We don't know the source of letters and numbers on which these symbols are constructed . Although human used the guts feelings at the time of using because our ancestors developed these symbols the machines don't have .
We are agree that statistics and meaning of different phrases of words used by machine but machine would depend on human for meaning and statistics of words due to lack of knowledge of construction of words by letters and many unsolved problem of mathematics like the Ferma theorem was unsolved up tp 1995.
NLP is (said in very superficial and generalized way) based on statistical relationships between words. In NLP models words are actually not stored as words but as embeddings (numerical vectors) that mean nothing to people. So NLP models do not even use letters and numbers in a way that we use them. They are stored in such a way that can be used for statistical analysis which essentially gives you how words relate to each other and what are most probable connections between words.
The problem is expecting to deconstruct utterances and find meaning within words. "The Jumbo has landed" was an every day phrase of the late C20th, whereas "The Eagle has landed" is probably the most iconic phrase of the C20th. Same structure, mostly the same words but an entirely different meaning.
My research, and open source software, shows that a reasoning HCI (https://youtu.be/5ra3P3DcpwY) can be constructed from the arbitrary mapping of large non-consecutive natural numbers, this moves us away from the verb-noun imperatives found in smart devices. Since we can generate large natural numbers using speech to text software, we can effectively talk to devices.
The reason why machines don’t have gut feelings, is that we ground our systems on small consecutive number systems (e.g. false=0, true=1) - there is no room for anything else. If the utterance ‘true’ was the value 102564, and ‘false’ 2626968, then anything else is your (negative) gut reaction. This is the Felicity found in Speech Act theory of J. L. Austin and John Searle, or in the NULL value of E. F. Codd's relational database system.
We are agree that statistics and meaning of different phrases of words used by machine but machine would depend on human for meaning and statistics of words due to lack of knowledge of construction of words by letters and many unsolved problem of mathematics like the Ferma theorem was unsolved up tp 1995.
- Badges
- Science topic
- Similar topics
- Automotive Engineering
- Machines