Make sure we have alt text for all images

[RuthBurns]

List of current alt text:

Welcome https://carpentries-incubator.github.io/machine-learning-librarians-archivists/01-welcome/index.html No images

Artificial Intelligence (AI) and Machine Learning (ML) in a nutshell

1) Image: Flow chart "Input, interpret, reason, decide, response, learn," Current Alt text? = Yes "alt="AI Process"

Machine Learning Modelling Concepts https://carpentries-incubator.github.io/machine-learning-librarians-archivists/03-AI-ML-concepts/index.html

2) Image: Screenshot of table with date highlighted and pop out boxes of how the data drills down. img src="../fig/ep-02-date_feature_engineering.png" alt="Date representations"

What is machine learning good at? https://carpentries-incubator.github.io/machine-learning-librarians-archivists/04-what-is-ml-good-at/index.html

3) Image with robot, and "input, feature maps, output" img src="../fig/ep-03-cnn-diagram.png" alt="Wikipedia CNN Diagram"

4) Image Classification- An image comparing image classication with object detection img src="../fig/ep-03-titanic_newspaper_cv.png" alt="Titanic Newspaper CV"

5) Activity: Photograph of an archive with drawers open showing many dead colourful birds. img src="../fig/ep-03-bird-collection.jpg" alt="USNM Bird Collection"

6) PixPlot image- a screenshot of image fields in the Meserve-Kunhardt Collection img src="../fig/ep-03-pixplot-github.png" alt="PixPlot screenshot from GitHub"

Understanding and managing bias https://carpentries-incubator.github.io/machine-learning-librarians-archivists/05-managing-data-bias/index.html

7)- Set of four wedding photos img src="../fig/ep-04-Google_Inclusive_Activity_1.png" alt="Four separate images depicting wedding ceremonies" 7b) The same image as 7 but with descriptive labelling below.

Applying Machine Learning https://carpentries-incubator.github.io/machine-learning-librarians-archivists/06-applying-machine-learning/index.html

8) Notepad scribble of processing pages to machine readable text img src="../fig/05-business-use-ocr.png" alt="Illustration of a business problem"

9) Notepad scribble of text to catalogue process img src="../fig/05-predictions-to-actions.png" alt="Two possible ways of using Machine Learning Predictions"

10) Notepad scribble of Digitized newspaper processes img src="../fig/05-model-choices.png" alt="Illustration of possible model choices"

11) Notepad scribble of photograph to preservation process img src="../fig/05-using-model.png" alt="A simplified GLAM workflow"

The Machine Learning Ecosystem https://carpentries-incubator.github.io/machine-learning-librarians-archivists/07-ecosystem/index.html No images

[PhilReedData]

Have we checked that all these images' licenses are included too? (CC BY 4.0 or similar)

[PhilReedData]

Define what makes the alt text fit for purpose. I've checked the Carpentries guidance.

We should check that the surrounding text (before or after) the image is aligned to. For example, describe the trays of birds in episode 5.

Flow of AI in episode1 could expand to say more.

[RuthBurns]

Artificial Intelligence (AI) and Machine Learning (ML) in a nutshell https://carpentries-incubator.github.io/machine-learning-librarians-archivists/02-AI-in-a-Nutshell/index.html

Image: Flow chart "Input, interpret, reason, decide, response, learn," Current Alt text? = Yes "alt="AI Process" Proposed alt text: A flowchart demonstrating AI Process

Machine Learning Modelling Concepts https://carpentries-incubator.github.io/machine-learning-librarians-archivists/03-AI-ML-concepts/index.html

img src="../fig/ep-02-date_feature_engineering.png" alt="Date representations" Proposed alt text: Table showing the data represented within a simple date, such as day of the week.

What is machine learning good at? https://carpentries-incubator.github.io/machine-learning-librarians-archivists/04-what-is-ml-good-at/index.html

img src="../fig/ep-03-cnn-diagram.png" alt="Wikipedia CNN Diagram" Licence CC BY-SA 4.0 Wikipedia Proposed alt text: Diagram depicting layers of a convolutional neural network

img src="../fig/ep-03-titanic_newspaper_cv.png" alt="Titanic Newspaper CV" Proposed Alt Text: An image of a newspaper front page comparing image classication with object detection. Proposed caption: Image classification identifies that there is 90% probability the image is of a newspaper, and 70% it is the New York Tribune. Object detection identifies parts of the newspaper, such as nameplate and headline.

img src="../fig/ep-03-bird-collection.jpg" alt="USNM Bird Collection" Proposed alt text: An overhead view of hundreds of drawers containing specimen trays of preserved birds of varying plumage, with one species per drawer.

img src="../fig/ep-03-pixplot-github.png" alt="PixPlot screenshot from GitHub" Proposed Alt Text- A collection of images in PixPlot clumped by similarity,

Understanding and managing bias https://carpentries-incubator.github.io/machine-learning-librarians-archivists/05-managing-data-bias/index.html

Set of four wedding photos img src="../fig/ep-04-Google_Inclusive_Activity_1.png" alt="Four separate images depicting wedding ceremonies" 7b) The same image as 7 but with descriptive labelling below.

Proposed Alt text: Four photos showing respectively 1) a man and a women in white clothes 2) a woman in a white dress, 3) a man in a tuxedo with a woman in a white dress, 4) a man in agbada and fila and a woman in wrapper, blouse and headpiece

Proposed Alt text: the same four images as previous image with the labels assigned by machine learning, demonstrating its biases.

Text that follows image should read as follows: The images are all from wedding photographs (donated by Googlers) for the Open Images dataset. The classifier’s label predictions are: 1) ceremony, wedding, bride, man, groom, woman, dress 2) bride, ceremony, wedding, dress, woman 3) ceremony, bride, wedding, man, groom, woman dress 4) person people What do you notice about the classifier text for (4)? How well did you do compared to the classifier’s label predictions?

Applying Machine Learning https://carpentries-incubator.github.io/machine-learning-librarians-archivists/06-applying-machine-learning/index.html

Notepad scribble of processing pages to machine readable text img src="../fig/05-business-use-ocr.png" alt="Illustration of a business problem" Proposed updates alt text- Diagram of workflow for using a paid for service to convert images to machine readable text.

Notepad scribble of text to catalogue process img src="../fig/05-predictions-to-actions.png" alt="Two possible ways of using Machine Learning Predictions" Proposed Alt text: Diagram of Workflow using machine learning to classify documents by language, with human oversight.

Notepad scribble of Digitized newspaper processes. img src="../fig/05-model-choices.png" alt="Illustration of possible model choices" Recommend splitting image and matching to bullet point order Proposed Image A) Diagram of machine learning models which are ‘deeply’ embedded into applications Proposed Image B) Diagram of Workflow using external closed source blackbox service. Proposed Image C) Diagram of Workflow of building your own system using existing resource components.

Diagram of Workflow using an application that incorporates AI. Notepad scribble of photograph to preservation process img src="../fig/05-using-model.png" alt="A simplified GLAM workflow" Proposed Alt text: Diagram of a simplfied GLAM workflow for publishing and preserving images.

[marionwalton]

I followed up on the Yoruba wedding alt text and found the correct names for the woman's attire as well

Four photos showing respectively 1) a man and a women in white clothes 2) a woman in a white dress, 3) a man in a tuxedo with a woman in a white dress, 4) a man in agbada and fila and a woman in iro ati buba with a gele. https://carpentries-incubator.github.io/machine-learning-librarians-archivists/05-managing-data-bias/index.html