21 June 2016
BL Labs are pleased to announce the two finalists of the BL Labs Competition (2016)!
The judging panel, consisting of the BL Labs Advisory Board and members of the British Library's Digital Scholarship team, definitely had their work cut out for them as the Competition had so many strong entries this year.
After much deliberation, BL Labs is excited to introduce the BL Labs Competition (2016) finalists and their two projects:
Black Abolitionist Performances and their Presence in Britain
Hannah-Rose Murray, PhD student at the University of Nottingham
Hannah-Rose describes her winning project for us:
My project will create an original and exciting window into Victorian society by analysing the African American presence in Britain and how their performances and lectures reached nearly every corner of the country. It asks, how can we uncover hidden black voices in the archive?
I have searched through the British Library’s online Newspaper Database to collect as much data as possible referring to formerly enslaved African American Frederick Douglass’ lectures in Britain and, for the first time, I have collated this information to provide a systematic and detailed analysis of his experiences. I have created a map showing some of Douglass’ lectures (image below with the red pins) and a second map listing lectures given by other black abolitionists (image below with multi-coloured pins). This is displayed on my website, www.frederickdouglassinbritain.com
Map: locations of Frederick Douglass’ lectures in the United Kingdom and Ireland
Searching the online newspaper database can only provide partial information, and there are thousands of lectures just waiting to be uncovered, and then plotted onto a map. But how can we achieve this? I want to build on the work of the previous BL Labs Competition winner, Katrina Navickas’ Political Meetings Mapper, and develop a process that searches for the individual rather than a record of the lecture in question. This will allow us to analyse the impact of black abolitionists on British society and how far they travelled to lecture against
American slavery. Mapping their movements has never been done before and we can use this visual representation to gather an estimate of how many lectures they gave and, most importantly, allow their hidden voices to be heard. We can explore their performances through their own words and follow how they interacted with British audiences to win support for abolition and combat the deeply entrenched racism of the period.
Hannah-Rose Murray is a second year PhD student with the Department of American and Canadian Studies, University of Nottingham. Her AHRC/M3C-funded PhD focuses on the legacy of formerly enslaved African Americans on British society and the different ways they fought British racism. Hannah-Rose received a first class Masters degree in Public History from Royal Holloway University and has a BA History degree from University College London (UCL). In Nottingham, Hannah-Rose works closely with the Centre for Research in Race and Rights and is one of the postgraduate directors of the Rights and Justice Research Priority Area, which includes the largest number of scholars (700) in the world working on rights and justice.
SherlockNet: Using Convolutional Neural Networks to automatically tag and caption the British Library Flickr collection
Luda Zhao, Brian Do and Karen Wang, students from Stanford University, California
The SherlockNet team tell us about their winning project:
Book illustrations, such as those in the British Library’s Flickr Commons 1 million collection, provide valuable insights into the cultural fabric of their time. However, large image collections are only useful and discoverable by researchers if they can be deeply explored, and the process can often be time-consuming, requires expertise and relies on humans to recognise patterns. Thus, a computationally guided approach providing automatic pattern recognition would allow art historians to study them in a more unbiased and efficient manner.
Machine learning can extract information and insights from data on a massive scale. In particular, deep learning methods such as convolutional neural networks (CNNs), inspired by biological neural networks in the brain, are particularly suited for understanding the content of images, due to the hierarchical nature of the “neurons” that make up the network. CNNs are composed of layers of neurons, where each neuron is a “mini-calculator” of sorts that takes in information from the layer above and outputs more information. Progressing deeper into the network means that neurons will be working on increasingly complex information, progressing from detect features like lines or curves in the shallower layers to detect objects like cars and planes in the deeper layers. The CNN could then combine these “high-level” features to make very accurate predictions on the contents of the image.
Team photo: SherlockNet presenting their initial work at Stanford University, California
In this project, we plan to develop and optimize CNN algorithms to accomplish two important tasks: tagging and captioning. In the first step, we will classify all images with general categorical tags (e.g. decorations, architecture, animals etc.). This will serve as the basis for us to develop new ways to facilitate rapid online tagging with user-defined sets of tags, including a tag suggestions interface that would continuously reward user input by doing on-the-fly online learning to improve model accuracy. In the second, we will build off recent results in Deep Learning research and automatically generate descriptive natural-language captions for all images (e.g. “A man in a meadow on a horse”). These natural-language captions will expand the richness of the images and provide more intuitive ways for researchers to discover and use these images.
We intend to make our tags and captions accessible and searchable by the public through a web-based interface. We also plan to provide supplementary visualisation tools to give researchers greater insight and access into the technical assumptions behind our work. Finally, we will use our text annotations to globally analyse trends in the British Library Flickr collection over time. Together, we hope our project will establish CNNs as a novel tool for image annotation and analysis, as well as encourage widespread adoption of neural networks in bibliology.
Luda Zhao is currently a Masters student studying Computer Science at Stanford University, living in Palo Alto, California. He is interested in using machine learning and data mining to tackle tough problems in a variety of real-life contexts, and he's excited to work with the British Library to make art more discoverable for people everywhere. In his spare time, Luda enjoys beaches, swimming, all-you-can-eat BBQs, reading, and collecting subway maps.
Brian Do grew up in sunny California and is a first-year MD/PhD student at Harvard Medical School. Previously he studied Computer Science and biology at Stanford. Brian loves using data visualisation and cutting edge tools to reveal unexpected things about sports, finance and even his own text message history. In his free time he enjoys cooking, running in the hot sun, and playing frisbee on the beach.
Karen Wang is currently a senior studying Computer Science at Stanford University, California. She also has an Art Practice minor. Karen is very interested in the intersection of computer science and humanities research, so this project is near and dear to her heart! She will be continuing her studies next year at Stanford in CS, Artificial Intelligence track.
BL Labs and the finalists are looking forward to working together on these exciting projects and we will be posting regular updates as these works progress.