top of page

FEELLLM: An Experimental Self-Report Surface for Sensory-Affective Modelling

Ajinomatrix is best known for its work on sensory digitization: taste, smell, texture, recipes, product development, sensory profiles, and machine-readable representations of perception.

But sensory intelligence does not stop at taste and aroma.


Human experience also includes affective response: comfort, aversion, attraction, overload, anticipation, trust, uncertainty, memory, and emotional context.



This is where FEELLLM enters the Ajinomatrix research landscape.


What FEELLLM is

FEELLLM is an experimental public self-report surface developed in the broader Ajinomatrix / Life-X ecosystem.

Its purpose is to help a person translate a difficult feeling, situation, or affective signal into more structured language.

It is not therapy, diagnosis, crisis support, or a medical device. It is not positioned as a health product.

It is a research-oriented interface for structured self-report.


Why Ajinomatrix is interested

Ajinomatrix works on the digitization and modelling of human perception.

In food, beverage, fragrance, product experience, and sensory R&D, perception is never purely chemical or mechanical. It is always contextual.

A flavour profile may be technically stable but perceived differently depending on:

  • context;

  • memory;

  • expectation;

  • culture;

  • emotional state;

  • environment;

  • timing;

  • prior exposure;

  • narrative.

FEELLLM explores one edge of this problem: how affective and hedonic self-report can be structured without pretending to be objective diagnosis.


Relationship to the broader AJX stack

FEELLLM connects conceptually with several Ajinomatrix research layers:

  • MP6 — structured representation of multi-sensory experience;

  • TTP / TasteTuner-public — public-facing hedonic and sensory response;

  • FEELLM / FEELLLM modelling — affective and contextual inference research;

  • BSPG — structured early-signal and risk-reflection work;

  • Life-X — broader experimental systems and governance layer.

The public PoC is not the whole system. It is one surface.


Affective self-report as a research object

Affective self-report is difficult.

It is subjective, context-dependent, and often ambiguous. But that does not make it useless. In product experience, wellbeing, consumer research, sensory science, and human-AI interaction, subjective reports can become valuable if they are structured carefully.

The challenge is to avoid two errors:

  1. reducing emotion to crude labels;

  2. overclaiming that an AI system can diagnose or understand a person from outside.

FEELLLM deliberately stays between those extremes.


What the public PoC does

The FEELLLM Public PoC helps users:

  • describe what they feel;

  • clarify context;

  • separate observation from interpretation;

  • produce more structured language;

  • prepare a possible human conversation.

Nothing is stored server-side in the current PoC.


Why this matters commercially and scientifically

For Ajinomatrix, FEELLLM is not a consumer-health product. It is a research-facing interface for understanding how structured self-report may later connect to:

  • hedonic response modelling;

  • sensory-context interpretation;

  • consumer experience research;

  • product perception studies;

  • affective layers of MP6-style sensory profiles;

  • human-in-the-loop validation.

It extends the logic of sensory digitization toward the broader question:


How do humans report experience, and how can those reports become usable without being distorted?


Try the public PoC

The FEELLLM Public PoC is available at:

For structured early-signal / BSPG work:

For consultation:

FEELLLM is experimental. It is not therapy, diagnosis or medical support.



It is part of Ajinomatrix’s broader exploration of perception, hedonic response, affective context and structured human self-report.

 
 
 

Comments

bottom of page