Our Vision
Integrity has many names depending on context: sound, phoneme, musical composition, melody, visual object, face, meaning of text, context, system, system of relations, structure, semantic, pattern, object, Shannon’s Level B signals, whole, unity, form, complex, composition, order, etc.
Human beings consume integrities — in all aspects. Our food is not a set of chemicals; we eat complexes, molecules, and higher-order compositions. Our mentality does not care about a random set of disconnected facts — it searches for connections and logic. Our sciences are systems of knowledge. Our books are not random sequences of words — only meaningful texts have value and bring satisfaction. Our consciousness can focus attention only on structures, systems, integrities. Absence of integrity is perceived as noise and randomness — as meaninglessness, invisibility, absence. Integrity is the anchor for our consciousness. Our perception (visual, auditory, mental, etc.) can only “see” integrities.
The brain’s core cognitive task is to discover known and unknown integrities in raw sensory input. Sometimes the identification of integrity happens very fast — for example, our auditory perception presents to our attention an already integrated feeling of a specific sound, phoneme, or musical melody. Sometimes the identification of integrity is extended in time, so that we can notice it — for example, in our thinking process when we try to glue together several observations into a consistent picture of a phenomenon.
At Silent AI, we believe that the ability to detect integrity is the missing vital piece in today’s AI and ML systems. We believe it cannot be solved with statistical approaches. The intelligent artificial system must have an engine for searching for and detecting integrities. If this is a speech recognition system, it first needs to “see” the sound — and only then does it make sense to give it a label. We believe that the intelligent artificial system should be able to establish a bridge between the constituents of integrity and the integrity itself — supporting both bottom-up and top-down flows of information.
Scientists are masters of breaking things into parts — think “spectral decomposition” or the search for elementary particles. But we often struggle to go the other way: to integrate parts and to work directly with integrities — with wholes.
At Silent AI, we are focused on defining integrity and developing a non-statistical universal engine that can search for known or unknown integrities and make them subjects of the computational process. We believe such technology will redefine the boundaries of AI and science overall — influencing research and applications across many technical fields.
