Platt Perspective on Business and Technology

Reconsidering Information Systems Infrastructure 4

Posted in business and convergent technologies, reexamining the fundamentals by Timothy Platt on May 15, 2018

This is the 4th posting to a series that I am developing here, with a goal of analyzing and discussing how artificial intelligence, and the emergence of artificial intelligent agents will transform the electronic and online-enabled information management systems that we have and use. See Ubiquitous Computing and Communications – everywhere all the time 2, postings 374 and loosely following for Parts 1-3. And also see two benchmark postings that I initially wrote just over six years apart but that together provided much of the specific impetus for my writing this series: Assumption 6 – The fallacy of the Singularity and the Fallacy of Simple Linear Progression – finding a middle ground and a late 2017 follow-up to that posting.

I began discussing what I have referred to as a possible “modest proposal” task for artificial intelligence-based systems in Part 1 and Part 2 of this series, as drawn from the pharmaceutical industry and representing what can be considered one of their holy grail goals: developing new possible drugs that would start out having a significant likelihood of therapeutic effectiveness based on their chemistry and a chemistry-level knowledge of the biological processes to be affected. And I have been using that challenge as a performance benchmark case-in-point example of how real world tasks might arise as problems that would be resolved by artificial intelligence, that go beyond the boundaries of what can be carried out using single, specialized artificial intelligence agents, at least in their current here-and-now state of development.

Then I further developed and explored that real world test-case challenge and what would go into effectively addressing it, in Part 3 where I began delving in at least preliminary-step detail, into the possibility of what might be achieved from developing and deploying what could become multiple hierarchical-level, complex problem solving arrays of what are individually only distinctly separate specialized, single task artificial agents, where some of them would carry out single specialized types of command and control functions, in coordinating the activities of lower level agents that report to them and that more directly work on solving specific aspects of the initial problem at hand.

Up to here, I have simply sidestepped the issue of whether such a system of individually simple, single algorithm agents, with its collective feedback and self-directing control systems could in some sense automatically qualify as having achieved some form of general intelligence: a level and type of reasoning capacity that would at least categorically match what is assumed by the term “human intelligence,” even if differing from that in detail. I categorically state here, that that type of presumptive conclusion need not in any way be valid or justified. And I cite, by way of well known example for justifying that claim, the feedback and automated control functionalities that have been built into complex steam powered systems going back to the 19th century, for regulating furnace temperatures, systems pressures and the like and even in tremendously complex systems. No one would deny that well designed systems of that sort do in fact manage and control their basic operational parameters for keeping them functioning and both effectively and safely. But it would be difficult to convincingly argue that a steam power plant and associated equipment in an old steam powered ship, for example, were in some sense intelligent and generally so. There, specialized and limited in overall aggregated form, is still specialized and limited, and even if at the level of managing larger and more complex problems than any of its single component-level parts could address.

With that noted, I add here that I concluded Part 3 of this series by stating that I have been attempting to offer a set of building block elements that would in all likelihood have to go into creating what would become an overall artificial intelligence system in general, and an arguably genuinely intelligent information management and communications network and system of such networks as a whole, in particular. Think of my discussion in this series up to here as focusing on “necessary but not sufficient” issues and systems resources.

I am going to turn to at least briefly discuss the questions and issues of infrastructure architecture in this, and how that would arise and how it would be managed and controlled, in what follows. (Think self-learning and self-evolving systems there, where a probably complex structured hierarchical system of agents would over time, optimize itself and effectively rewire itself in that process.)

And my goal there will be to at least offer some thoughts as to what might go into the “sufficient” side of intelligent and generally intelligent systems. And as part of that, I will more fully consider at least some basic-outline requirements and parameters for functionally defining an artificial general intelligence system per se, and in what would qualify as more operational terms (and not just in more vaguely stated conceptual terms as are more usually considered for this.)

Let’s start addressing all of that with those “higher” level but still simple, single algorithm agents that function higher up in a networked hierarchy of them, and that act upon and manage the activities of “lower” level agents, and how they in fact connect with them and manage them, and the question of what that means. In that, let’s at least start with the type of multi-agent system that I have been discussing in the context of my above noted modest proposal example, and build from there. And I begin that by raising what might be considered a swear word in this type of narrative for how it can be expansively and misleadingly used: “emergent properties.” And let me begin addressing that, by in-effect reframing Turing’s hypothesis in at least somewhat operationalized terms:

• A system of simple, pre-intelligent components does not become intelligent as whole because some subset of its component building block elements does. It becomes intelligent because it reaches a point in its overall development where examination of that system as a whole, and as a black box system, indicates that it is no longer possible to distinguish between it and its informational performance output, and that of a benchmark presumed-general intelligence and its output that it would be compared to (e.g. in Turning’s terms, a live and conscious person who has been deemed to be of at least average intelligence, when tested against a machine.)

And with a simple conceptual wave of the hands, an artificial construct becomes at least arguably intelligent. But what does this mean, when looking past the simply conceptual, and when peering into the black box of such a system? I am going to at least begin to address that question starting in a next series installment, by discussing two fundamentally distinct sets of issues that I would argue enter into any valid answer to it:

• The concept of emergent properties as they might be more operationally defined, and
• The concept of awareness, as a process of information processing level (e.g. pre- or non- directly empirical) simulation and model building.

And as part of that, I will address the issues of specific knowledge based expert systems, and of granularity in the scope and complexity of what a system might be in-effect hardwired to address, as for example in a Turing test context.

Meanwhile, you can find this and related postings and series at Ubiquitous Computing and Communications – everywhere all the time and its Page 2 continuation. And you can also find a link to this posting, appended to the end of Section I of Reexamining the Fundamentals as a supplemental entry there.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: