the future history of artificial general intelligence - ben goertzel - h+ summit @ harvard
DESCRIPTION
The Future History of Artificial General Intelligence: Cognitive Robotics and the Global Brain, 2010-2040 After a brief survey of the state of the art in AGI, cognitive robotics and distributed Internet intelligence, we will extrapolate current technology directions forward, and look at some of the more plausible and interesting future possibilities. Special attention will be paid to the scenario in which global brain technology helps provide both background knowledge, collective hybrid human/digital computing power and human-friendly goal content for non-human-like, self-improving AGI systems. Dr. Ben Goertzel is CEO of AI software company Novamente LLC and bioinformatics company Biomind LLC; leader of the open-source OpenCog AI software project; Vice Chairman of Humanity+; Advisor to the Singularity University and Singularity Institute; and External Research Professor at Xiamen University, China. His research work encompasses artificial general intelligence, natural language processing, cognitive science, data mining, machine learning, computational finance, bioinformatics, virtual worlds and gaming and other areas. He has published a dozen scientific books, nearly 90 technical papers, and numerous journalistic articles. Before entering the software industry he served as a university faculty in several departments of mathematics, computer science and cognitive science, in the US, Australia and New Zealand. He has three children and too many pets, and in his spare time enjoys creating avant-garde fiction and music, and the outdoors.TRANSCRIPT
The Future History of Artificial General Intelligence:
2010-2045
Dr. Ben GoertzelCEO, Novamente LLC and Biomind LLC
CTO, Genescient Corp.Vice Chairman, Humanity+
External Research Professor, Xiamen UniversityAdvisor, Singularity Institute and Singularity University
2010
“Narrow AI”, rather than“Artificial General Intelligence” (AGI)
still dominates the AI field
Each “Narrow AI” program is (in the ideal case) highly competent at carrying out certain complex tasks in certain environments
In 2005, Biomind’s AI found a ~100% accurate predictor of Parkinson’s Disease based on
heteroplasmic mitochondrial mutations
Narrow AI helps fight disease
tumor suppression
heat shock (stress response)
metabolism
2009: Biomind AI infers aging networks from gene expression data from Genescient’s “Methuselah flies”
Narrow AI helps fight aging
2010: AGI is on the Rise
agi-conf.org/2010
AGI-11August 2011
Google CampusMountain View CA
late 2010
one practical approach to
advanced AGI:
OpenCog
At Xiamen University, OpenCog is being used to control a Nao robot
2010:The Robot Revolution
Begins
2020
2020:Rise of the
Robot Children
animal/childlike
AGI+
useful narrow AI =
useful early-stage AGI,
such as virtual AI scientists
2020:Rise of the
AI Scientists
DeclarativeProbabilistic
Logic Networks,concept
ProceduralMOSES
(probabilistic evolutionary
learning),
Sensoryhierarchy of
memory/processing
units
Episodicinternal world
simulation engine
Attentional/Intentionaleconomic attention networks,
A truly powerful artificial scientist will emerge from combining
childlike, commonsense-savvy AGI with multiple science-specific narrow-AI algorithms
one path to a powerful global brain: AGI + narrow-AI + humans + communication networks
2020:Emergence of
a more reflective,deliberative global brain
spanning AGIs,narrow AIs and humans
2030
2030:One Big
AI Network
2030:Jacked into
the AI Network
2040
Kurzweil projectstechnological Singularity
around 2045
Text
2010