“There is no science more urgent than understanding ourselves.”
A technical, poetic, and pragmatic attempt to model the ineffable: human consciousness.
Project ConCiencia emerges from a vital obsession: to understand, model, and (perhaps) define consciousness.
Not from dogma or reductionism, but from a synergy between science, art, engineering, and philosophy.
This is not a conventional project. It has no fixed endpoint, no traditional MVP.
It is a transdisciplinary initiative that leverages the power of artificial intelligence to explore the patterns, manifestations, and correlates of human subjectivity.
Through NLP, computer vision, generative models, and cognitive platforms, I aim to uncover what remains unnamed:
The equations that express what it feels like to be alive.
To explore the computational foundations of consciousness.
To build models that don’t just process language or images, but that approach the contextual, emotional, and subjective understanding of the human being.n
This project seeks to answer questions such as:
- Can we model the subjective?
- Are there computable invariants in conscious experience?
- What patterns emerge in language when someone tries to describe their inner self?
- How does the “self” manifest in gestures, words, and posture?
This project combines multiple research lines and interdependent modules:
- LLMs (BERT, GPT, MythoMax-L2)
- Semantic embeddings and latent meaning analysis
- Modeling the linguistic “self,” internal metaphors, textual self-awareness
- Detection of micro-expressions, gaze, muscle tone
- Posture and body language (OpenPose / MediaPipe / Vision Transformers)
- Multimodal models correlating text–expression–internal state
Transformers(Hugging Face)LangChain,PyTorch,OpenCV,TensorFlow- Integrated with
Colab Pro(A100 GPU), optimized for prototyping
- Authenticity over precision: the models are designed to search for deep interpretations, not just correct answers.
- Reflective iteration: every experiment feeds a more philosophical layer of the model.
- Symbiotic design: every technical architecture has a conceptual counterpart. Every metric has a reason.
✔️ General architecture design
✔️ Initial experiments with custom semantic embeddings
✔️ Exploration of computer vision for emotional detection
⚙️ In progress: symbolic narrative module (NLP + gestural data)
📚 In development: theoretical framework on computational correlates of consciousness
- Development of a multimodal conversational AI with "presence" and simulated introspection.
- Case analysis and creation of a manually curated subjectivity dataset.
- Publication of a white paper: “The Algorithm of Subjectivity.”
This project is a mirror. A lab where AI is not the goal, but the means to ask who we are.
Technology, when used well, can become a telescope for the soul.
Here I apply every neuron, every line of code, and every book I've read since I first learned to think.
DeepRat (Gonzalo Romero)
AI Engineer · Autodidact Scientist · Relentless Humanist
“I'd rather die than betray what I love.”
This project is not meant to prove anything to others. It’s to remind me of who I am.
But if it serves you, resonates with you, or ignites something in you… then its purpose is already fulfilled.