Questions, experiments, and measured results.

CORTEX-MS

Active Research

Can molecular identification from mass-spectrometry evidence benefit from a learned architecture that maintains and updates competing molecular hypotheses rather than treating identification solely as one-shot classification or retrieval?

1 experiments · 5 findings · 1 artifacts

Cortex

Experimental Architecture

Can useful computation be organized around selective local processing, bounded global integration, persistent hypotheses, and learned gating rather than applying equivalent global computation at every stage?

1 experiments · 2 findings · 0 artifacts

BNLM

Client-Side Language Model & In-Browser Training Engine

What architecture, systems design, and compute-dispatch patterns allow a Transformer language model to be trained from random initialization entirely inside a single browser tab under client-side memory and compute constraints?

1 experiments · 2 findings · 0 artifacts

TinyCoherent

Active Research / Pure C LLM Engine

Can an inspectable glass-box Transformer language model be trained, instruction-tuned, and scaled entirely in pure C directly on Apple Silicon without dependency on Python/PyTorch runtimes, while maintaining rigorous reasoning token curation?

4 experiments · 4 findings · 1 artifacts

Physiological Modeling Research

Experimental / Null & Mixed Results Preserved

Does incorporating patient-specific vascular elastance parameters yield statistically and clinically meaningful improvements in Mean Arterial Pressure (MAP) forecasting over baseline autoregression?

1 experiments · 3 findings · 0 artifacts