Update

_ai/the-atomic-human-churchill.md

@@ -52,7 +52,7 @@ date: 2023-10-23
 geometry: ["a4paper", "margin=2cm"]
 papersize: a4paper
 transition: None
-ppt: 
+ppt: 2023-10-03_the-atomic-human.pptx
 ---
 \define{noSlideTitle}
 

_ai/the-atomic-human-st-andrews.md

@@ -0,0 +1,113 @@
+---
+title: "The Atomic Human"
+subtitle: "Understanding Ourselves in the Age of AI"
+abstract: |
+  A vital perspective is missing from the discussions we're having
+  about Artificial Intelligence: what does it mean for our identity?
+ 
+  Our fascination with AI stems from the perceived uniqueness of human
+  intelligence. We believe it's what differentiates us. Fears of AI
+  not only concern how it invades our digital lives, but also the
+  implied threat of an intelligence that displaces us from our
+  position at the centre of the world.
+ 
+  Atomism, proposed by Democritus, suggested it was impossible to
+  continue dividing matter down into ever smaller components:
+  eventually we reach a point where a cut cannot be made (the Greek
+  for uncuttable is 'atom'). In the same way, by slicing away at the
+  facets of human intelligence that can be replaced by machines, AI
+  uncovers what is left: an indivisible core that is the essence of
+  humanity.
+ 
+  By contrasting our own (evolved, locked-in, embodied) intelligence
+  with the capabilities of machine intelligence through history, The
+  Atomic Human reveals the technical origins, capabilities and
+  limitations of AI systems, and how they should be wielded. Not just
+  by the experts, but ordinary people. Either AI is a tool for us, or
+  we become a tool of AI. Understanding this will enable us to choose
+  the future we want.
+ 
+  This talk is based on Neil's forthcoming book to be published with
+  Allen Lane in June 2024. Machine learning solutions, in particular
+  those based on deep learning methods, form an underpinning of the
+  current revolution in "artificial intelligence" that has dominated
+  popular press headlines and is having a significant influence on the
+  wider tech agenda.
+  
+  In this talk I will give an overview of where we are now with
+  machine learning solutions, and what challenges we face both in the
+  near and far future. These include practical application of existing
+  algorithms in the face of the need to explain decision making,
+  mechanisms for improving the quality and availability of data,
+  dealing with large unstructured datasets.
+author:
+- family: Lawrence
+  given: Neil D.
+  gscholar: r3SJcvoAAAAJ
+  institute: University of Cambridge
+  twitter: lawrennd
+  url: http://inverseprobability.com
+venue: Churchill College Science Society
+date: 2023-10-23
+geometry: ["a4paper", "margin=2cm"]
+papersize: a4paper
+transition: None
+ppt: 
+---
+\define{noSlideTitle}
+
+
+\include{_books/includes/the-atomic-human.md}
+\include{_ai/includes/henry-ford-intro.md}
+\include{_ai/includes/embodiment-factors-short.md}
+
+\include{_ai/includes/cuneiform.md}
+
+
+<!-- Faster horse -->
+
+<!-- Embodiment Factors -->
+\include{_ai/includes/embodiment-factors-celsius.md}
+
+<!-- Information Triangle -->
+\include{_data-science/includes/new-flow-of-information.md}
+\include{_data-science/includes/societal-effects.md}
+
+<!-- AI Fallacy -->
+\include{_ai/includes/the-great-ai-fallacy.md}
+
+<!-- Mathematical Statistics -->
+\include{_data-science/includes/lies-damned-lies.md}
+
+<!-- Conversation -->
+\include{_ai/includes/conversation.md}
+
+<!-- Fritz Heider -->
+\include{_ai/includes/heider-simmel.md}
+\include{_ai/includes/baby-shoes.md}
+<!-- Conversation LLM -->
+\include{_ai/includes/conversation-computer.md}
+\include{_ai/includes/conversation-probability.md}
+\include{_ai/includes/human-computers-interacting.md}
+\include{_ai/includes/human-culture-interacting.md}
+\include{_data-science/includes/number-data-theatre.md}
+<!--\include{_psychology/includes/selective-attention-bias.md}-->
+<!--include{_data-science/includes/data-selection-attention-bias.md}-->
+\include{_ai/includes/conversation-llm.md}
+\include{_ai/includes/human-analogue-machines.md}
+
+\include{_physics/includes/richard-feynmann-doubt.md}
+
+<!-- Interfaces AI for Science -->
+<!--include{_ai/includes/interfaces-ai-for-science.md}-->
+
+\subsection{Conclusions}
+
+\notes{The probabilistic modelling community has evolved in an era where the assumption was that ambiguous conclusions are best shared with a (trained) professional through probabilities. Recent advances in generative AI offer the possibility of machines that have a better understanding of human subjective ambiguities and therefore machines that can summarise information in a way that can be interogated rather than just through a series of numbers.}
+
+\slides{* HAMs change how we share ambiguous information.
+* We need to think about how that effects our sharing of proabilities.}
+
+\thanks
+
+\references

_posts/2023-10-23-the-atomic-human-churchill.html

@@ -47,6 +47,7 @@
 week: 0
 reveal: 2023-10-23-the-atomic-human-churchill.slides.html
 transition: None
+ppt: 2023-10-03_the-atomic-human.pptx
 layout: talk
 categories:
 - notes
@@ -1284,7 +1285,7 @@ <h2 id="human-analogue-machine">Human Analogue Machine</h2>
 suggests the brain is neither a digital or an analogue computer, and the
 same can be said of the modern neural network systems that are being
 tagged as “artificial intelligence”.</p>
-<p>{I believe a better term for them is “human-analogue machines”,
+<p>I believe a better term for them is “human-analogue machines”,
 because what we have built is not a system that can make intelligent
 decisions from first principles (a rational approach) but one that
 observes how humans have made decisions through our data and
@@ -1308,23 +1309,12 @@ <h2 id="human-analogue-machine">Human Analogue Machine</h2>
 auto-regressive manner.</p>
 </div>
 </div>
-<ul>
-<li><p>A human-analogue machine is a machine that has created a feature
-space that is analagous to the “feature space” our brain uses to
-reason.</p></li>
-<li><p>The latest generation of LLMs are exhibiting this charateristic,
-giving them ability to converse.</p></li>
-</ul>
 <p>The perils of developing this capability include counterfeit people,
 a notion that the philosopher <a
 href="https://www.theatlantic.com/technology/archive/2023/05/problem-counterfeit-people/674075/">Daniel
 Dennett has described in <em>The Atlantic</em></a>. This is where
 computers can represent themselves as human and fool people into doing
 things on that basis.</p>
-<p>{* Perils of this include <em>counterfeit people</em>. * Daniel
-Dennett has described the challenges these bring in <a
-href="https://www.theatlantic.com/technology/archive/2023/05/problem-counterfeit-people/674075/">an
-article in The Atlantic</a>.</p>
 <h2 id="intellectual-debt">Intellectual Debt</h2>
 <div style="text-align:right">
 <span class="editsection-bracket" style="">[</span><span
@@ -1377,6 +1367,10 @@ <h2 id="intellectual-debt">Intellectual Debt</h2>
 1958)</span> the machine in a manner where humans can communicate
 without special training. This in turn offers the opportunity to
 overcome the challenge of <em>intellectual debt</em>.</p>
+<p>Despite the lack of interpretability of machine learning models, they
+allow us access to what the machine is doing in a way that bypasses many
+of the traditional techniques developed in statistics. But understanding
+this new route for access is a major new challenge.</p>
 <div class="figure">
 <div id="new-flow-of-information-4-figure" class="figure-frame">
 <object class="svgplot " data="https://inverseprobability.com/talks/./slides/diagrams//data-science/new-flow-of-information004.svg" width="70%" style=" ">

slides/2023-10-23-the-atomic-human-churchill.slides.html

@@ -856,13 +856,6 @@ <h2></h2>
 </section>
 <section id="section-56" class="slide level2">
 <h2></h2>
-<p>{I believe a better term for them is “human-analogue machines”,
-because what we have built is not a system that can make intelligent
-decisions from first principles (a rational approach) but one that
-observes how humans have made decisions through our data and
-reconstructs that process. Machine learning is more empiricist than
-rational, but now we n empirical approach that distils our evolved
-intelligence.</p>
 <div class="figure">
 <div id="human-analogue-machine-figure" class="figure-frame">
 <div class="centered" style="">
@@ -879,22 +872,42 @@ <h2></h2>
 </section>
 <section id="section-57" class="slide level2">
 <h2></h2>
-<p>{* Perils of this include <em>counterfeit people</em>. * Daniel
-Dennett has described the challenges these bring in <a
-href="https://www.theatlantic.com/technology/archive/2023/05/problem-counterfeit-people/674075/">an
-article in The Atlantic</a>.</p>
+<div class="figure">
+<div id="human-analogue-machine-2-figure" class="figure-frame">
+<object class="svgplot " data="https://inverseprobability.com/talks/./slides/diagrams//ai/processor-ham.svg" width="40%" style=" ">
+</object>
+</div>
+</div>
+<aside class="notes">
+</aside>
+<ul>
+<li><p>A human-analogue machine is a machine that has created a feature
+space that is analagous to the “feature space” our brain uses to
+reason.</p></li>
+<li><p>The latest generation of LLMs are exhibiting this charateristic,
+giving them ability to converse.</p></li>
+</ul>
 </section>
 <section id="section-58" class="slide level2">
 <h2></h2>
 <ul>
+<li>Perils of this include <em>counterfeit people</em>.</li>
+<li>Daniel Dennett has described the challenges these bring in <a
+href="https://www.theatlantic.com/technology/archive/2023/05/problem-counterfeit-people/674075/">an
+article in The Atlantic</a>.</li>
+</ul>
+</section>
+<section id="section-59" class="slide level2">
+<h2></h2>
+<ul>
 <li><p>But if correctly done, the machine can be appropriately
 “psychologically represented”</p></li>
 <li><p>This might allow us to deal with the challenge of
 <em>intellectual debt</em> where we create machines we cannot
 explain.</p></li>
 </ul>
 </section>
-<section id="section-59" class="slide level2">
+<section id="section-60" class="slide level2">
 <h2></h2>
 <div class="figure">
 <div id="new-flow-of-information-4-figure" class="figure-frame">
@@ -913,7 +926,7 @@ <h2></h2>
 science emerge when that relationship is being mediated by the machine.
 </aside>
 </section>
-<section id="section-60" class="slide level2">
+<section id="section-61" class="slide level2">
 <h2></h2>
 <div class="figure">
 <div id="new-flow-of-information-ham-figure" class="figure-frame">
@@ -925,7 +938,7 @@ <h2></h2>
 The HAM now sits between us and the traditional digital computer.
 </aside>
 </section>
-<section id="section-61" class="slide level2">
+<section id="section-62" class="slide level2">
 <h2></h2>
 <blockquote>
 <p>One thing is I can live with is doubt, and uncertainty and not
@@ -937,15 +950,15 @@ <h2></h2>
 <!-- Interfaces AI for Science -->
 <!--include{_ai/includes/interfaces-ai-for-science.md}-->
 </section>
-<section id="section-62" class="slide level2">
+<section id="section-63" class="slide level2">
 <h2></h2>
 <ul>
 <li>HAMs change how we share ambiguous information.</li>
 <li>We need to think about how that effects our sharing of
 proabilities.</li>
 </ul>
 </section>
-<section id="section-63" class="slide level2 scrollable">
+<section id="section-64" class="slide level2 scrollable">
 <h2 class="scrollable"></h2>
 <ul>
 <li><p>twitter: <a
@@ -961,7 +974,7 @@ <h2 class="scrollable"></h2>
 Damned Lies and Big Data</a></p></li>
 </ul>
 </section>
-<section id="section-64" class="slide level2 unnumbered scrollable">
+<section id="section-65" class="slide level2 unnumbered scrollable">
 <h2 class="unnumbered scrollable"></h2>
 <div id="refs" class="references csl-bib-body hanging-indent"
 role="list">