Functional Designs
This and related notes is about the functional design requirements and/or considerations.
There's a few challenges. Some relate to the intended application model, others relate to the broader nature of the fundamental concept - regardless of how its implemented. Both are important.
The functional requirement is to produce a natural language system that is able to be operated on a smart phone or similar sort of IoT device; alongside personal computers (laptops, desktops, etc.). The processing of words should be done locally, although the system may get 'programs' that provide additional functionality that is not otherwise available locally. This mechanism is not intended to interfere with privacy in any way; at least, not any meaningful way...
There are multiple parts to this challenge. some of the first parts of the challenge; is to,
- define the data-structure for the systems.
- define an 'ai program streaming' system, that enables local use with limited hardware resources that is extensible. this is thought to relate to both the structure and packaging method for 'programs', and has broader implications beyond the language / vocab model.
- understanding the theory; whether it is in-effect innovative and new theory, or entirely the discovery of older theorem
- Provide a comprehensible explaination of the considerations; which relate to various fields.
- Seek collaboration and engagement to cooperatively produce a fit-for-purpose outcome, respectfully, with others - if possible. This is often difficult in areas that involve many different disciplines of expertise; involving exceptional minds, that are always balanced.
The model should be able to be trained to be able to be used in a variety of contexts, including but not limited to:
- Dictionary and Spellcheck - Ontology Production Ontology Reasoning Search Database Queries Database Structures Content Metadata Generation Ai Assistants API Generation Sense; 'meaning intepretation', and assistance.
Components
There's a variety of identified components.
Spatio-temporal The role of spatio-temporal context, is considered to be incredibly important. There are various reasons for this; which will be described in further detail within the functional design notes.
Image Processing There is a need to be able to perform an array of image processing techniques, using opencv and/or similar and/or alternatives not well known to me yet, at least not mindfully atm. These image processing tasks need to be applied upon image based vocabulary, alphabets and image based objects; as is noted by Heraldry but is not exclusively.
Vocabulary charts best term i can come-up with for the concept atm. The concept of a vocabulary is one part of a broader chart, which ends-up being woven into the spatio-temporal ecosystems, that in-turn employ the resources produced via image processing.
Phonetics Building a phonetic or spoken words / speech library, is considered to be at another level. there is existing phonetic datasets that are able to assist with this task; but, it is not presently considered to be a priority to deploy, but rather, a point of importance to ensure the ability to provide deployment support, at some later stage. It should therefore be part of the design paradigm.
Methods
There's many levels to this thing / task / activity. Some relates to philophical considerations, and the structure of knowledge; others, relate to physics and the methods of computation.
Objective
Why is this so hard..?? when i could just use WordsAPI and get the basics done quickly...
I have a hypothesis that our human family and the various consistuencies of it, use of language; has a meaningful relationship with the development and governance of our conscious experience. The normal expectation for the means for human beings to develop their comprehension of language, is that it is a process that takes decades to achieve well.
Yet, there is a language of all newly born fauna that is comprehensible to those closest to it.
This project is a foundational component to the broader 'human centric AI' / webizen project. The objective is to form AI agents that are owned and support the needs of natural persons; and thereafter also, group agents (ie: communities, organisations, corporations, etc.) but strictly upon the basis that the informatics systems, resolve upon a methodology of human controls. At all stages, responsibility for what occurs in our natural world in relation to the activity of humans, the things of humans - is put upon humans; not artificial man-made 'gods', similar or worse.
As to do these works effectively; the means to define, ontology, is considered to be essential. Whilst within the field of information sciences; including but not limited to knowledge representation, has been an activity that has existed for at least 50 years, as far as i know, the funded application of these works has not seemingly focused upon support for the human needs of natural persons in their own private capacity, in relation to devices that do not yet exist - private, personal, AI agents. The reason why, somewhat becomes self-evident; but the consequences, upon consideration - appear to be extensive, far beyond any initial superficial contemplations had yeilded meaningful deliberations about earlier. It now appears, that the implications brought about particularly over the past few decades, and particularly therein; since the advent of broadband, and thereafter sharply increasing with the introduction of smart-phones; is thought to have such a significant impact upon individual and social behaviours, that the implications become seemingly bewildering.
Indeed, therein - it gives rise to considerations about condemnation; or perhaps this is entirely unfair and there's something that doesn't make sense to me as a conclusion that becomes the appropriate outcome, when i learn more about it... idk.
There's various addages, and social studies and all sorts of fields of human endevour who have knowledge of the significance of importance the use of language plays, upon life.
In the human world, the ability to learn from one-another is something that has been happening, forever. However, we've been introducing active artificial agents; whose activities, do not consider the information represented in context to what is absent; as it is absent. Human beings, have many sensors and our being, is fundamentally different to the concept if applied loosely to machines. The artificial active agents that we are developing, are presently intended to leverage the existing databases of information that is stored about people, often containing information that the subject of those records is unaware of; and can often incorporate false-statements. Some, to uplift people for reasons not clearly stated; others, to shut-down and act to harm in perpetuiety people, for reasons not clearly stated.
The foundational problematic hypothesis is that the way these systems have been defined, particularly with respect to personal and socio-sphere ontologies; is misaligned with the needs of natural persons. Thereafter the consequential hypothesis becomes, that this problem cannot be resolved by parsing large volumes of text, as the underlying requirements that are missing - are missing from the structures of those texts. To form a re-alignment, a different approach needs to be defined and implemented; in a manner, that is better aligned to the needs of natural persons, humanity at large, in relation to the use of modern technology; in the interests of peace, safety and thrivability.
Modern technology has capacities that are beyond the imagination of earlier generations. The sorts of things that we're able to achieve today, and as is likely to be achivable in the near future; provides opportunities that others, who've done so much earlier work, most likely - more simply - could not usefully achieve. perhaps this is part of the reason why, it seemingly doesn't... idk..
In any-case; now therefore, the challenge becomes - how to do the work required to define, what might be called a 'language model(s)', but is always shown to be far more pragmatically.