0 Hybrid Architectures: combining the various agent
0 E.g. combining reflex components with knowledge‐
0 Compilation: convert declarative information at the
deliberate level into more efficient representations,
eventually reaching the reflex level.
0 Soar (1987), Theo (1990).
0 Solutions to new problems are derived by deliberation,
but are compiled and stored at the reflex level for
solving future occurrences of similar problems.
0 Increasingly more complex environments: partially
observable, multiagent, stochastic, sequential,
dynamic and/or semidynamic, continuous, unknown.
0 Satisficing: deliberating long enough to come up with
an answer that is good enough.
0 ‘Good’ is context‐dependent, i.e. with respect to a
0 Anytime Algorithms: there is always an output,
which becomes better over time.
0 Bounded rationality: forms of rational behavior that
adhere to the constraints of satisficing.
0 Bounded optimality: the best algorithm given the
computational resource constraints.
0 Decision‐Theoretic Metareasoning: deciding which
computations to perform (in what order) and which not.
0 Computations are characterized by costs and benefits:
0 Cost: time, memory, and knowledge input needed.
0 Benefit: the projected quality improvement of the system’s
behavior due to the computation.
0 Computation ordering: an earlier computation’s benefit
may be the knowledge input to later computations.
0 Reflective Architecture: combining environment state
and computation state. An agents’ reasoning process
becomes part of the environment (what is observed).
0 Swarm intelligence: performance measure applied
to collective behavior.
0 Decentralized representation
0 Emergent behavior
0 Weak emergence: the qualities of the system are
reducible to the system's constituent parts.
0 Strong emergence: e.g. qualia.
0 The concepts of utility and rationality change!
0 What is rational for the swarm is not the sum of what is
rational for all individuals.
Prisoner B silent Prisoner B betray
Prisoner A silent A:1 day, B: 1 day A:10 year, B:0
Prisoner A betray A:0, B:10 years A:9 years, B:9 years
Two suspects are arrested. If one testifies against the other (betray) and
the other remains silent, the betrayer goes free and the silent accomplice
receives the full 10‐year sentence. If both remain silent, both prisoners
are sentenced to only 1 day for not cooperating with the investigation. If
each betrays the other, each receives a 9‐year sentence. How should the
• No matter what the other player does, a player will always
gain a greater payoff by playing betray.
• Since in any situation betraying is more beneficial than
remaining silent, all rational players will betray.
0 DARPA Grand Challenge: competition for long
distance, driverless vehicles.
0 Goal: making 1/3 of the military ground units
0 2005, STANLEY: driverless robotic car wins the
DARPA Grand Challenge, 212km off‐road course.
0 2006, BOSS: driverless robotic car in urban area.
0 2000, RemoteAgent: on‐board autonomous planning,
scheduling operations for spacecraft. NASA.
0 2004, MAPGEN: NASA’s Mars Exploration Rovers.
0 2008, MEXAR2, European Space Agency’s Mars
0 Speech recognition and dialogue handling in call
0 Automatic Translation
0 Hands‐free computing
0 Speaker identification
0 AI’s main interest in the 1960’s.
0 1997, Deep Blue (IBM): the first computer to beat the
world champion in chess, 3.5‐2.5.
0 “The brain’s last stand.” [Newsweek]
0 IBM’s stock increased by $18.000.000.000.
0 2004, Hydra defeats world champion Ponomariov 2‐0.
0 Elo rating 2850‐3000, higher than that of any human.
0 2007, Rybka defeats several grand masters, even
when giving them certain advantages.
Checkers: 1994, Chinook becomes the world champion.
Go: MoGo, still performs at advanced amateur level.
0 There is still work to do on some games.
0 Pattern recognition
0 Discovering complicated interactions: e.g. drug use,
0 Handwriting recognition
0 Biometric identification
0 Credit card usage
0 Document classification
0 Classifying 1.000.000.000 spam messages every day.
0 1991, DART (Dynamic Analysis and Replanning Tool)
scheduled transportation during the Persian Gulf
0 Starting points, destinations, routes, conflict resolution.
0 Over 50.000 vehicles, cargo, people.
0 According to DARPA, this system alone paid back 30
years of investment in AI.
0 Blue Brain Project: A neurologically realistic model
0 Blue Gene supercomputer
0 NEURON software
0 2007, simulating a rat’s neocortical column
(functional part of the neocortex).
0 2009, director Henry Markram announced to build a
human brain in 10 years time.
0 Think of what you learned about the history of AI. What
is wrong with Markram’s announcement?