Re: Cognitive Systems Paradigm - P Langley

Thanks for the pointer.  

I very much concur with the importance of a systems level perspective. Work in linguistics, for instance, has focused on syntactic details of language, but language is essentially a means for one mind to communicate meaning to another. Building executable end to end models of communication via language would provide a firm foundation for evaluating theories of how language is learned and evolves through use, turning linguistics from a descriptive science into an experimental one.

Langley comments on the popularity of formal guarantees, and I’ve long noticed that many publications have included mathematical models, lemmas and proofs, as a way to strengthen the case they make, despite starting from dubious assumptions. A related point is the faith in mathematical logic and formal semantics as being superior to other approaches.

I also believe in the value of functional level requirements as a basis for decoupling higher level theories from the underlying implementations. This corresponds to transforming problems into models that are easier to work with, something that has been a boon to physics. Of course we still want to understand how the brain approximates these higher level models. This is where simulations of pulsed neural networks can work in hand with improvements in neuroscience. Work on Deep Learning has diverged from biologically plausible models, and unsurprisingly, fails to mirror the capabilities of the human brain in major ways.

Another term in vogue is “human-like AI” which has the benefit of being immediately understandable to a broad audience, along with the implications of a broader scope than cognition. Moreover, experimental data of human subjects can guide theories and their evaluation, e.g. eye tracking data in relation to natural language processing.

Langley concludes that "we need demonstrations of flexible, high-level cognition in less constrained settings that require the combination of inference, problem solving, and language into more complete intelligent systems”.

This is very much in line with the ambitions of the W3C Cognitive AI CG.

> On 12 Feb 2021, at 05:55, Paola Di Maio <paola.dimaio@gmail.com> wrote:
> 
> This week  Reader
> important points, ie in the beginning, there was no separation between AI and Cognitive Systems
> 
>   The early days of artificial intelligence were guided by a common vision: understanding and reproducing, in computational systems, the full range of intelligent behavior that we observe in humans. Many researchers continued to share these aims until the 1980s and 1990s, when AI began to fragment into a variety of specialized subdisciplines, each with far more limited objectives. This produced progress in each area, but, in the process, many abandoned the field’s original goal. Rather than creating intelligent systems with the same breadth and flexibility as humans, most recent research has produced impressive but narrow idiot savants. The field’s central goal was to understand the nature of the mind. This is one of the core aims of science, on an equal footing with questions about the nature of the universe, the nature of matter, and the nature of life. As such, it deserves the same respect and attention it received during discipline’s initial periods. However, since mainstream AI has largely abandoned this goal, we require a new name for research that remains committed to the original vision. For this purpose, I propose the phrase cognitive systems, which Brachman and Lemnios (2002) championed at DARPA in their efforts to encourage research in this early tradition. As we will see later, this label incorporates some key ideas behind the movement  
> 
> 
>   The Cognitive Systems Paradigm  
> P Langley
> http://www.cogsys.org/pdf/paper-1-2.pdf <http://www.cogsys.org/pdf/paper-1-2.pdf>
> 

Dave Raggett <dsr@w3.org> http://www.w3.org/People/Raggett
W3C Data Activity Lead & W3C champion for the Web of things