diff --git a/schedule.md b/schedule.md
index 65ecee9e882c7c653580ab801d21482a70c7ed2a..6abf090cefdba71cfc188bb6d9489179387d91fe 100644
--- a/schedule.md
+++ b/schedule.md
@@ -265,6 +265,45 @@ Finish making laser cuttable designs and print
{% enddetails %}
+{: .week}
+# Week 7 (11/6 - 11/10): Accessible AI
+{% details Learning Goals & Plan %}
+## Learning Goals
+- Sources of Bias in AI based systems
+- Applications of AI for Accessibility
+
+## Lecture Plan
+
+**Monday Slides** {% include slide.html title="AI and Accessibility" loc="bias-in-machine-learning.html" %}
+
+{: .homework} Required Reading and Reflection (for Wednesday)
+:
+- **Required: Respond to the Reading Questions and Preparation Requirements.**
+ - [Areas of Strategic Visibility: Disability Bias in Biometrics](https://arxiv.org/abs/2208.04712 - [Explaining Explanations: An Approach to Evaluating Interpretability of Machine Learning](https://www.cs.purdue.edu/homes/bb/nit/Lalana-Explainations%20of%20IAS.pdf)
+ - [Increasing Data Equity in Accessibility](https://arxiv.org/abs/2210.01902)
+- **If you want to go deeper**
+ - [The Future of Urban Accessibility for People with Disabilities: Data Collection, Analytic, Policy, and Tools](https://dl.acm.org/doi/fullHtml/10.1145/3517428.3550402)
+
+**Wednesday Slides** {% include slide.html title="Designing for and with people with disabilities" loc="designing.html" %}
+
+{: .homework} [Final Project Proposal](assignments/project-proposals.html) Assigned: Prepare your final project proposal (individual)
+:
+
+**Thursday: Section**: Practice with making data accessible
+
+
+## Friday 11/10: HOLIDAY
+{: .holiday}
+
+
+
+{: .week}
+
+
+{% enddetails %}
+
+# Module 3: Everything Everywhere All at Once
+
{: .draft}
# Everything after this is draft
diff --git a/slides/bias-in-machine-learning.html b/slides/bias-in-machine-learning.html
index b50f09345a770e56bf8ac705fa783e254688004d..163154a26c24b196fd08af09b8f05ce16729540f 100644
--- a/slides/bias-in-machine-learning.html
+++ b/slides/bias-in-machine-learning.html
@@ -1,13 +1,13 @@
---
layout: presentation
-title: Bias in Machine Learning --Week 6--
-description: Comparison of Assessment Approaches
+title: AI Accessibility Bias --Week 6--
+description: Discussion of Bias in AI
class: middle, center, inverse
---
background-image: url(img/people.png)
.left-column50[
-# Week 6: Bias in Machine Learning
+# Week 6: Automating Accessibility
{{site.classnum}}, {{site.quarter}}
]
@@ -27,33 +27,25 @@ class:
[//]: # (Outline Slide)
# Learning Goals for Today
-- What is Machine Learning (ML)?
+- How do AI & ML algorithms work?
-- What are the components of ML?
+- How do we minimize disability bias in automated systems?
- How do we collect data? Who do we collect the data from?
-- Is the data "good"?
+- How do we know whether the data is "good"?
-- How do we minimize disability bias?
---
-# Machine Learning
-
-
-
-.center[**But really, *what is it*?**]
-
----
-# Machine Learning
+# AI & Machine Learning
-Machine Learning changes the way we think about a problem.
+Both can change the way we think about a problem.
But *how*?
- What is the *traditional* approach to solve a problem?
-- How does Machine Learning solve a problem?
+- How does AI solve a problem?
---
@@ -87,8 +79,8 @@ Old Approach: Create software by hand
New Approach: Collect data and train algorithms
- Will still do the above, but will also have some functionality based
-on ML
-- *Collect lots of examples and train a ML algorithm*
+on AI
+- *Collect lots of examples and train a AI algorithm*
- *Statistical way of thinking*
---
@@ -126,28 +118,24 @@ classDef normal fill:#e6f3ff,stroke:#333,stroke-width:2px;
class U,C,A,I,S,E,Act,Act2 normal
</div>
]
----
-# This is *Machine Intelligence*
-
-Often used to process sensor data
-Goal is to develop systems that can improve
-performance with more experience
-- Can use "example data" as "experience"
-- Uses these examples to discern patterns
-- And to make predictions
+---
+# Basic Approach Of All AI
-Not really intelligent, just my word for Machine Learning, AI, Neural Programming, etc etc
+- Collect data (and lots and lots of it!)
+- Discern patterns
+- Make predictions
---
-# Components of Machine Intelligence
-
-- **Collect data (and lots and lots of it!)**
+# Pause and Discuss
+How could disability bias affect these?
+- Collect data (and lots and lots of it!)
- Discern patterns
-
- Make predictions
+(Post on [Ed]({{site.discussion}}/3805159)
+
---
# Data Collection
@@ -158,11 +146,13 @@ Not really intelligent, just my word for Machine Learning, AI, Neural Programmin
- Who do we collect data from?
---
-# Problems with Data
+# Problems with Data (1/2)
- System timeouts that are trained on movement speeds of <q>typical</q> people
- Biometrics that cannot function on a person who isn't still for long enough
- Inferencing about people that doesn't account for height; stamina; range of motion; or AT use (e.g. wheelchairs)
+---
+# Problems with Data (2/2)
When groups are historically
marginalized and underrepresented, this is
.quote[imprinted in the data that shapes AI
@@ -189,32 +179,34 @@ We may need to rethink <q>fairness</q> in terms of individual rather than group
Defining such unbiased metrics requires careful thought and domain knowledge, and scientific research will be essential to defining appropriate procedures for this.
----
-# Small Group Discussion [Post on Ed]({{site.discussion}}2514887)
+<!-- --- -->
+<!-- # Small Group Discussion [Post on Ed]({{site.discussion}}TBD) -->
-Who might be excluded in the data set you found?
+<!-- Who might be excluded in the data set you found? -->
-How was fairness measured in the data set you found, if it was discussed?
+<!-- How was fairness measured in the data set you found, if it was discussed? -->
-How would you go about testing for fairness in that data?
+<!-- How would you go about testing for fairness in that data? -->
---
-# Best Practices For Data Fairness
+# Best Practices For Data Fairness (1/2)
-How do we motivate and ethically compensate disabled people to give their data?
+- How do we motivate and ethically compensate disabled people to give their data?
-What should we communicate at data collection time?
+- What should we communicate at data collection time?
-Is the data collection infrastructure accessible? Does it protect sensitive information about participants adequately given the heterogeneous nature of disability?
+- Is the data collection infrastructure accessible? Does it protect sensitive information about participants adequately given the heterogeneous nature of disability?
-Does the meta data collected oversimplify disability? Who is labeling the data and do the have biases affecting labeling?
+---
+# Best Practices For Data Fairness (2/2)
+- Does the meta data collected oversimplify disability? Who is labeling the data and do the have biases affecting labeling?
- Whittaker (2019) discusses the example of clickworkers who label people
as disabled <q>based on a hunch</q>.
---
-# Components of Machine Intelligence
+# Basic Approach Of All AI
- Collect data (and lots and lots of it!)
@@ -237,20 +229,7 @@ they are about)
![:fa history] Can learn over time or train up front
---
-# Our Focus: Supervised Learning
-
-![:fa eye] *Supervised learning* (we have lots of examples of what should be
- predicted)
-
-![:fa eye-slash] *Unsupervised learning* (e.g. clustering into groups and inferring what
-they are about)
-
-![:fa low-vision] Can combine these (semi-supervised)
-
-![:fa history] Can learn over time or train up front
-
----
-# Supervised Learning
+# Machine Learning
.left-column50[
## Training Process
@@ -277,7 +256,7 @@ Symbolic requires feature engineering (humans deciding how to *summarize* data u
]
---
-# Supervised Learning
+# Large Language Models
.left-column50[
## Training Process
@@ -303,28 +282,11 @@ Neural approaches (e.g. ChatGPT) use massive amounts of data to train a network
]
---
-# Supervised Learning
-
-.left-column50[
-## Training Process
+# Both
-<div class="mermaid">
-graph TD
-
-L(Label) --> MI(Training Algorithm)
-D(Input Data) --> MI
-MI --> C(Symbolic/Neural Predictor)
-
-classDef normal fill:#e6f3ff,stroke:#333,stroke-width:2px;
-
-class D,U,C,A,I,S,E,Act,Act2 normal
-</div>
-]
-
-
-.right-column50[
## Prediction Process
+.left-column[
<div class="mermaid">
graph TD
@@ -335,82 +297,31 @@ classDef normal fill:#e6f3ff,stroke:#333,stroke-width:2px;
class P,D,U,C,A,I,S,E,Act,Act2 normal
</div>
-]
-
-
----
-# How do we Evaluate Predictors?
-
-Compare to Prior probabilities
-- Probability before any observations (ie just guessing)
-- Ex. ML classifier to guess if an animal is a cat or a ferret based on the ear location
- - Assume all pointy eared fuzzy creatures are cats (some percentage will be right)
-
-Compare to simplistic algorithms
-- Ex. Classifying cats vs ferrets based on size
-- Your model needs to do better than these too
-
-Surprising how often this doesn't happen in published work/before deployment
-
-???
-We did this to study gender's impact on academic authorship; doctors reviews
----
-# Adding Nuance
-.left-column50[
-## <q>Confusion Matrix</q>
-
-]
-.right-column50[
-Don't just measure accuracy (percent correct)
-
-Lots of other metrics based on false positives and negatives
-- Precision = TP / (TP+FP) Intuition: Of the positive items, how many right?
-- Recall = TP / (TP+FN) Intuition: Of all things that should have been positive, how many actually labeled correctly?
-- ... Many More
]
-
---
-# Using Proper Methods
-
-.left-column50[
-**Symbolic Methods Can Easily Overfit**
-
-When your ML model is too specific for data you have, it might not generalize well
-
-Best test is a data set you haven't seen before
-
-
-
-]
-
-.right-column50[
-**Neural Methods Can Have Hidden Biases**
+# How do we Evaluate Predictors/Predictions?
-
-
-]
-
----
-# Disability Biases to Watch Out For
-
-Norms are baked deeply into algorithms which are designed to learn about the most common cases. As human judgment is increasingly replaced by biometrics, *norms* become more strictly enforced.
+Norms are baked deeply into algorithms which are designed to learn about the most common cases. As human judgment is increasingly replaced by AI, *norms* become more strictly enforced.
- Do outliers face higher error rates?
- Do they disproportionately represent and misrepresent people with disability?
- How does this impact allocation of resources?
---
-# How does norming harm people with disabilities?
+# How does norming harm people with disabilities? (1/2)
Machine intelligence already being used to track allocation of assistive technologies, from CPAP machines for people with sleep apnea (Araujo 2018) to prosthetic legs (as described by Jullian Wiese in
Granta and uncovered in Whittaker et al 2019), deciding who is <q>compliant enough</q> to deserve them.
+---
+# How does norming harm people with disabilities? (2/2)
+
Technology may also fail to recognize that a disabled person is even present (Kane, 2020), thus <q>demarcating what it means to be a legible human and
whose bodies, actions, and lives fall outside... [and] remapping and calcifying the boundaries
of inclusion and marginalization</q> (Whittaker, 2019).
---
-# How does norming harm people with disabilities?
+# How does norming harm people with disabilities? (3/2)
Many biometric systems gatekeep access based on either individual identity, identity as a human, or class of human, such as <q>old enough to buy cigarettes.</q>
Examples:
@@ -419,7 +330,7 @@ Examples:
- at work, activity tracking may define <q>success</q> in terms that exclude disabled workers. (may also increase the likelihood of work-related disability, by forcing workers to work at maximal efficiency)
---
-# Components of Machine Intelligence
+# Basic Approach Of All AI
- Collect data (and lots and lots of it!)
@@ -428,99 +339,90 @@ Examples:
- **Make predictions**
---
-# Concerns at Prediction Time
+# Stopp and Discuss
+What do you think could go wrong at prediction time?
-Denial of insurance and medical care, or threaten employment (Whittaker,
+Add a second (Post on [Ed]({{site.discussion}}/3805238)
+
+
+---
+# Concerns at Prediction Time (1/5)
+
+- Denial of insurance and medical care, or threaten employment (Whittaker,
2019, p. 21).
-- HireVue, an AI based video
+ - HireVue, an AI based video
interviewing company has a patent on file to detect disability (Larsen, 2018).
-- This is illegal under the ADA, which
- - forbids asking about disability
+ - This is illegal under the ADA, which
+ - forbids asking about disability
status in a hiring process (42 U.S.C. § 12112(a))
- - forbids <q>using qualification
+ - forbids <q>using qualification
standards, employment tests or other selection criteria that screen out or tend to screen out
an individual with a disability</q> (42 U.S.C. § 12112(b)(6)).
---
-# Concerns at Prediction Time
+# Concerns at Prediction Time (2/5)
-Denial of insurance and medical care, or threaten employment
-
-Disability identification
-- Examples: detect Parkinsons from gait (Das, 2012), and mouse movement (Youngmann,
+- Denial of insurance and medical care, or threaten employment
+- Disability identification
+ - Examples: detect Parkinsons from gait (Das, 2012), and mouse movement (Youngmann,
2019); detecting autism from home videos (Leblanc, 2020).
-- What are the ethics of doing this without consent?
-- Many of these algorithms encode medical model biases
-
-Relatedly, failure to identify disability
-- Legally under the ADA, if you are treated as disabled, you are disabled. Yet biometrics cannot detect how people are treated.
-
----
-# Concerns at Prediction Time
-
-Denial of insurance and medical care, or threaten employment
+ - What are the ethics of doing this without consent?
+ - Many of these algorithms encode medical model biases
-Disability Identification / Failure to Identify
-
-Apps that Harm
-- Example: Training behaviors in <q>support</q> of autistic individuals without regard to debates about agency and independence of the target audience [Demo, 2017];
-- As with regular accessibility apps, AI based aps can harm, be disability dongles, etc
-- As with regular apps, AI based apps may not be accessible
+- Relatedly, failure to identify disability
+ - Legally under the ADA, if you are treated as disabled, you are disabled. Yet biometrics cannot detect how people are treated.
---
+# Concerns at Prediction Time (3/5)
-# Concerns at Prediction Time
-
-
+- Denial of insurance and medical care, or threaten employment
+- Disability Identification / Failure to Identify
+- Apps that Harm
+ - Example: Training behaviors in <q>support</q> of autistic individuals without regard to debates about agency and independence of the target audience [Demo, 2017];
+ - As with regular accessibility apps, AI based aps can harm, be disability dongles, etc
+ - As with regular apps, AI based apps may not be accessible
---
-# Concerns at Prediction Time
-Denial of insurance and medical care, or threaten employment
-
-Disability Identification / Failure to Identify
-
-Apps that Harm
-
-AI with Baked in Biases
-- Consequences of biased data and lack of control over training results more nuanced than just accuracy (as with headlines we just read)
-- Privacy can also be a concern.
- - For rare conditions, an algorithm may learn
+# Concerns at Prediction Time (4/5)
+- Denial of insurance and medical care, or threaten employment
+- Disability Identification / Failure to Identify
+- Apps that Harm
+- AI with Baked in Biases
+ - Consequences of biased data and lack of control over training results more nuanced than just accuracy (as with headlines we just read)
+ - Privacy can also be a concern.
+ - For rare conditions, an algorithm may learn
to recognize the disability, rather than the individual, reducing
security when used for access control, allowing multiple people with
similar impairments to access the same data.
---
-# Concerns at Prediction Time
-
-Denial of insurance and medical care, or threaten employment
+# Concerns at Prediction Time (5/5)
-Disability Identification / Failure to Identify
-
-Apps that Harm
-
-AI with Baked in Biases
-
-Transparency and Accountability
-- Power differences between builders and users
-- Representation of disabled people among builders
-- Algorithms that are not *interpretable* or *correctable*
-- Users of algorithms whose use them to enforce larger societal harms
-
----
-# Small Group Discussion [Post on Ed]({{site.discussion}}2515387)
-
-Revisit the data set you chose
-
-Do you know what sort of predictions it was used for if any?
-
-What possible harms could be done with those predictions?
-
-Reminder of our list
- Denial of insurance and medical care, or threaten employment
- Disability Identification / Failure to Identify
- Apps that Harm
- AI with Baked in Biases
- Transparency and Accountability
+ - Power differences between builders and users
+ - Representation of disabled people among builders
+ - Algorithms that are not *interpretable* or *correctable*
+ - Users of algorithms whose use them to enforce larger societal harms
+
+<!-- --- -->
+<!-- # Small Group Discussion [Post on Ed]({{site.discussion}}2515387) -->
+
+<!-- Revisit the data set you chose -->
+
+<!-- Do you know what sort of predictions it was used for if any? -->
+
+<!-- What possible harms could be done with those predictions? -->
+
+<!-- Reminder of our list -->
+<!-- - Denial of insurance and medical care, or threaten employment -->
+<!-- - Disability Identification / Failure to Identify -->
+<!-- - Apps that Harm -->
+<!-- - AI with Baked in Biases -->
+<!-- - Transparency and Accountability -->
---
diff --git a/slides/designing.html b/slides/designing.html
index d890ed087831ab16799d3bf5f793ba2d0e5076d1..6ad6f49dca3a17a5f20c936c6ca6fe81e28b9508 100644
--- a/slides/designing.html
+++ b/slides/designing.html
@@ -1,6 +1,6 @@
---
layout: presentation
-title: Evaluation --Week 4--
+title: Picking Problems --Week 7--
description: Designing for ad With People with Disabilities
class: middle, center, inverse
---