Journal of Computer and Communications, 2014, 2, 38-44
Published Online September 2014 in SciRes.
How to cite this paper: Diamant, E. (2014) Cognitive Robotics: For Never Was a Story of More Woe than This. Journal of
Computer and Communications, 2, 38-44.
Cognitive Robotics: For Never Was a Story of
More Woe than This
Emanuel Diamant
VIDIA-mant , KiriatOno, Israel
Received Ju ly 2014
We are now on the verge of the next technical revolutionrobots are going to invade our lives.
However, to interact with humans or to be incorporated into a human “collective” robots have to
be provided with some human-like cognitive abilities. What does it mean?Nobody knows. But,
robotics research communities are trying hard to find out a way to cope with this problem. Mean-
while, despite abundant funding these efforts did not lead to any meaningful result (only in Eu-
rope, only in the past ten years, Cognitive Robotics research funding has reached a ceiling of 1.39
billion euros). In the next ten years, a similar budget is going to be spent to tackle the Cognitive
Robotics problems in the frame of the Human Brain Project. There is no reason to expect that this
time the result will be different. We would like to try to explain why we are so unhappy about this.
Cognitive Robotics, Inform ati on, Ph ysical Infor ma tio n, Semantic Information
1. Introduction
From the beginning, it was a fascinating idea: to create human-like living beings that would help and assist us in
our tedious everyday duties. The history has preserved many famous stories about such undertakings Pygma-
lion and his Galatea, Talos the guard of Crete (both from the ancient Greek mythology), Maharal’s Golem from
the late 16th century Prague, Frankenstein’s monster of the early 19th century.
In the year 1920, that was Karel Capek who gave them their present-day nameRobots. In 1942, Isaac Azi-
mov was the first who introduced the termRobotics. In 1959, the first real, not imagined and not legendary,
industrial robot had entered the factory floor and, strictly speaking, has heralded the beginning of the robotics
era. Then, at the end of the past century, robots start to appear in our human surroundings.
It has immediately become clear that, to work with humans (in cooperation and in tight interaction with them),
robots have to possess some human-like cognitive abilities. What does it mean “to possess human-like cognitive
abilities”? —Nobody knew then, nobody knows today. But that does not matterthe robotics research commu-
nity enthusiastically started to cope with the challenge, endorsed with ample budget funding provided by the
USA Defence Advanced Research Projects Agency (DARPA) and the European Union Research and Techno-
logical Development (EU RTD) programme. We would like to provide a short account of these efforts.
E. Diamant
2. March to the Glorious Future
As it was just said above, the DARPA in USA and the European Commission in Europe are today the most
prominent supporters of scientific and technological progress, which are operating worldwide and are promoting
an extensive range of critically important research initiatives. In the past 10 - 15 years, Cognitive Robotics was
certainly one among them.
2.1. DARPA’s Projects on Cognitive Robotics
The DARPA has always posited itself as an authority aimed to address a wide range of technological opportuni-
ties directed to meet the national security challenges. Endorsed with a budget of up to $2.8 billion (in FY 2013),
it pursues its objectives through a wide range of R&D programs [1]. Cognitive Robotics does not appear in
DARPA’s programme as a bundle of programs grouped by a common theme; on the contrary, in DARPA’s
practice Cognitive Robotics is handled as a collection of separate programs that share a common target issue.
The list of Cognitive Robotics and Cognitive-Robotics-related programs launched in the years 2001-2013 can be
seen in Table 1.
DARPA’s efforts on robotics are focused primarily on military and defence-related applications with a clear
goal to bring real-time, integrated, multi-source intelligence to the battlefield. DARPA does not strive to replace
the warrior with a robot, but it believes that it is possible to improve the abilities of individual warfighter by
combining technological achievements with human brain cognitive capacities thus making information under-
standing and decision-making far more effective and efficient for military users. So, it tries hard, on one hand, to
revolutionize the underlying technologies (for unmanned sensor systems and battlefield information gathering)
and, on the other hand, to merge them with the next generation computational systems that will have some hu-
man-l ike cognitive capabilities (such as reasoning and learning capabilities) and levels of autonomy which are
beyond those of the today’s systems. The spectrum of programs presented in Table 1 reliably represents this
DARPA’s approach to Cognitive Robotics R&D.
2.2. The European Programs on Cognitive Robotics
European Union research is conducted in a frame of research programmes called Framework Programmes for
Research and Technological Development, in short Framework Programmes farther abbreviated as FP1 to FP8.
Cognitive Robotics related issues start to appear in the FP5 programme and then, respectively, continue to
evolve and expand in the following FP6 and FP7 work programmes. Contrary to DARPA’s approach, EU R&D
activities are clustered to several main “themes” that are further segmented into “challenges”, which are further
divided into “objectives” in frame of which the individual projects are carried out. Cognitive Robotics in the
Frame Programmes FP6 and FP7 is represented as a Challenge (Challenge 2) of the Information and Communi-
cation Technologies (ICT) theme (Theme 3). At the time of the transition from FP5 to FP6, when subdivision to
Challenges was not yet introduced, Cognitive Robotics and its related issues such as Cognitive Vision and four
other particular items appear straight as objectives in the Information Society Technologies (IST) theme (see
Table 2).
Table 1. DARPA’s projects on cognitive robotics.
Cognitive Computing Systems (CoGS) 2008-…....
DARPAs Neovision project (NEOVISION2) 2009-……
Video and Image Retrieval and Analysis Tool (VIRAT) 2011-……
Autonomous Robotics Manipulation Program (ARM) 2011-……
DARPA Robotics Challenge (DRC) 2012-……
DARPA’s Insight Program (DIP) 2013-…....
Biologically Inspired Cognitive Architectures (BICA) 2005-2007
The Cognitive Technology Threat Warning System (CT2WS)
Cognitive Assistant that Learns and Organizes (CALO) 2003-2008
Personalized Assistant that Learns (PAL) 2003-2008
Augmented Cognition (AugCog) 2001-2006
E. Diamant
Juxtaposing Table 2, Table 3, and Table 4, it can be seen how from a Cognitive Vision objective in FP5 (a
Cognitive Robotics related topic) Cognitive Robotics has evolved to a full-blown Challenge (Challenge 2) in the
FP6 and FP7 programmes, steadily growing from 190 M€ in FP5 [2] [3] to near 1.2 billion euros in the next FP6
and FP7 programmes (435 M€ for FP6 [4]-[6] + 768 M€ for FP7 [7]-[10]).
3. Deceived Expectations
During all these times, the declared goals of Cognitive Robotics programmes were: (As it follows from
Table 2. Robotics inFP5.
Area No. of projects Total cost (M€) Total EC funding (M€)
IST Demining 8 30.6 15.6
IST FET Neuro-IT 15 32.4 23.1
IST FET General 17 39.2 25.7
IST Cognitive Vision 8 24.2 17.3
GROWTH etc. 24 63.2 34.9
Total 72 189.7 116.5
Table 3. Cognitive robotics in FP6.
Year Obje ctive Total cost (M€)
Work Progra mme IST2002-IV.2.1 Cognitive vision systems
IST2002-VI.2.2 Presence Research: Cognitive sciences and future media ?
Work Progra mme
2.3 .1.7. S em an t i c-based Knowledge Systems
2.3 .1.8. Networked Audiovisual systems and home platforms
2.3 .2.4. Cognitive Systems Bio-inspired Intelligent Information Systems
Proactive initiatives (i) Beyond robotics
Work Programme
2.4 .6. Networked Audio Visual Systems and Home Platforms
2.4 .7. Semantic -based Knowledge and Content Systems
2.4 .8. Cognitive Systems
2.4.11. Integrated biomedical information for better health
Total 435
Table 4. Cognitive Robotics in FP7.
Year Ca ll Objective Budget (M€)
2007 Call 1 ICT-2007.2.1 Cognitive systems, interaction, robotic
ICT-2007.4.2 Intelligent content and semantics
ICT-2007.8.3 Bio-ICT convergence
2008 Call 3
ICT-2007.2.2 Cognitive systems, interaction, robotics
ICT-2007.4.3 Digital libraries and technology-enhanced learning
ICT-2007.4.4 Intelligent content and semantics
ICT-2007.8.5 Embodied Intelligence
2009 Call 4 ICT-2009.2.1: Cognitive Systems and Robotics
ICT-2009.2.2: Language-Based Interaction 73
2010 Call 6 ICT-2009.2.1: Cognitive Systems and Robotics
ICT-2009.2.2: Language-Based Interaction 80
2011 Call 7 ICT-2011-7 Cognitive Systems and Robotics 73
2012 Call 9 ICT-2011-9 Cognitive Systems and Robotics 82
2013 Call 10 IC T-2013.2.1 Robotics, Cognitive Systems & Smart Spaces, Symbiotic Interaction
ICT-2013.2.2 Robotics use cases & Accompanying measures 67
Total 768
E. Diamant
DARPA’s News Releases) “to create adaptable, integrated intelligence systems aimed to augment intelligence
analysts’ capabilities to support time-sensitive operations on the battlefield” [11]. And in another document
“The objectives (of DARPA’s programs) are not to replace human analysts, but to make them more effective
and efficient by reducing their cognitive load and enabling them to search for activities and threats quickly and
easily” [12].
Objectives of Challenge 2 programs in the EU research initiative have been far more ambitious—The FP6
Workprogramme for years 2003-2004 states: (The objective is) “to construct physically instantiated or embodied
systems that can perceive, understand (the semantics of information conveyed through their perceptual input)
and interact with their environment, and evolve in order to achieve human-like performance in activities requir-
ing context-(situation and task) specific knowledge, etc. The development of cognitive robots whose “purpose in
life” would be to serve humans as assistants or “companions”. Such robots would be able to learn new skills and
tasks in an active open-ended way and to grow in constant interaction and co-operation with humans” [4].
These objectives (almost in similar words) are repeatedly declared in all further Work programmes. For ex-
ample, the 2011-2012 Workprogramme says that in these words: “Challenge 2 focuses on artificial cognitive
systems and robots that operate in dynamic, nondeterministic, real-life environments… Actions under this C ha l-
lenge support research on engineering robotic systems and on endowing artificial systems with cognitive capa-
bilities” [8].
Careful examination of the outcome that results from both the DARPA’s programs and from the FP5-FP7 ob-
jectives leads to a univocal conclusion—the announced goals of all these programs have never been reached!
The explanation of this phenomenon is very simplepeople try to provide robots with human-like cognitive
abilities, but at the same time the same people are devoid of even a slightest understanding about what does the
notion of “human-like cognitive abilities” really mean.
During the past years, the problem has become obvious and has been even mentioned in the 2011-2012
Workprogramme: “Hard scientific and technological research issues still need to be tackled in order to make
robots fit for rendering high-quality services, or for flexible manufacturing scenarios. Sound theories are requi-
site to underpinning the development of robotic systems and providing pertinent design paradigms, also in-
formed by studies of natural cognitive systems (as in the neuro- and behavioural sciences) [8].
Even more definite was the statement of the year 2013 Work programme“An additional research focus tar-
geted under this challenge will address symbiotic human-machine relations, which aims at a deeper understand-
ing of human behaviour during interaction with ICT, going beyond conventional approaches. The work on cog-
nitive systems and smart spaces and on symbiotic human-machine relations is not restricted to robotics” [13].
This promise was also left unfulfilled. At the end of 2013, Cognitive Robotics research has moved to and has
tightened itself with the human brain research activities.
4. New Hopes
At the beginning of year 2014, both Europe and USA will launch ambitious programmes for human brain re-
search. In the USA, the programme is called the Brain Research through Advancing Innovative Neurotechnolo-
gies (BRAIN) Initiative and it was announced by President Barack Obama on April 2013. Its accomplishment
will be led by the National Institutes of Health (NIH), DARPA, and the National Science Foundation (NSF) [14].
In Europe, the Human Brain Project is a ten-year project, consisting of a thirty-month ramp-up phase, funded
under FP7, with support from a special flagship ERANET, and a ninety-month operational phase, to be funded
under Horizon 2020 programme. The project, which will have a total budget of over 1 billion Euros, is Euro-
pean-led with a strong element of international cooperation. The goal of the project is to build a completely new
ICT infrastructure for neuroscience, and for brain-related research in medicine and computing, catalysing a
global collaborative effort to understand the human brain and its diseases and ultimately to emulate its comput a-
tional capabilities [15].
The main features of the two projects are collected in the Table 5.
As it follows from the Table 5, Cognitive Robotics is not among the main goals of the two Flagship initia-
tives, but it is definitely among their main purposes. In the European Human Brain Project it appears as the
“Cognitive Architectures” line in the list of the HBP topics. In the American BRAIN Project Cognitive Robotics
issues are hidden behind the “Link neuronal activity to behaviour” topic.
E. Diamant
Table 5. European and american human brain projects.
Parameter European HB Project American BRAIN Project
Dura ti on 10 years. long-lasting programme
Funding $ 1.35 billion. $110 million in the 2014 fiscal year supposed to ramp up
this commitment in subsequent years
Main Topics
Human and mouse neural channelomics.
Genotype to phenotype mapping the brain.
Identifying, gathering and organizing neuroscience
dat a .
Cognitive architectures.
Novel methods for rule-based clustering of medical
dat a .
Neural configurations for neuromorphic computing
Virtual robotic environments, agents, sensory &
motor systems.
Theory of multi-scale circuits.
Generate a census of brain cell types
Create structural maps of the brain
Develop new, large-scale neural network recording
Develop a suite of tools for neural circuit manipulation
Link neuronal activity to behavior
Integrate theory, modeling, statistics and computation
with neuroscience experiments
Delineate mechanisms underlying human brain imag-
ing technologies
Create mechanisms to enable collection of human data
for scientific research
Disseminate knowledge and training
5. An Attempt to Predict the Future
In attempt to predict the future results of these two projects, let us juxtapose them with something that is well
known to us and that we are quite familiar with. We mean the enduring and persistent study of National Eco-
nomics. While the human nervous system can be seen as the driving force behind the behaviour of a single hu-
man, national economics can be seen as the driving force behind the behaviour of a whole human society. Both
are complex distributed systems whose efficient operation is supported by an all-embracing communication
system. In the Human brain that is the Nervous system, in the National Economics this is the Transportation
s ystem.
From Table 6, one can see that the principal features of the Transportation system are well reflected in both
brain research projects. Only one feature “What is being transported?” is missing in the future brain studies. A
proper answer to the question “What is being transported in the Nervous system between different brain parts?”
should be “Information”. But, for unknown reasons, that is left undefined in both future mega-projects. And the
consequences of this omission are predictable.
On the other hand, the reason of this omission is also fully understandable: we don’t know what Information
is and how it is being transported (processed) in the brain. (That the brain is an information processing system is
a widely accepted hypothesis in the scientific community). So, it will be wise to try to understand what informa-
tion is.
6. What Is Information
While a consensus definition of information does not exist, we would like to propose a definition of our own
(borrowed and extended from the Kolmogorov’s definition of information, first introduced in the mid-sixties of
the past century):
Information is a linguistic description of structures observable in a given data set [ 1 6].
Two types of structures could be distinguished in a data setprimary and secondary data structures. The first
are data elements aggregations whose agglomeration is guided by natural physical laws; the others are aggrega-
tions of primary data structures which appear in the observer’s brain guided by the observer’s customs and hab-
its. Therefore, the first could be called Physical data structures, and the second, Meaningful or Semantic data
structures. And their descriptions should be accordingly called Physical Information and Semantic Informa-
This subdivision is usually overlooked in the contemporary data processing approaches leading to mistaken
and erroneous data handling methods and techniques.
In [17], E. Diamant presents a list of publications on the subject and a more extended explanation of informa-
tion description duality can be found. Meanwhile, it is important to explain the consequences that immediately
pop up from this assertion. And which are critically important for the right definition of the notion of cognition.
E. Diamant
Table 6. Juxtaposing human brain projects.
Economic system Human brain system
Transportation system Nervous system
System’s Features American BRAIN Project European HB Project
Network topology
(road and pathway maps) The Brain Connectome Project Neural channelomics
Neuroinformatics platform
Network dynamics (Traffic)
Transportation means, time tables,
hubs, congestions The DARPA’s SyNAPSE Project Neuromorphic computing platform
What is being transported
(through the network)
Raw materials, Goods, Freights. (Information) (Information)
In the light of this just acquired knowledge, we can certainly posit that cognitive ability is the ability to
process information. And that is what our brains are doing, and that is what we are striving to replicate in our
Cognitive Robots designs.
First of all, physical information is carried by the data and therefore can be promptly extracted from it. At the
same time, semantic information is a description of observer’s arrangement of the physical data structures and
therefore it can not be extracted from the data, because semantics is not a property of the data, it is a property of
an observer that is watching and scrutinizing the data. As such, semantics is always subjective and it is always a
result of mutual agreements and conventions that are established in a certain group of observers, or a future
group of robots and humans that act as a team sharing a common semantic understanding (semantic information)
about their environment. An important sequel of this is that the semantic information can not be learned auto-
nomously, but it should be provided to a cognitive robot from the outside (semantics has to be taught and not
learned, as it is usually requested by all work programmes).
Another important corollary that follows from the new understanding of information nature is that informa-
tion description is always a linguistic description, that is, a string of symbols which can take a form of a mathe-
matical formula (don’t forget that mathematics is a sort of a language) or a natural language itema word, a
sentence, or a piece of text. That is a very important outcome of the new theory considering that contemporary
approaches to the problem of information processing are assuming computer involvement in the processing task.
However, contemporary computers are data processing machines which are not supposed to process natural
language texts which are carrying semantic information.
Finally, we would like to provide some examples of widespread misunderstandings that appear in the Calls of
proposals issued by DARPA and EU Commission: In the “ICT Work Programme 2009/2010, (C(2009) 5893)”
[7], in its “Part 4.2 Challenge 2: Cognitive Systems, Interaction, Robotics” the problem that Robotic systems
have to cope with is specifies as “extracting meaning and purpose from bursts of sensor data or strings of com-
puter code…” This is a false and a misleading statementsensor data does not possess semantics, and therefore,
meaning and purpose can not be extracted from it.
DARPA’s Document “Deep Learning” (RFI SN08-42) states that: “DARPA is interested in new algorithms
for learning from unlabeled data in an unsupervised manner to extract emergent symbolic representations from
sensory input”Again, that is a false and a misleading statement symbolic representations (semantics) could
not be learned from data.
7. Conclusion
Cognitive Robotics R&D is a very important branch of contemporary science that is paving the road to the next
technological revolutionsmart robots that are invading our everyday lives. Until now, the extensive research
efforts of the Cognitive Robotics field investigators have been derailed by a wrong understanding about the es-
sence of information, in general, and semantic information, in particular. We hope that the paper will contribute
to some changes in this situation.
[1] DARPA’s Mission in a Changing World, DARPA Framework 2013.
E. Diamant
[2] Karp , P. and Skordas, T. (2003 ) Robotics in the IST Programme. 22nd JCF Meeting, Madrid, Spain, 16-18 October
[3] 2002 Work Progra mme.
[4] 2003-2004 Workprogramme.
[5] 2005-06 Work Programme.
[6] 2006 Work Programme Fourth Update.
[7] Updated Work Programme 2009 and Work Programme 2010.
[8] Updated Work Programme 2011 and Work Programme 2012.
[9] Work Programme 2013. f
[10] ICT Challenge 2: Cognitive Systems and Robotics.
[11] DARPA Advances Video Analysis Tools.
[12] Unified Military Intelligence Picture Helping to Dispel the Fog of War.
[13] ICT Work Programme 2013.
[14] BRAIN Working Group Interim Report.
[15] The Human Brain Project.
[16] Diamant, E. (2012) Let Us First Agree on what the Term S emant icsMeans: An Unorthodox Approach to an
Age-Old Debate. In: Afzal, M.T., Ed., SemanticsAdvances in Theories and Mathematical Models, InTech Publisher,
[17] Diamant, E. (2014) Brain, Vision, Robotics, and Artificial Intelligence. fo