From 8cdfaa8f934f78e3946b15316b69f84f2a16b40b Mon Sep 17 00:00:00 2001
From: Ross Gayler <ross@rossgayler.com>
Date: Mon, 12 Jun 2023 22:02:17 +1000
Subject: [PATCH] Add Possible Implications slide to Statistical Interpretation
 of Algebraic Terms

---
 ...ntation_Gayler_MidnightVSA_2023-06-15.html | 17 +++++++++++
 ...entation_Gayler_MidnightVSA_2023-06-15.qmd | 29 +++++++++++++------
 2 files changed, 37 insertions(+), 9 deletions(-)

diff --git a/presentation_Gayler_MidnightVSA_2023-06-15.html b/presentation_Gayler_MidnightVSA_2023-06-15.html
index 3740fd5..1b3aade 100644
--- a/presentation_Gayler_MidnightVSA_2023-06-15.html
+++ b/presentation_Gayler_MidnightVSA_2023-06-15.html
@@ -625,6 +625,23 @@ <h2>Terms as rotated features</h2>
 </ul></li>
 </ul></li>
 </ul>
+</section>
+<section id="possible-implications" class="slide level2">
+<h2>Possible implications</h2>
+<ul>
+<li>Operation of VSA regression/classification systems can be understood/analysed with respect to terms in hypervector
+<ul>
+<li>E.g. Integer Echo State Network builds standard sequence representation (Interpretable as set of lagged inputs)</li>
+</ul></li>
+<li>Representations can be designed to achieve objectives
+<ul>
+<li><p>What features needed for standard regression?</p></li>
+<li><p>Create algebraic terms that implement those features</p>
+<ul>
+<li>E.g. Epileptic Seizure Challenge needed interactions of time-series features with time of day (bindings)</li>
+</ul></li>
+</ul></li>
+</ul>
 </section></section>
 <section id="understand-everything-at-the-element-level" class="title-slide slide level1 center">
 <h1>Understand everything at the element level</h1>
diff --git a/presentation_Gayler_MidnightVSA_2023-06-15.qmd b/presentation_Gayler_MidnightVSA_2023-06-15.qmd
index 1bf8223..f8dbfd4 100644
--- a/presentation_Gayler_MidnightVSA_2023-06-15.qmd
+++ b/presentation_Gayler_MidnightVSA_2023-06-15.qmd
@@ -212,15 +212,26 @@ Interpret hypervector as specifying a set of indistinguishable realities rather
 
 ## Terms as rotated features
 
-* In statistical modelling, we could rotate the data space
-  * Features are spread over columns
-  * Works if rotated features are orthogonal in the data space
-* VSA uses regression to extract predictions from hypervectors
-  * Hypervectors often expressible as sum of algebraic terms
-  * Algebraic terms are often orthogonal (rotated features)
-    * Simple terms represent main effects
-    * Bindings represent interactions
-    
+-   In statistical modelling, we could rotate the data space
+    -   Features are spread over columns
+    -   Works if rotated features are orthogonal in the data space
+-   VSA uses regression to extract predictions from hypervectors
+    -   Hypervectors often expressible as sum of algebraic terms
+    -   Algebraic terms are often orthogonal (rotated features)
+        -   Simple terms represent main effects
+        -   Bindings represent interactions
+
+## Possible implications
+
+-   Operation of VSA regression/classification systems can be understood/analysed with respect to terms in hypervector
+    -   E.g. Integer Echo State Network builds standard sequence representation (Interpretable as set of lagged inputs)
+-   Representations can be designed to achieve objectives
+    -   What features needed for standard regression?
+
+    -   Create algebraic terms that implement those features
+
+        -   E.g. Epileptic Seizure Challenge needed interactions of time-series features with time of day (bindings)
+
 # Understand everything at the element level
 
 # Permutation and indices