Artificial Intelligence: Comparative A.I.
UPMANA - Knowledge By Comparison
Also compare with ChatGPT/Gemini/Claude
You can select up to 5 metrics.
0 Selected
You can select 2 to 5 geographies..
0 Selected
Knowledge is gained from understanding and using Information (data/facts). All Knowledge can be classified on the basis of its Information sources – Evidence (Pramana) based, Estimation (Anumana) based, and Comparison (Upmana) based. Indian logic (Tarkashastra) classifies Knowledge (Gyana) as – Pramana, Anumana, Upmana, and additionally, a Non-Apparent (Apratyaksh) form of Knowledge known as Shabda-Gyan (Word Power).
Prarang’s Upmana is a tool to gain Knowledge by Comparison. Underlying the tool is Prarang’s Analytics with Information classified (Evidence) and designed (Estimated) from reliable World & Country data providers. Prarang does no surveys itself.
Upmana offers new insights through four dependent layers of Comparisons:
Countries, States, Districts and Cities can be compared. It can even be used to do unconventional (“Apples and Oranges comparison” in English phrase!) geographic comparisons. For example – Cities/Districts vs Countries. A maximum of 5 Geographies can be simultaneously selected.
A unique ranking system is the DNA of Prarang’s Analytics. All conventional Metrics of 195 Countries are ranked from top to bottom, as well as ranked for deviation from Average (Samana). Similarly, all conventional metrics of approximately 800 cities/districts of India are also ranked in this manner.
Unconventional comparisons create new insights. Upmana allows you to actually compare Apple production in one country with Orange production in another and/or Gold production in a third! A maximum of 5 different Metrics can be simultaneously selected.
The Comparison output generated by Prarang’s Upmana can also be compared (on the same Prompt instruction) with GenAI (Generative Artificial Intelligence) outputs from Meta’s Llama, Google’s Gemini, Microsoft’s ChatGPT, and others, on a single click.
Your research using Upmana’s Comparison outputs can be easily saved and/or shared with others.
A prompt is a question, command, or statement used to interact between a human and an GenAI (Generative Artificial Intelligence) model that allows the model to produce the intended output. An GenAI model can provide several outputs based on how the prompt is phrased. The purpose of the prompt is to provide the GenAI model with enough information so that it can produce output relevant to the prompt.
Upmana assists you in generating a more precise and meaningful prompt for GenAIs by helping select some geographies and some measurable aspect (from a vast array of world and country metrics).
Information is not Knowledge. Knowledge is not Intelligence. Intelligence is not Wisdom.
Information is structured & organized data & facts, which are useful. Understanding information and related concepts creates Knowledge. Application of knowledge at the right time & place, is Intelligence. A broad understanding based on experience & acquired knowledge, used for the betterment of self & society, is Wisdom.
Generative Artificial Intelligence (GenAI), using LLMs (Large Language Models), is a method to reformat & restructure Information using a layer of new sentences/language (based on word & word usage statistics on digitized content). It essentially seems to skip the step of creating knowledge from Information. GenAI thus avoids "knowing". Instead, based on the query/prompt, it tries to answer promptly, even mimicking human “reasoning” to some extent.
Using brute computation power, GenAI is fast approaching a state of memorising all known digitized Information & all known Languages of Mankind, with a man-designed ability of connecting the two. This code/algorithms (man-designed) is where biases & errors are coming in.
Prarang’s Upmana focuses on creating knowledge by comparison for its users. It is not an GenAI yet, although it mimics the GenAI input prompt & its output is comparable with GenAI outputs for the same prompt. Both GenAI & Upmana are still work in progress. The difference is as follows:
| GenAI (Generative Artificial Intelligence) |
Upmana (Knowledge by Comparison) |
|
|---|---|---|
| Prompt Input | Can write & ask any question | Must select some Geographical & some Metric (any Measured) for Upmana to generate text |
| Ability to do Precis writing/summary. | No ability to size/resize output | |
| Prompt Output | Text, data & images content, as instructed | Text & data content, as instructed |
| Unstable Response: Unique sentences created for identical prompts, every attempt. | Stable Response: Consistent output for same prompt, every attempt. | |
| Hallucinations / errors sometimes in text & data | No hallucinations | |
| Does not provide clear sources for its data | Always cites specific data source & year | |
| Ability to write software code. | Unable to write any software code | |
| Architecture | Generative AI | Traditional AI |
| Large Language Model (LLM) | Small Language Model (Language Localisation design) | |
| Data sources – Screen-Scraping without citations | Large data Aggregation from reliable sources | |
| LLM – NLP based on Western Linguistics | Analytics – Indian Linguistics (Tarkashastra) based | |
| Word as Token – High computation need | Phoneme as Token – Under development, unknown computation need | |
| Photos – Recognition & Production: GANs and Diffusion Models | Photos – Under development |
