Identify specific AI technologies that can address critical records and archives challenges
Determine the benefits and risks of using AI technologies on records and archives
Ensure that archival concepts and principles inform the development of responsible AI
Validate outcomes from Objective 3 through case studies and demonstrations
Muhammad focussed on trustworthiness as an issue for Archives. They are looking at using AI to assess and verify the authenticity of Archives through time. The essential research question: Can we develop artificial intelligence for carrying out competently and efficiently all records and archives functions while respecting the nature and ensuring the continuing trustworthiness of the record.
He noted that a fundamental difference between analog and digital records is the fact that analogue materials can be proven and verified on face value and rarely need extrinsic evidence. However for digital materials, extrinsic elements such as metadata are needed. They rely on ‘circumstantial’ evidence such as the integrity of the hosting system as well as the politics, procedures and technology surrounding the digital record.
Muhammad suggests that off-the-shelf tools are not well suited to archives, so within the Archives profession we will have to develop the systems ourselves. We are the only ones who know what to do because we are the professionals. Developers need to talk to archives professionals to find out what they want and design appropriate AI tools for them. The tools need to respect the trustworthiness of the records. The project is looking to influence the development of responsible tools.
The project looks to provide a wealth of tools and code. A very important aspect of the project is training the community. Muhammad suggested that the Archives profession will have to do a great deal of training to engage with AI tools and its possibilities.
Linking AI to Archives and Records, Peter Sullivan
The aim of the talk was to look at combining archival concepts and principles with AI. Peter used the lens of Diplomatic to consider AI solutions and how AI may interact with different components of the record including the context, act, persons, procedure, form and archival bond. Which parts of the archival record are impacted by AI and how does this inform the design of AI tools that respect diplomatic theory?
The most important component is the ‘archival bond’ which covers how aspects of records are related to each other. AI may be poor at looking at records in context of other records, and may not be able to respect the archival bond. Also, AI may not respect the context of the creation of the records and may not be aware of different levels of appraisal used.
AI may be helpful where there are different variations of names and fuzzy matching can be used to reconcile names. This aligns with the Archives Hub Names project. Dealing with records in aggregate may be somewhere AI is able to help, using topic modelling and clustering techniques. This is a use case we have identified ourselves and something we are looking at with the Archives Hub Labs Project. Finally he mentioned the interesting question of how we will archive the artefacts of AI developments themselves.
Model for an AI-Assisted Digitisation Project, Peter Sullivan
Peter talked about how AI is being used to help with the archiving of audio recordings, providing AI generated metadata enrichment. He noted this is very time-consuming to do by hand. Different types of recordings create very different challenges to AI to analyse . For UNESCO audio they are using four models, one for language translation and three for text extraction and text summarisation.
AI and Archives: Basic Requirements, Pilar Campos and Eloi Puertas
The project is aiming to provide a resource for archive professionals to assess AI solutions to help guide decision-making and create recommendations. They will provide a check list to assess AI tool performance. The rationale behind this is that there is a huge amount of interest and concern regarding AI, but a scarcity of implementation examples, along with a lack of knowledge of AI solutions for the professionals in the archives domain. There is also a degree of mistrust of the results of AI.
The expected results of the project are to provide AI knowledge in the archive domain and a list of potential risks for archivists. A SWOT analysis about AI from the Archives viewpoint will be provided, along with an assessment of the balance between our expectations of risk.
Automated Transcription: Palaeography and AI, Thiara Alves and Leonardo Fontes
The talk was essentially about using AI for automated transcription. The speakers talked about using Transkribus for transcription of text from images of documents. They found that most algorithms weren’t good at detecting old versions of Portuguese and Spanish words. The speakers felt that the context provided by the archivist was necessary for the transcribers transcriptions to be useful.
First Steps and Main Expectations from CRDI’s Experience of AI, David Inglésias
David talked about a project looking at being able to search images that haven’t been catalogued, so they don’t have metadata unless it is created by using AI. This ability is very useful for a photographic archive. They work with the Europeana Kaleidoscope project to attempt to provide archival context for images.
AI also allows for innovative new approaches to presenting photographs in addition to the standard historical ways of doing so. AI can be used for clustering photos that appear to be similar or related in someway. This could be something that the Archives Hub could look at also.
In this post I will go through the steps we took to create a human labelled dataset (i.e. naming objects within images), applying the labels to bounding boxes (showing where the objects are in the image) in order to identify objects and train an ML model. Note that the other approach, and one we will talk about in another post, is to simply let a pre-trained tool do the work of labelling without any human intervention. But we thought that it would be worthwhile to try the human labelling out before seeing what the out-of-the-box results are.
I used Amazon SageMaker for this work. In SageMaker you can set up a labelling job using the Ground Truth service, by giving the location of the source material – in this case, the folder containing the Jamson photographs. Images have to be jpg, or png, so if you have tif images, for example, they have to be converted. You give the job a name and provide the location of the source material (in our case an S3 bucket, which is the Amazon Simple Storage Service).
I then decide on my approach. I trained the algorithm with a random sample of images from this collection. This is because I wanted this sample to be a subset of the full Jamson Archive dataset of images we are working with. We can then use the ML model created from the subset to make object detection predictions for the rest of the dataset.
Once I had these settings completed, I started to create the labels for the ‘Ground Truth’ job. You have to provide the list of labels first of all from which you will select individual labels for each image. You cannot create the labels as you go. This immediately seemed like a big constraint to me.
I went through the photographs and decided upon the labels – you can only add up to 50 labels. It is probably worth noting here that ‘label bias’ is a known issue within machine learning. This is where the set of labelled data is not fully representative of the entirety of potential labels, and so it can create bias. This might be something we come back to, in order to think about the implications.
I chose to add some fairly obvious labels, such as boat or church. But I also wanted to try adding labels for features that are often not described in the metadata for an image, but nonetheless might be of interest to researchers, so I added things like terraced house, telegraph pole, hat and tree, for example.
Once you have the labels, there are some other options. You can assign to a labelling team, and make the task time bound, which might be useful for thinking about the resources involved in doing a job like this. You can also ask for automated data labelling, which does add to the cost, so it is worth considering this when deciding on your settings. The automated labelling uses ML to learn from the human labelling. As the task will be assigned to a work team, you need to ensure that you have the people you want in the team already added to Ground Truth.
Those assigned to the labelling job will receive an email confirming this and giving a link to access to the labelling job.
You can now begin the job of identifying objects and applying labels.
First up I have a photograph showing rowing boats. I didn’t add the label ‘rowing boat’ as I didn’t go through every single photograph to find all the objects that I might want to label, so not a good start! ‘Boat’ will have to do. As stated above, I had to work with the labels that I created, I can’t add more labels at this stage.
I added as many labels as I could to each photograph, which was a fairly time intensive exercise. For example, in the image below I added not only boat and person but also hat and chimney. I also added water, which could be optimistic, as it is not really an object that is bound within a box, and it is rather difficult to identify in many cases, but it’s worth a try.
I can zoom in and out and play with exposure and contrast settings to help me identify objects.
Here is another example where I experimented with some labels that seem quite ambitious – I tried shopfront and pavement, for example, though it is hard to classify a shop from another house front, and it is hard to pin-point a pavement.
The more I went through the images, drawing bounding boxes and adding labels, the more I could see the challenges and wondered how the out-of-the-box ML tools would fare identifying these things. My aim in doing the labelling work was partly to get my head into that space of identification, and what the characteristics are of various objects (especially objects in the historic images that are common in archive collections). But my aim was also to train the model to improve accuracy. For an object like a chimney, this labelling exercise looked like it might be fruitful. A chimney has certain characteristics and giving the algorithm lots of examples seems like it will improve the model and thus identify more chimneys. But giving the algorithm examples of shop fronts is harder to predict. If you try to identify the characteristics, it is often a bay window and you can see items displayed in it. It will usually have a sign above, though that is indistinct in many of these pictures. It seems very different training the model on clear, full view images of shops, as opposed to the reality of many photographs, where they are just part of the whole scene, and you get a partial view.
There were certainly some features I really wanted to label as I went along. Not being able to do this seemed to be a major shortcoming of the tool. For example, I thought flags might be good – something that has quite defined characteristics – and I might have added some more architectural features such as dome and statue, and even just building (I had house, terraced house, shop and pub). Having said that, I assume that identifying common features like buildings and people will work well out-of-the-box.
Running a labelling job is a very interesting form of classification. You have to decide how thorough you are going to be. It is more labour intensive than simply providing a description like ‘view of a street’ or ‘war memorial’. I found it elucidating as I felt that I was looking at images in a different way and thinking about how amazing the brain is to be able to pick out a rather blurred cart or a van or a bicycle with a trailer, or whatever it might be, and how we have all these classifications in our head. It took more time than it might have done because I was thinking about this project, and about writing blog posts! But, if you invest time in training a model well, then it may be able to add labels to unlabelled photographs, and thus save time down the line. So, investing time at this point could reap real rewards.
In the above example, I’ve outlined an object that i’ve identified as a telegraph pole. One question I had is whether I am are right in all of my identifications, and I’m sure there will be times when things will be wrongly identified. But this is certainly the type of feature that isn’t normally described within an image, and there must be enthusiasts for telegraph poles out there! (Well, maybe more likely historians looking at communications or the history of the telephone). It also helps to provide examples from different periods of history, so that the algorithm learns more about the object. I’ve added a label for a cart and a van in this photo. These are not all that clear within the image, but maybe by labelling less distinct features, I will help with future automated identification in archival images.
I’ve added hat as a label, but it strikes me that my boxes also highlight heads or faces in many cases, as the people in these photos are small, and it is hard to distinguish hat from head. I also suspect that the algorithm might be quite good with hats, though I don’t yet know for sure.
I used ‘person’ as a label, and also ‘child’, and I tended not to use ‘person’ for ‘child’, which is obviously incorrect, but I thought that it made more sense to train the algorithm to identify children, as person is probably going to work quite well. But again, I imagine that person identification is going to be quite successful without my extra work – though identifying a child is a rather more challenging task. In the end, it may be that there is no real point in doing any work identifying people as that work has probably been done with millions of images, so adding my hundred odd is hardly going to matter!
I had church as a label, and then used it for anything that looked like a church, so that included Beverly Minster, for example. I couldn’t guarantee that every building I labelled as a church is a church, and I didn’t have more nuanced labels. I didn’t have church interior as a label, so I did wonder whether labelling the interior with the same label as the exterior would not be ideal.
I was interested in whether pubs and inns can be identified. Like shops, they are easy for us to identify, but it is not easy to define them for a machine.
A pub is usually a larger building (but not always) with a sign on the facade (but not always) and maybe a hanging sign. But that could be said for a shop as well. It is the details such as the shape of the sign that help a human eye distinguish it. Even a lantern hanging over the door, or several people hanging around outside! In many of the photos the pub is indistinct, and I wondered whether it is better to identify it as a pub, or whether that could be misleading.
I found that things like street lamps and telegraph poles seemed to work well, as they have clear characteristics. I wanted to try to identify more indistinct things like street and pavement, and I added these labels in order to see if they yield any useful results.
I chose to label 10% of the images. That was 109 in total, and it took a few hours. I think if I did it again I would aim to label about 50 for an experiment like this. But then the more labels you provide, the more likely you will get results.
The next step will be to compare the output using the Rekognition out of the box service with one trained using these labels. I’m very interested to see how the two compare! We are very aware that we are using a very small labelled dataset for training, but we are using the transfer learning approach that builds upon existing models, so we are hopeful we may see some improvement in label predications. We are also working on adding these labels to our front end interface and thinking about how they might enhance discoverability.
Thanks to Adrian Stevenson, one of the Hub Labs team, who took me through the technical processes outlined in this post.
There are many ways of utilising the International Image Interoperability framework (IIIF) in order to deliver high-quality, attributed digital objects online at scale. One of the exploratory areas focused on in Images and Machine Learning – a project which is part of Archives Hub Labs – is how to display the context of the archive hierarchy using IIIF alongside the digital media.
Two of the objectives for this project are:
to explore IIIF Manifest and IIIF Collection creation from archive descriptions.
to test IIIF viewers in the context of showing the structure of archival material whilst viewing the digitised collections.
We have been experimenting with two types of resource from the IIIF Presentation API. The IIIF Manifest added into the Mirador viewer on the collection page contains just the images, in order to easily access these through the viewer. This is in contrast to a IIIF Collection, which we have been experimenting with. The IIIF Collection includes not only the images from a collection but also metadata and item structure within the IIIF resource. It is defined as a set of manifests (or ‘child’ collections) that communicate hierarchy or gather related things (for example, a set of boxes that each have folders within them, and photographs within those folders). We have been testing whether this has the potential to represent the hierarchy of an archival structure within the IIIF structure.
Creating a User Interface
Since joining the Archives Hub team, one of the areas I’ve been involved in is building a User Interface for this project that allows us to test out the different ways in which we can display the IIIF Images, Manifests and Collections using the IIIF Image API and the IIIF Presentation API. Below I will share some screenshots from my progress and talk about my process when building this User Interface.
This web application is currently a prototype and further development will be happening in the future. The programming language I am using is Typescript. I began by creating a Next.js React application and I am also using Tailwind CSS for styling. My first task was to use the Mirador viewer to display IIIF Collections and Manifests, so I installed the mirador package into the codebase. I created dynamic pages for every contributor to display their collections.
I also created dynamic collection pages for each collection. Included on the left-hand side of a collection page is the archives hub record link and the metadata about the collection taken from the archival EAD data – these sections displaying the metadata can be extended or hidden. The right-hand side of a collection page features aMirador viewer. A simple IIIF Manifest has been added for all of the images in each collection. This Manifest is used to help quickly navigate through and browse the images in the collection.
Mirador has the ability to display multiple windows within one workspace. This is really useful for comparison of images side-by-side. Therefore, I have also created a ‘Compare Collections’ page where two Manifests of collection images can be compared side-by-side. I have configured two windows to display within one Mirador viewer. Then, two collections can be chosen for comparison using the dropdown select boxes seen in the image below.
There are three key next steps for developing the User Interface –
We have experimented with the Mirador viewer, and now we will be looking at how the Universal Viewer handles IIIF Collections.
From the workshop feedback and from our exploration with the display of images, we will be looking at how we can offer an alternative experience of these archival images – distinct from their cataloguing hierarchy – such as thematic digital exhibitions and linking to other IIIF Collections and Manifests that already exist.
As part of the Machine Learning aspect of this project, we will be utilising the additional option to add annotations within the IIIF resources, so that the ML outputs from each image can be added as annotations and displayed in a viewer.
Labs IIIF Workshop
We recently held a workshop with the Archives Hub Labs project participants in order to get feedback on viewing the archive hierarchy through these IIIF Collections, displayed in a Mirador viewer. In preparation for this workshop, Ben created a sample of IIIF Collections using the images kindly provided by the project participants and the archival data related to these images that is on the Archives Hub. These were then loaded into the Mirador viewer so our workshop participants could see how the collection hierarchy is displayed within the viewer. The outcomes of this workshop will be explored in the next Archives Hub Labs blog post.
Thank you to Cardiff University, Bangor University, Brighton Design Archives at the University of Brighton, the University of Hull, the Borthwick Institute for Archives at the University of York, Lambeth Palace (Church of England) and Lloyds Bank for providing their digital collections and for participating in Archives Hub Labs.
For our Machine Learning experiments we are using Amazon Web Services (AWS). We thought it would be useful to explain what we have been doing.
AWS, like most Cloud providers, gives you access to a huge range of infrastructure, services and tools. Typically, instead of having your own servers physically on your premises, you instead utlitise the virtual servers provided in the Cloud. The Cloud is a cost effective solution, and in particular it allows for elasticity; dynamically allocating resources as required. It also provides a range of features, and that includes a set of Machine Learning services and tools.
One of the services available is Amazon Rekognition. This is what we have used when writing our previous blog posts.
One of the things Rekognition does is object detection. We have written about using Rekognition in a previous post.
Our initial experiments were done on the basis of uploading single images at a time and looking at the output. The next step is to work out how to submit a batch of images and get output from that. AWS doesn’t have an interface that allows you to upload a batch. We have batches of images stored in the Cloud (using the ‘S3’ service), and so we need to pass sets of images from S3 to the Rekognition service and store the resulting label predictions (outputs). We also need to figure out how to provide these predictions to our contributors in a user friendly display.
After substantial research into approaches that we could take, we decided to use the AWS Lambda and DynamoDB services along with Rekognition and S3. Lambda is a service that allows you to run code without having to set up the virtual machine infrastructure (it is often referred to as a serverless approach). We used some ‘blueprint’ Lambda code (written in Python) as the basis, and extended it for our purposes.
Using something like AWS does not mean that you get this type of facility out of the box. AWS provides the infrastructure and the interfaces are reasonably user friendly, but it does not provide a full blown application for doing Machine Learning. We have to do some development work in order to use Rekognition, or other ML tools, for a set of images.
Lambda is set up so the code will run every time an image is placed in the S3 bucket. It then passes the output (label prediction) to another AWS service, called DynamoDB, which is a ‘NoSQL’ database.
In the above image you can see an excerpt from the output from running the Lambda code. This is for image U DX336-1-6.jpg (see below) and it has predicted ‘tree’ with a confidence level of 94.51 percent. Ideally we wanted to add the ‘bounding box’ which provides the co-ordinates for where the object is within the image.
We spent quite a bit of time trying to figure out how to add bounding boxes, and eventually realised that they are only added for some objects – Amazon Rekognition Image and Amazon Rekognition Video can return the bounding box for common object labels such as cars, furniture, apparel or pets, but the information isn’t returned for less common object labels. Quite how things are classed as more or less common is not clear. At the moment we are working on passing the bounding box information (when there is any) to our database output.
Clearly for this image, it would be useful to have ‘memorial’ and ‘cross’ as label predictions, but these terms are absent. However, sometimes ML can provide terms that might not be used by the cataloguer, such as ‘tree’ or ‘monument’.
So we now have the ability to submit a batch of images, but currently the output is in JSON (the above output table is only provided if you upload the image individually). We are hoping to read the data and place the labels into our IIIF development interface.
The next step is to create a model using a subset of the images that our participants have provided. A key thing to understand is that in order to train a model so that it makes better predictions you need to provide labelled images. Therefore, if you want to try using ML, it is likely that part of the ML journey will require you to undertake a substantial amount of labelling if you don’t already have labelled images. Providing labelled content is the way that the algorithm learns. If we provided the above image and a batch of others like it and included a label of ‘memorial’ then that would make it more likely that other non-labelled images we input would be identified correctly. We could also include the more specific label ‘war memorial’ – but it would seem like a tall order for ML to distinguish war memorials from other types. Having said that, the fascinating thing is that often machines learn to detect patterns in a way that surpasses what humans can achieve. We can only give it a go and see what we get.
Thanks to Adrian Stevenson, one of the Hub Labs team, who took me through the technical processes outlined in this post.
As part of the Archives Hub Labs ‘Images and Machine Learning’ project we are currently exploring the challenges around implementing IIIF image services for archival collections, and also for Archives Hub more specifically as an aggregator of archival descriptions. This work is motivated by our desire to encourage the inclusion of more digital content on Archives Hub, and to improve our users’ experience of that content, in terms of both display and associated functionality.
Before we start to report on our progress with IIIF, we thought it would be useful to capture some of our current ideas and objectives with regards to the presentation of digital content on Archives Hub. This will help us to assess at later stages of the project how well IIIF supports those objectives, since it can be easy to get caught up in the excitement of experimenting with new technologies and lose sight of one’s starting point. It will also help our audience to understand how we’re aiming to develop the Hub, and how the Labs project supports those aims.
Crucial part of modern research and engagement with collections, especially after the pandemic
Another route into archives for researchers
Contributes to making archives more accessible
Will enable us to create new experiences and entry points within Archives Hub
To support contributing archives which can’t host or display content themselves
The Current Situation
At the moment our contributors can include digital content in their descriptions on Archives Hub. They add links to their descriptions prior to publication, and they can do this at any level, e.g. ‘item’ level for images of individually catalogued objects, or maybe ‘fonds’ or ‘collection’ level for a selection of sample images. If the links are to image files, these are displayed on the Hub as part of the description. If the links are to video or audio files, or documents, we just display a link.
There are a few disadvantages to this set up: it can be a labour-intensive process adding individual links to descriptions; links often go dead because content is moved, leading to disappointment for researchers; and it means contributing archives need to be able to host content themselves, which isn’t always possible.
Where images are included in descriptions, these are embedded in the page as part of the description itself. If there are multiple images they are arranged to best fit the size of the screen, which means their order isn’t preserved.
If a user clicks on an image it is opened in a pop out viewer, which has a zoom button, and arrows for browsing if there is more than one image.
The embedded image and the viewer are both quite small, so there is also a button to view the image in fullscreen.
The viewer and the fullscreen option both obscure all or part of the decription itself, and there is no descriptive information included around the image other than a caption, if one has been provided.
As you can see the current interface is functional, but not ideal. Listed below are some of the key things we would like to look at and improve going forwards. The list is not intended to be exhaustive, but even so it’s pretty long, and we’re aware that we might not be able to fix everything, and certainly not in one go.
Documenting our aims though is an important part of steering our innovations work, even if those aims end up evolving as part of the exploration process.
Display and Viewing Experience
❐ The viewer needs updating so that users can play audio and video files in situ on the Hub, just as they can view images at the moment. It would be great if they could also read documents (PDF, Word etc).
❐ Large or high-resolution image files should load more quickly into the viewer.
❐ The viewer should also include tools for interacting with content, e.g. for images: zoom, rotate, greyscale, adjust brightness/contrast etc; for audio-visual files: play, pause, rewind, modify speed etc.
❐ When opened, any content viewer should expand to a more usable size than the current one.
❐ Should the viewer also support the display of descriptive information around the content, so that if the archive description itself is obscured, the user still has context for what they’re looking at? Any viewer should definitely clearly display rights and licensing information alongside content.
Search and Navigation
❐ The Archives Hub search interface should offer users the option to filter by the type of digital content included in their search results (e.g. image, video, PDF etc).
❐ The search interface should also highlight the presence of digital content in search results more prominently, and maybe even include a preview?
❐ When viewing the top level of a multi-level description, users should be able to identify easily which levels include digital content.
❐ Users should also be able to jump to the digital content within a multi-level description quickly – possibly being able to browse through the digital content separately from the description itself?
❐ Users should be able to begin with digital content as a route into the material on Archives Hub, rather than only being able to search the text descriptions as their starting point.
❐ Perhaps Archives Hub should offer some form of hosting service, to support archives, improve availability of digital content on the Hub, and allow for the development of workflows around managing content?
❐ Ideally, we would also develop a user-friendly method for linking content to descriptions, to make publishing and updating digital content easy and time-efficient.
❐ Any workflows or interfaces for managing digital content should be straightforward and accessible for non-technical staff.
❐ The service could give contributors access to innovative but sustainable tools, which drive engagement by highlighting their collections.
❐ If possible, any resources created should be re-usable within an archive’s own sites or resources – making the most of both the material and the time invested.
❐ We could look at offering options for contributors to curate content in creative and inventive ways which aren’t tied to cataloguing alone, and which offer alternative ways of experiencing archival material for users.
❐ It would be exciting for users to be able to ‘collect’, customise or interact with content in more direct ways. Some examples might include:
Creating their own collections of content
Creating annotations or notes
Publicly tagging or commenting on content
❐ Develop the experience for users with things like: automated tagging of images for better search; providing searchable OCR scanned text for text within images; using the tagging or classification of content to provide links to information and resources elsewhere.
One of the challenges that we face with our Labs project is presentation of the Machine Learning results. We thought there would be many out of the box tools to help with this, but we have not found this to be the case.
If we use the AWS console Rekognition service interface for example, we get presented with results, but they are not provided in a way that will readily allow us and our project participants to assess them. Here is a screenshot of an image from Cardiff University – an example of out of the box use of AWS Rekognition:
This is just one result – but we want to present the results from a large collection of images. Ideally we would run the image recognition on all of the Cardiff images, and/or on the images from one collection, assess the results within the project team and also present them back to our colleagues at Cardiff.
The ML results are actually presented in JSON:
Here you can see some of the terms identified and the confidence scores.
These particular images, from the University archive, are catalogued to item level. That means they may not benefit so much from adding tags or identifying objects. But they are unlikely to have all the terms (or ‘labels’ in ML parlance) that the Rekognition service comes up with. Sometimes the things identified are not what a cataloguer would necessarily think to add to a description. The above image is identified as ‘outdoors’, ‘ground’ and ‘soil. These terms could be useful for a researcher. Just identifying photographs with people in them could potentially be useful.
Another example below is of a printed item – a poem.
Strange formatting of the transcript aside, the JSON below shows the detected text (squirrels), confidence and area of the image where the word is located.
If this was provided to the end user, then anyone interested in squirrels in literature (surely there must be someone…) can find this digital content.
But we have to figure out how to present results and what functionality is required. It reminds me of using Open Refine to assess person name matches. The interface provides for a human eye to assess and confirm or reject the results.
We want to be able to lead discussions with our contributors on the usefulness, accuracy, bias – lack of bias – and peculiarities of machine learning, and for that a usable interface is essential.
How we might knit this in with the Hub description is something to consider down the line. The first question is whether to use the results of ML at all. However, it is hard to imagine that it won’t play a part as it gets better at recognition and classification. Archvists often talk about how they don’t have time to catalogue. So it is arguable that machine learning, even if the results are not perfect, will be an improvement on the backlogs that we currently have.
AWS Rekognition tools
We have thought about which tools we would like to use and we are currently creating a spreadsheet of the images we have from our participants and which tools to use with each group of images.
Some tools may seem less likely, for example, image moderation. But with the focus on ethics and sensitive data, this could be useful for identifying potentially offensive or controversial images.
The Image Moderation tool recognises nudity in the above image.
This could be carried through to the end user interface, and a user could click on ‘view content’ if they chose to do so.
The image moderation tool may classify images art images as sensitive when they are very unlikely to cause offence. The tools may not be able to distinguish offensive nudity from classical art nudity. With training it is likely to improve, but when you think about it, it is not always an easy line for a human to draw.
Face comparison could potentially be useful where you want to identify individuals and instances of them within a large collection of photographs for example, so we might try that out.
However, we have decided that we won’t be using ‘celebrity recognition’, or ‘PPE detection’ for this particular project!
Text and Images
We are particularly interested in text and in text within images. It might be a way to connect images, and we might be able to pull the text out to be used for searching.
Suffice to say that text will be very variable. We ran Transkribus Lite on some materials.
We compared this to use of AWS Text Rekognition.
These examples illustrate the problem with handwritten documents. Potentially the model could be trained to work better for handwriting, but this may require a very large amount of input data given the variability of writing styles.
Transkribus has transcribed this short typescript text from the same archive well. One word ‘house’ has been transcribed as ‘housd’ and ‘idea’ caused a formatting issue, but overall a good result.
The above example is Transkribus Lite on a poster from the University of Brighton Design Archives. In archives, many digital items are images with text – particularly collections of posters or flyers. Transkribus has not done well with this (though this is just using the Lite version out of the box).
We also tried this with the AWS Rekognition Text tool, and it worked well.
Another example of images with text is maps and plans.
Above are two examples of places identified from the plan output in JSON. If we can take these outputs and add them to our search interface, an end user could search for ‘clerkenwell’ or ‘northampton square’ and find this plan.
Questions we currently have:
How do we present the results back to the project team?
How do we present the results to the participants?
Do we ask participants specific questions in order to get structured feedback?
Will we get text that is useful enough to go to the next step?
Which images provide good text and which don’t?
How might they results be used on the Archives Hub to help with discovery?
As we progress the work, we will start to think about organising a workshop for participants to get their feedback on the ML outputs.
Thanks to Adrian Stevenson, one of the Hub Labs team, who took me through the technical processes outlined in this post.
Given that Hub Labs is looking at AI and we’re also writing blog posts, it seemed like a good opportunity to try out one of the AI blog post writing applications that keep appearing in my various streams. The one that seems to get the best recommendations and is also free to trial is Writesonic. I’ve been meaning to write a post about our initial steps to build the AWS cloud infrastructure for Labs, so I thought I would use this as the basis for the experiment. You essentially provide a title, in this case I chose ‘Getting Started with Machine Learning in the Cloud’, and it will then provide some intros and outlines that you can accept or reject before it goes on to produce some copy. The best copy I could get it to produce given the options provided is pasted in below. You can see it’s a bit marketing speak and it was no great surprise that it’s not proved to be of much use for my experiment at least. For some reason it wanted to assume we are using Google Cloud and didn’t offer any AWS copy. It was interesting to try it though, and I could see it might be useful for more general posts.
The AI generated copy (italised by me) including the image is pasted below:
Getting Started with Machine Learning in the Cloud
Image Source: Freepik
The practice of machine learning involves training computer algorithms to recognize patterns in data and make predictions based on those patterns. It takes a lot of resources to implement machine learning in your business. There are many different datasets, tools, and software packages that you need to understand and choose from before you canput them to work analyzing your data and building your algorithm. This blog post discusses how you can use cloud services for machine learning with an example using Google Cloud Machine Learning (Cloud ML) and its APIs for Natural Language Processing (NLP).
What is Machine Learning?
There are many definitions of machine learning, including the IEEE’s “A computer program is said to learn from experience E with respect to some class of tasks T and some performance measure P if its performance at tasks in T, as measured by P, improves with experience E.” Or, as silly as it sounds, you can think about machine learning as a computer program that takes a bunch of data and spits out a prediction. You’re the one that has to tell the program how to do it.
Cloud ML with its APIs for NLP
Google Cloud Machine Learning workflows let you build and manage machine learning models in Google Cloud. Cloud ML supports a variety of machine learning frameworks, including TensorFlow, Caffe 2, and Apache MXNet. Cloud ML also provides a set of Natural Language APIs to support natural language processing (NLP) tasks, including sentiment analysis and entity extraction. Cloud ML’s NLP APIs can be accessed programmatically through its REST APIs.
Create a dataset using BigQuery
When you’re collecting data for machine learning, it’s important that your data be as accurate as possible. Google’s BigQuery lets you create a dataset in the cloud, and then you can use Google’s Natural Language API to analyze that dataset and get a breakdown of the different parts of speech. BigQuery is Google’s fully managed, enterprise data warehouse for business analytics. With BigQuery, you can query large data sets using standard SQL and get results in seconds. It’s also fully integrated with Google’s other analytical and collaboration tools, such as Google Sheets, Google Data Studio, and Google Cloud Platform.
Create a model using TensorFlow
With your dataset collected and a model for the Natural Language API built, you can create the model with TensorFlow. In this example, you’ll be building a model for sentiment analysis, which means you’ll be using the API’s sentiment analysis functionality. If you’re new to TensorFlow, don’t worry. There are some libraries out there that make it super easy to get started. One of them is tf-python-client, which is a library that allows you to write python code that uses TensorFlow.
Run the model in a Cloud Dataflow job
At this point, you’ve built a model and have everything ready to go. All that’s left to do is plug everything into Cloud Dataflow. Cloud Dataflow is Google’s fully managed, distributed framework for executing streaming data pipelines. Cloud Dataflow lets you build scalable, easy-to-manage data processing pipelines that transform and enrich your data. With Dataflow, you can build data processing and machine learning pipelines in a single, unified programming model. This means that you can build a model that performs data analysis, trains the model, and then deploys it in one workflow instead of three different ones.
Run the model in an API server
Your model is ready to go. Now you’re just missing the code to plug into an API server that can handle your data and send it to BigQuery. The Google Cloud Natural Language API lets you understand the meaning of text. You can use the API to detect entities, sentiments, and other linguistic properties of text. The API supports 82 languages.
Now that you know what machine learning is and have gotten started with Cloud ML and its NLP APIs, you can put this knowledge to use. You can use the NLP APIs to perform sentiment analysis or entity extraction on a dataset. This can be helpful for creating articles, social media posts, or marketing materials that are more effective.
With Cloud ML and its NLP APIs, you can create machine learning models that are more accurate and make more informed predictions. It takes a lot more effort to create models without cloud services, and it’s often difficult to keep track of all of the resources needed. With cloud services, you can focus on the model and not worry about the infrastructure and resources needed to run it.
A recent OCLC paper by Thomas Padilla highlights the need for ‘Pilot collaborations between institutions with representative collections’ and working ‘to share source data and produce “gold standard” training data.‘
However, it is also surely true that cultural heritage will need to engage with the broader AI and ML communities to understand and benefit fully from the range of ML services such as translation, transcription, object identification and facial recognition:
‘Advances in all of these areas are being driven and guided by the government or commercial sectors, which are infinitely better funded than cultural memory; for example, many nation-states and major corporations are intensively interested in facial recognition. The key strategy for the cultural memory sector will be to exploit these advantages, adapting and tuning the technologies around the margins for its own needs.’ From a short blog post by Dr Clifford Lynch from the CNI which is well worth reading.
People often criticise Machine Learning for being biased. But bias and mis-representation is essentially due to embedded bias in the input training data. The algorithm learns with what it has. So one of the key tasks for us as an archives community is to think about training data. We need algorithms that are trained to work for us to give us useful outputs.
Gathering training data in order to create useful models is going to be a challenge. Machine Learning is not like anything else that we have done before – we don’t actually know what we’ll get – we just know that we need to give the algorithm data that educates it in the way that we want. A bit like a child in school, we can teach it the curriculum, but we don’t know if it will pass the exam.
It certainly seems a given that we will need to use well labelled archival material as training data, so that the model is tailored specifically to the material we have. We will need to work together to provide this scale of training data. We have many wonderfully catalogued collections, with detail down to item level; as well as many collections that are catalogued quite basically, maybe just at collection level. If we join together as a community and utilise the well-catalogued content to train algorithms, we may be able to achieve something really useful to help make all collections more discoverable.
If an algorithm is trained on a fairly narrow set of data, then it is questionable whether it will have broad applicability. For example, if we train an algorithm on letters written in the 18th century, but just authored by two or three people, then it is unlikely to learn enough to be of real use with transcription; but if we train it on the handwriting of fifty people or more, then it could be a really useful tool for recognising and transcribing 18th century letters To do this training, we will need to bring content together. We will need to share the Machine Learning journey. The benefits could be massive in terms of discoverability of archives; effective discovery for all those materials that we currently don’t have time to catalogue. The main danger is that the resulting identification, transcription, tagging or whatever, is not to the standard that we want. We can only experiment and see what happens if we trial ML with a set of data (which is what we are doing now with our Labs project). One benefit could actually be much more consistency across collections. As someone working on aggregating data from 350 organisations, I can testify that we are not consistent! – and this lack of consistency impairs discovery.
Archival content is likely to be distinct in terms of both quality and subject. Typescripts might be old and faded, manuscripts might be hard to read, photographs might be black and white and not as high resolution as modern prints. Photographs might be of historical artefacts that are not recognised by most algorithms. We have specific challenges with our material, and we need the algorithms to learn from our material, in order to then provide something useful as we input more content.
In terms of subject, the Lotus and Delta shoe shops are a good example of a specific topic. They are represented in the Joseph Emberton papers, at the University of Brighton Design Archives, with a series of photographs. Architecture is potentially an interesting area to focus on. ML could give us some outputs that provide information on architectural features. It could be that the design of Lotus and Delta shops can be connected to other shops with similar architectures and shop fronts. ML may pick out features that a cataloguer may not include. On the other hand, we may find that it is extremely hard to train an algorithm on old black and white and potentially low resolution photographs in order for it to learn what a shop is, and maybe what a shoe shop is.
In this collection a number of the photographs are of exteriors. Some are identified by location, and some are not yet identified.
These photographs have been catalogued to item level, and so researchers will be able to find these when searching for ‘shops’ and particularly ‘shoe shops’ on the Hub, e.g. a search for ‘harrogate shoe shop‘ finds the exterior of a shop front in Harrogate. There may not be much more that could be provided for searching this collection, unless machine learning could label the type of shop front, the type of windows and signage for example. This seems very challenging with these old photographs, but presumably not impossible. With ML it is a matter of trying things out. You might think that if artificial intelligence can master self-driving cars it can master shop exteriors….but it is not a foregone conclusion.
If the model was trained with this set of photographs, then other shop fronts could potentially be identified in photographs that aren’t catalogued individually. We could potentially end up with collections from many different archives tagged with ‘shop front’ and potentially with ‘shoes’. Whether an unidentified shop front could be be identified is less certain, unless there are definite contextual features to work with.
Shop interiors are likely to be even more of a challenge. But it will be exciting to try things like this out and see what we get.
Commercial providers offer black box solutions, and we can be sure they were not trained to work well with archives. They may be adapted to new situations, but it is unlikely they can ever work effectively for archival content. I explored this to an extent in my last blog post. However, it is worth considering that a model not trained on archival material may highlight objects or topics that we would not think of including in a catalogue entry.
The Archives Hub and Jisc could play a pivotal role in co-ordinating work to create better models for archival material. Aggregation allows for providing more training material, and thus creating more effective models.
‘To date, most ML projects in libraries have required bespoke data annotation to create sufficient training data. Reproducing this work for every ML project, however, risks wasting both time and labor, and there are ample opportunities for scholars to share and build upon each other’s work.’ (R. Cordell, LC Labs report)
We can have a role to play in ‘data gathering, sharing, annotation, ethics monitoring, and record-keeping processes‘ (Eun Seo Jo, Timnit Gebru, https://arxiv.org/abs/1912.10389). We will need to think about how to bring our contributors into the loop in order to check and feedback on the ML outputs. This is a non-trivial part of the process that we are considering at the moment. We need an interface that displays the results of our ML trials.
One of the interesting aspects of this is that collections that have been catalogued in detail will provide the training data for collections that are not. Will this prove to be a barrier, or will it bring us together as a community? In theory the resources that some archives have, which have enabled them to catalogue to item level, can benefit those with minimal resources. Would this be a free and open exchange, or would we start to see a commercial framework developing?
It is also important that we don’t ignore the catalogue entries from our 350 contributors. Catalogues could provide great fodder for ML – we could start to establish connections and commonalities and increase the utility of the catalogues considerably.
The issue of how to incorporate the results of ML into the end user discovery interface is yet another challenge. Is it fundamentally important that end users know what has been done through ML and what has been done by a human? I can’t help thinking that over time the lines will blur, as we become more comfortable with AI….or as AI simply becomes more integrated into our world. It is clear that many people don’t realise how much Artificial Intelligence sits behind so many systems and processes that we use on an everyday basis. But I think that for the time being, it would be useful to make that distinction within our end user interfaces, so that people know why something has been catalogued or described in a certain way and so that we can assess the effectiveness of the ML contribution.
In subsequent posts we aim to share some initial findings from doing work at scale. We will only be able to undertake some modest experiments, but we hope that we are contributing to the start of what will be a very big adventure for archives.
Machine Learning is a sub-set of Artificial Intelligence (AI). You might like to look at devopedia.org for a short introduction to Machine Learning (ML).
Machine Learning is a data-oriented technique that enables computers to learn from experience. Human experience comes from our interaction with the environment. For computers, experience is indirect. It’s based on data collected from the world, data about the world.
Definition of Machine Learning from devopedia.org
The idea of this and subsequent blog posts is to look at machine learning from a specifically archival point of view as well as update you on our Labs project, Images and Machine Learning. We hope that our blog posts help archivists and other information professionals within the archival or cultural heritage domain to better understand ML and how it might be used.
At the Archives Hub we are particularly focussed on looking at Machine Learning from the point of view of archival catalogues and digital content, to aid discoverability, and potentially to identify patterns and bias in cataloguing.
Machine Learning to aid discoverability can be carried out as supervised or unsupervised learning. Supervised learning may be the most reliable, producing the best results. It requires a set of data that contains both the inputs and the desired outputs. By ‘outputs’ we mean that the objective is provided by labelling some of the input data. This is often called training data. In a ‘traditional’ scenario, code is written to take input and create output; in machine learning, input and output is provided, and the part done by human code is instead done by machine algorithms to create a model. This model is then used to derive outputs from further inputs.
So, for example, taking the Vickers instruments collection from the Borthwick: https://dlib.york.ac.uk/yodl/app/collection/detail?id=york%3a796319&ref=browse. You may want to recognise optical instruments, for example, telescopes and microscopes. You could provide training data with a set of labelled images (output data) to create a model. You could then input additional images and see if the optical instruments are identified by the model.
Of course, the Borthwick may have catalogued these photographs already (in fact, they have been catalogued), so we know which are telescopes and which are micrometers or lenses or eye pieces. If you have a specialist collection, essentially focused on a subject, and the photographs are already labelled, then there may be less scope for improving discoverability for that collection by using machine learning. If the Borthwick had only catalogued a few boxes of photographs, they might consider using machine learning to label the remaining photographs. However, a big advantage is that the enhanced telescope recognising model can now be used on all the images from the Archives Hub to discover and label images containing telescopes from other collections. This is one of the great advantages of applying ML across the aggregated data of the Archives Hub. The results of machine learning are always going to be better with more training data, so ideally you would provide a large collection of labelled photographs in order to teach the algorithm. Archive collections may not always be at the kind of scale where this process is optimised. Providing good training data is potentially a very substantial task, and does require that the content is labelled. It is possible to use models that are already available without doing this training step, but the results are likely to be far less useful.
Another scenario that could lend itself to ML is a more varied collection, such as Borthwick’s University photograph collection. These have been catalogued, but there is potential to recognise various additional elements within the photographs.
The above photograph has been labelled as a construction site. ML could recognise that there are people in the photograph, and this information could be added, so a researcher could then look for construction site with people. Recognising people in a photograph is something that many ML tools are able to do, having already been trained on this. However, archive collections are often composed of historic documents and old photographs that may not be as clear as modern documents. In addition, the models will probably have been trained with more current content. This is likely to be an issue for archives generally. For models to be effective, they need to have been trained with content that is similar to the content we want to catalogue.
The benefits of adding labels to photographs via ML to potentially enhance the catalogue and help with discoverability is going to depend upon a number of factors: how well the image is already catalogued, whether training data can be provided to improve the algorithm, how well ML can then pick out features that might be of use.
The drawings of fossil fish at the Geological Society are another example of a very subject specific collection. We put a few of these through some out-of-the-box ML tools. These tools have been pre-trained on large diverse datasets, but we have not done any additional training ourselves yet, so you could see them as generalists in recognising entities rather than specialists with any particular material or topic.
In this case the drawing has been tagged with ‘fossil’, which could be useful if you wanted to identify fossil drawings from a varied collection of drawings. It has also tagged this with archaeology and art, both of which could potentially be useful, again depending upon the context. The label of soil is a bit more problematic, and yet it is the one that has been added with 99.5% certainty. However, a bit of training to tell the algorithm that ‘soil’ is not correct may remove this tag from subsequent drawings.
This example illustrates the above point that a subject specific collection may be tagged with labels that are already provided in the catalogue description. It also shows that machine learning is unlikely to ever be perfectly accurate (although there are many claims it outperforms humans in a number of areas). It is very likely to add labels that are not correct. Ideally we would train the model to make less mistakes – though it is unlikely that all mistakes will be eliminated – so that does mean some level of manual review.
Tagging an image using ML may draw out features that would not necessarily be added to the catalogue – maybe they are not relevant to the repository’s main theme, and in the end, it is too time-consuming for cataloguers themselves to describe each photo in great detail as part of the cataloguing process.
The above image is a simple one with not too much going on. It will be discoverable on the Queen’s website through a search for ‘china’ or ‘robert hart’ for example, but tagging could make it discoverable for those interested in plants or architectural features. Again, false positives could be a problem, so a key here is to think about levels of certainty and how to manage expectations.
As mentioned above, archival images are often difficult to interpret. They may be old and faded, and they may also represent features or items that an algorithm will not recognise.
In the above example from Brighton Design Archives, the photograph is from a set made of an exhibition of 1947, Things In Their Home Setting. The AWS image Rekognition service has no problem with the chair, but it has confidently identified the oven as a refrigerator. This could probably be corrected by providing more training data, or giving feedback to improve the understanding of the algorithm and its knowledge of 1940’s kitchen furniture. But by the time you have given enough training data for the model to recognise a cooker from a fridge from a washing machine, it might have been easier simply to do the cataloguing manually.
Another option for machine learning is optical character recognition. This has been around for a while, but it has improved substantially as a result of the machine learning approach. Again, one of the challenges for archives is that many items within the collections are handwritten, faded, and generally not easily readable. So, can ML prove to be better with these items than previous OCR approaches?
A tool like Transkribus can potentially offer great benefits to archives, and is seen as a community-driven effort to create, gather and share training data. We hope to try out some experiments with it in the course of our project.
The above plan is from Lambeth Palace Library’s 19th century ecclesiastical maps. It can already be found searching for ‘clerkenwell’ or ‘st james parish’. But ML could potentially provide more searchable information.
The words here are fairly clear, so the character recognition using the Microsoft Azure ML service is quite good. Obviously the formatting is an issue in terms of word order. ‘James’ is recognised as ‘Iames’ due to the style of writing. ‘Church’ is recognised despite the style looking like ‘Chvrch’ – this will be something the algorithm has learnt. This analysis could potentially be useful to add to the catalogue because an end user could then search for ‘pentonville chapel’ or ‘northampton square’ and find this plan.
As well as looking at digital archives, we will be trying out examples with catalogue text. A great deal of archival cataloguing is legacy data, and archivists do not always have the time to catalogue to item level or to add index terms, which can substantially aid discoverability. So, it is tempting to look at ML as a means to substantially improve our catalogues. For example, to add to our index terms, which provide structured access points for end users searching for people, organisations, places and subjects.
In a traditional approach to adding subject terms to a catalogue, you might write rules. We have done this in our Names Project – we have written a whole load of rules in order to identify name, life dates, and additional data within index terms. We could have written even more rules – for example, to try to identify forename and surname. But it would be very difficult because the data does not present the elements of names consistently. We could potentially train an ML model with a load of names, tagging the parts of the name as forename, surname, dates, titles, epithets. But could an algorithm then successfully work out the parts of any subsequent names that we feed into it? It seems unlikely because there is no real consistency in how cataloguers input names. The algorithm might learn, for example, that a word, then a comma, then another word is surname, forename (Roberts, Elizabeth). But two words followed by a comma and another word could be surname + forename or forename + surname, (Vaughan Williams, Ralph; Gerald Finzi, composer). In this scenario, the best option may be to aim to use source data (e.g. the Virtual International Authority File) to compare our data to, rather than try to train a machine to learn patterns, when there really isn’t a model to provide the input.
We may find that analysing text within a catalogue offers more promise.
Here is an example from an administrative history of the British Linen Group, a collection held by Lloyds Banking Group. The entity recognition is pretty good – people’s names, organisations, dates, places, occupations and other entities can be picked out fairly successfully from catalogues. Of course that is only the first step; it is how to then use that information that is the main issue. You would not necessarily want to apply the terms as index terms for example, as they may not be what the collection is substantially about. But from the above example you could easily imagine tagging all the place names with a ‘place’ tag, so that a place search could find them. So, a general search for Stranraer would obviously find this catalogue entry, but if you could identify it as a place name it could be included in the more specific place name search.
With machine learning it is very difficult and sometimes impossible to understand exactly what is happening and why. By definition, the machine learns and modifies its output. Whilst you can provide training data to give inputs and desired outputs, machine learning will always be just that….a machine learning as it goes along, and not simply working through a programme that a human has written. Supervised learning provides for the most control over the outputs. Unsupervised learning, and deep learning, are where you have much less control (we’ll come onto those in later posts).
It is only by understanding the algorithms and what they are doing that you can set up your environment for the best results. But that is where things can get very complicated. We are going to try to run some experiments where we do prepare the data, but learning how to do this is a non-trivial task. Hence one of the questions we are asking is ‘is Machine Learning worth the effort required in order to improve archival discoverability?’ We hope to get at least some way along the road to answering that question.
There are, of course, other pressing questions, not least the issue of bias, and concerns about energy use with machine learning as well as how to preserve the processes and outputs of ML and document the decision making. But there could be big wins in terms of saving time that can then be dedicated to other tasks. The increasing volumes of data that we have to process may make this a necessity. We hope to touch upon some of these areas, but this is a fairly small scale project and Machine Learning it is one huge topic.
Under our new Labs umbrella, we have started a new project, ‘Images and Machine Learning’ it has three distinct and related strands.
We will be working on these themes with ten participants, who already contribute to the Archives Hub, and who have expressed an interest in one or more of these strands: Cardiff University, Bangor University, Brighton Design Archives at the University of Brighton, Queens University Belfast, the University of Hull, the Borthwick Institute for Archives at the University of York, the Geological Society, the Paul Mellon Centre, Lambeth Palace (Church of England) and Lloyds Bank.
This project is not about pre-selecting participants or content that meet any kind of criteria. The point is to work with a whole variety of descriptions and images, and not in any sense to ‘cherry pick’ descriptions or images in order to make our lives easier. We want a realistic sense of what is required to implement digital storage and IIIF display, and we want to see how machine learning tools work with a range of content. Some of the participants will be able to dedicate more time to the project, others will have very little time, some will have technical experience, others won’t. A successful implementation that runs beyond our project and into service will need to fit in with our contributors needs and limitations. It is problematic to run a project that asks for unrealistic amounts of time from people that will not be achievable long-term, as trying to turn a project into a service is not likely to work.
Over the years we have been asked a number of times about hosting content for our contributors. Whilst there are already options available for hosting, there are issues of cost, technical support, fit for purpose-ness, trust and security for archives that are not necessarily easily met.
Jisc can potentially provide a digital object store that is relatively inexpensive, integrated with the current Archives Hub tools and interfaces, and designed specifically to meet our own contributors’ requirements. In order to explore this proposal, we are going to invest some resource into modifying our current administrative interface, the CIIM, to enable the ingest of digital content.
We spent some time looking at the feasibility of integrating an archival digital object store with the current Jisc Preservation Service. However, for various reasons this did not prove to be a practical solution. One of the main issues is the particular nature of archives as hierarchical multi-level collections. Archival metadata has its own particular requirements. The CIIM is already set up to work with EAD descriptions and by using the CIIM we have full control over the metadata so that we can design it to meet the needs of archives. It also allows us to more easily think about enabling IIIF (see below).
The idea is that contributors use the CIIM to upload content and attach metadata. They can then organise and search their content, and publish it, in order to give it web address URIs that can be added to their archival descriptions – both in the Archives Hub and elsewhere.
It should be noted that this store is not designed to be a preservation solution. As said, Jisc already provides this service, and there are many other services available. This is a store for access and use, and for providing IIIF enabled content.
The metadata fields have not yet been finalised, but we have a working proposal and some thoughts about each field.
mandatory? individual vs batch?
preferably structured, options for approx. and not dated.
possibly a URI. option to add institution’s rights statement.
controlled list. values to be determined with participants. could upload a thesaurus. could try ML to identify type.
enable digital objects to be grouped e.g by topic or e.g. ‘to do’ to indicate work is required
unpublished/published. May refer to IIIF enabled.
unique URI of image (at individual level)
Proposed fields for the Digital Object Store
We need to think about the workflow and user interface. The images would be uploaded and not published by default, so that they would only be available to the DAO Store user at that point. On publication, they would be available at a designated URL. Would we then give the option to re-size? Would we set a maximum size? How would this fit in with IIIF and the preference for images of a higher resolution? We will certainly need to think about how to handle low resolution images.
International Image Interoperability Framework
IIIF is a framework that enables images to be viewed in any IIIF viewer. Typically, they can be sequenced, such as for a book, and they are zoomable to a very high resolution. At the heart of IIIF is the principle that organisations expose images over the web in a way that allows researchers to use images from anywhere, using any platform that speaks IIIF. This means a researcher can group images for their own research purposes, and very easily compare them. IIIF promotes the idea of fully open digital content, and works best with high resolution images.
There are very good reasons for the Archives Hub to get involved in IIIF, but there are challenges being an aggregator that individual institutions don’t face, or at least not to the same degree. We won’t know what digital content we will receive, so we have to think about how to work with images of varying resolutions. Our contributors will have different preferences for the interface and functionality. On the plus side, we are a large and established service, with technical expertise and good relationships with our contributors. We can potentially help smaller and less well-resourced institutions into this world. In addition, we are well positioned to establish a community of use, to share experiences and challenges.
One thing that we are very convinced by: IIIF is a really effective way to surface digital content and it is an enormous boon to researchers. So, it makes total sense for us to move into this area. With this in mind, Jisc has become a member of the IIIF Consortium, and we aim to take advantage of the knowledge and experience within the community – and to contribute to it.
This is a huge area, and it can feel rather daunting. It is also very complicated, and we are under no illusions that it will be a long road, probably with plenty of blind alleys. It is very exciting, but not without big challenges.
It seems as if ML is getting a bad reputation lately, with the idea that algorithms make decisions that are often unfair or unjust, or that are clearly biased. But the main issue lies with the data. ML is about machines learning from data, and if the data is inadequate, biased, or suspect in some way, then the outcomes are not likely to be good. ML offers us a big opportunity to analyse our data. It can help us surface bias and problematic cataloguing.
We want to take the descriptions and images that our participants provide and see what we can do with ML tools. Obviously we won’t do anything that affects the data without consulting with our contributors. But it is best with ML to have a large amount of data, and so this is an area where an aggregator has an advantage.
This area is truly exploratory. We are not aiming for anything other than the broad idea of improved discoverability. We will see if ML can help identify entities, such as people, places and concepts. But we are also open to looking at the results of ML and thinking about how we might benefit from them. We may conclude that ML only has limited use for us – at least, as it stands now. But it is changing all the time, and becoming more sophisticated. It is something that will only grow and become more embedded within cultural heritage.
Over the next several months we will be blogging about the project, and we would be very pleased to receive feedback and thoughts. We will also be holding some webinar sessions. These will be advertised to contributors via our contributors list, and advertised on the JiscMail archives-nra list.