implementing the good old fashioned TV broadcasting model on a different medium. They also differ from other AI courses as they propose a new paradigm approach to embodied cognition (a.k.a. AI and Robotics). It is a kind of Copernican revolution with respect to GOFAI and its robotics application — and thus a research program for the coming decades.

This year I coordinated the lectures, with help from Prof. Rolf Pfeifer and Dr. Nathan Labhart at the University of Zurich . Rolf Pfeifer and I provided the context (introduction, moderation, and conclusion). As always, there were 2-3  invited guest lectures each week.