Program

Thursday 9 November

• 18h30: opening exhibition
• 19h & 20h30: guided tour of the exhibition

Friday 10 November : Lectures

Mike Kestemont (UA) on Generative Models and the Digital Humanities: Towards Synthetic Literature

Mike Kestemont is assistant professor in the department of literature at the University of Antwerp in Belgium. He is a researcher in computational text analysis, in particular for historic texts. Authorship attribution is one of his main areas of expertise. He designs computational algorithms which can automatically identify the authors of anonymous texts through the quantitative analysis of individual writing styles. For the yearly event ‘The Netherlands Read!’ he co-designed Asibot, a writing tool trained with Recurrent Neural Networks, based on +4000 Dutch novels. The tool was used by Dutch novelist Ronald Giphart to write an extra fiction story for the re-edition of Asimov’s ‘I Robot’ in the beginning of November. Mike will present recent advances in Machine Learning - and its changing cultural status - with an emphasis on generative models, i.e. models that synthetizes new artificial data, instead of mere modelling pre-existing data. This will include a survey of some ongoing ethical discussions in the world of AI.

Amir Sarabadani on Wikipedia’s ORES-project

Software engineer Amir Sarabadani will present the ORES-project. “The Objective Revision Evaluation Service” is a web service and API that provides machine learning as a service for Wikimedia projects maintained by the Scoring Platform team. The system is designed to help automate critical wiki-work - for example: vandalism detection and removal. Currently, the two general types of scores that ORES generates are in the context of ‘edit quality’ and ‘article quality’. Amir has been active for Wikipedia since 2006, as sysop, bureaucrat and check user for the Persian Wikipedia, and developer for Wikimedia projects. He is operator of Dexbot and one of the developers of the pywikibot framework, and works as a software engineer in Wikimedia Germany. Amir was born in 1992 in Tehran, Iran and studied physics. He currently lives in Berlin, Germany.

Saturday 11 November

In the framework of Algoliterary Encounters Nicolas Malevé proposes a workshop on computer vision. Language, words, writing, descriptions and formulations are intimately linked to the way the millions of images on the internet are organised. Over the years, algorithmic techniques have evolved creating a new articulation of the relations between vision, information and knowledge. The recent breed of algorithms that power computer vision make heavy use of the techniques of machine learning. As other algorithms, machine learning algorithms need to be programmed, but they also need to be trained. Contemporary artificial intelligence aims to “teach” machines the cognitive abilities of humans.

But how do computer scientists understand human vision and how do they translate it in a concept they can work with? They are interested in a very specific aspect of human vision: the glimpse, the glance, the moment in perception that allow to take immediate decisions, a near reflex perception.

Nicolas introduces a method to assign relationships between images and words, as described in « What do we perceive in a glance of a real-world scene? » (Fei Fei et al. 2007). By proposing a variation on this method, he shifts the focus of the experiment, which is not so much to collect quantitative data from the participants but to discuss with the participants what is at stake in the experiment and how it models vision.

Participation is limited to 15 persons.

Sunday 12 November

‘We Are A Sentiment Thermometer’ is one of the installations in the exhibition of Algoliterary Encounters. It will be the starting point for the workshop.

It asks questions to commonly used language models based on machine learning, like GloVe and word2vec. Using part of the Internet as training data, these models are considered to learn from ‘collective intelligence’.

‘We Are A Sentiment Thermometer’ scores given written sentences as more positive or negative. This technique, also called sentiment analysis, is widely used to measure the success of marketing and other campaigns. The method is based on a script by software engineer Rob Speer, who shows the racist bias built into it, once the model starts judging culture specific sentences.

‘We Are A Sentiment Thermometer’ is presented as a case to raise questions of very different sorts. In this workshop members of Algolit will introduce the script and extend on the different steps and components that are used. Depending on the interests and skills, the script can be used to ask other questions to this algorithmically constructed unconsciousness. Or it can be modified in different ways, to offer better understanding and alternative points of view.

With: Manetta Berends, Cristina Cochior, Gijs De Heij, Hans Lammerant, An Mertens