This very recent paper was authored by Yoshua Bengio, Yann LeCun, Timothy Lillicrap among others.
Some excerpts:
- "... we must invest in fundamental research in NeuroAI ..."
[Maybe not!] - "... as profound as the computer revolution of the last half century ..."
[That is funny! Did the computer revolution not start already in World War II, then there were the mainframes (e.g. IBM) in the 1960s/70s/80s] - "... to reach this potential, we still require advances that will make AI more human-like in its capabilities. Historically, neuroscience has been a key driver and source of inspiration ..."
[AI does not have to be human like] - "... many front-line AI researchers believe that major new breakthroughs are needed before we can build artificial systems capable of doing all that a human ... Today’s AI systems cannot compete with the sensorimotor capabilities of a four-year old child, or even simple animals. ... A growing number of AI researchers doubt that merely scaling up current approaches will overcome these limitations. ..."
- "... Over the last decades, through efforts such as the NIH BRAIN initiative and others, we have amassed an enormous amount of knowledge about the brain. ... We believe it is the right time for a large-scale effort to identify and understand the principles of biological intelligence ..."
- "... Words only have meaning for these models by virtue of their statistical co-occurrence, as opposed to their grounding in real-world experiences, so even the most advanced language models, despite their increasing power, continue to struggle with some basic aspects of physical common sense ..."
- "... We therefore propose an expanded Turing test, one that includes advanced sensorimotor abilities. ... An expanded embodied Turing test would benchmark and compare the interactions with the world of artificial systems versus humans and other animals. Because each animal has its own unique set of abilities, each animal defines its own embodied Turing test: An artificial beaver might be tested on its ability to build a dam, and an artificial squirrel on its ability to jump through trees. Nonetheless, many core sensorimotor capabilities are shared by almost all animals, and the ability of animals to rapidly evolve the sensorimotor skills needed to adapt to new environments suggests that these core skills provide a solid foundation. ..."
[The authors may have watched too many of Sir Attenboroughs natural history documentary videos.] - "... Although machine learning has been pursuing approaches for sidestepping this tabula rasa limitation, including self-supervised learning, transfer learning, continual learning, meta learning, one-shot learning and imitation learning ..., none of these approaches comes close to achieving the flexibility found in most animals. Thus, we argue that understanding the neural circuit-level principles that provide the foundation for behavioral flexibility in the real-world, even in simple animals, has the potential to greatly increase the flexibility and utility of AI systems. ... we can greatly accelerate our search for general-purpose circuits for real-world interaction by taking advantage of the optimization process that evolution has already engaged in ..."
- "... Achieving these goals will require significant resources and also contribution across many disciplines beyond traditional AI and neuroscience, including psychology, engineering, linguistics, etc. ..."
- "... Moreover, explicit design of new training programs can ensure that the NeuroAI research community reflects the demographics of society as a whole. ..."
[Ah the ideology of diversity, equity, an inclusion shines through! Meritocracy combined with equal opportunity is sufficient!] - "... large investment in a shared computational platform will be required. Much like a particle accelerator in physics or large telescope in astronomy, this sort of large-scale shared resource will be essential for moving the brain-inspired AI research agenda forward. It will require a major organizational effort, with government and ideally also industry support, that has as its central goal scientific progress on animal and human-like intelligence. ..."
[More large scale government projects would certainly not hurt!]
Remarks:
- Anthropomorphism permeates this paper. The naturalistic viewpoint dominates. The human brain is the model to imitate!
- We ought to follow biology and evolution to further develop AI.
Unfortunately, humans do not have millions of years to develop better AI.
Humans may very well transcend nature & evolution to find better solutions - The authors kind of suggest that current AI efforts have hit a wall and new directions and approaches on larger scale and across several areas of scientific research (interdisciplinary approach) are necessary
- These researchers are right about the enormous energy and resource consumption of today's advanced AI systems compared with the very energy efficient human brain.
- Future artificial intelligence may very well not be based on silico or other solid materials, but on biological molecules or single atoms or other yet to be discovered chemistry and physics. Going beyond nature is destiny!
No comments:
Post a Comment