Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision		Previous revision
projects:quantum:seqnn [2026/05/21 07:18] – kymkiprojects:quantum:seqnn [2026/05/21 08:45] (current) – [Introduction] kymki
Line 105: Line 105:
  
  
-====== SEQNN Stage 5: Entangled Generative Estimation ======+====== Self-Estimating Quantum Neural Networks ======
  
 //Exploratory research note. This work is being shaped toward a publishable paper, but the results described here should be read as "work under validation". Github/codeberg link coming once I have had a strong coffee and cleaned up the repo.// //Exploratory research note. This work is being shaped toward a publishable paper, but the results described here should be read as "work under validation". Github/codeberg link coming once I have had a strong coffee and cleaned up the repo.//
Line 111: Line 111:
 ===== Introduction ===== ===== Introduction =====
  
-Kalman filters are fascinating. How can they be applied to QML models to provide some sort of "self-estimation" of state to be used in parameter updates? This was the fundamental question here that drove me to go on a rather extensive exploration of this topic. In this post I will describe the current position of Self-Estimating Quantum Neural Networks (SEQNN).+Kalman filtering is fascinating and has a wide area of application. How can it be applied to QML models to provide some sort of "self-estimation" of state? This was the fundamental question here that drove me to go on a rather extensive exploration of this topic. In this post I will describe the current position of Self-Estimating Quantum Neural Networks (SEQNN).
  
 SEQNN, is an attempt to treat the parameters of a variational quantum model as latent states to be estimated sequentially, rather than as weights managed by a global optimizer. SEQNN, is an attempt to treat the parameters of a variational quantum model as latent states to be estimated sequentially, rather than as weights managed by a global optimizer.
Line 132: Line 132:
   * each node maintains both an estimate and an uncertainty.   * each node maintains both an estimate and an uncertainty.
  
-The important difference is not only "which update rule is used." The claim being explored is structural: when the model is written as an estimator, uncertainty, identifiability, calibration, and measurement attribution become first-class objects instead of after-the-fact diagnostics.+The important difference is not only "which update rule is used." Rather, when the model is written as an estimator, uncertainty, identifiability, calibration, and measurement attribution become terms we can use to reason with and build our algorithms around. Is that helpful? Lets see what the outcome is here.
  
-===== Roadmap position =====+===== Roadmap position (needs improvements) =====
  
 ^ Stage ^ Status ^ ^ Stage ^ Status ^
Line 234: Line 234:
 ===== Model ladder ===== ===== Model ladder =====
  
-The project deliberately moved through increasingly demanding observation models. The key pattern is: do not add parameters, entanglement, or topology until the observation equation says what can actually be observed.+The project moved through increasingly demanding observation models. The key pattern is: do not add parameters, entanglement, or topology until the observation equation says what can actually be observed.
  
 ^ Stage/model ^ Observation structure ^ What it tests ^ Claim status ^ ^ Stage/model ^ Observation structure ^ What it tests ^ Claim status ^