@UniversityLeeds Eric Atwell is discussing resources for those interested in Arabic corpus linguistics. One of his points of research is corpus libguistics using the Quran. #WIDH2018
Standard corpus linguistics and machine learning tools: sketchengine.eu #WIDH2018
There are tools out there DH researchers can use rather than building from scratch. #WIDH2018
Part of his research has been a chat bot that has been trained on text from the Quran. #WIDH2018
Alternative natural language processing methods and representations for capturing the semantics of texts (zoom in quite a bit) #WIDH2018 : 
“Classical Arabic texts, in particular the Quran, are special cases for machine learning...[as] there has already been a lot of groundwork done on interpreting the meanings of the words within the corpora.” #WIDH2018
Another resource: Wekr #WIDH2018
Multiple projects coming out of his lab. One studies non-Arabic pronunciation, a few study Quran ontology, one is the afore-mentioned Quran as a training corpus for an ML chatbot, another studies the hadīth and is developing a part of speech tagger for Arabic. #WIDH2018
Another looks at the Chinese translation of the Quran #WIDH2018
Modern NLP methods derived from modern Arabic (trained on Twitter, for example) can be adapted to Classical Arabic. #WIDH2018
Multiple lessons from this colloquium that Prof. Atwell will be taking back to his research group: identifying text reuse is a key issue #WIDH2018
Is DH really humanities? If not, maybe it is AI research? The participants here are developing gold standard datasets for ML training and evaluation. #WIDH2018
Ideas for the future: Islamicate Digital Humanities can provide challenges for AI #WIDH2018
Run a contest to solve your ML problems. alt.qcri.org/semeval2019/in… #WIDH2018
Present your work to Corpus Linguists: WACL Workshop in Arabic Corpus Linguistics, CL ‘2019 Cardiff 22-26 Jul 2019. #WIDH2018
Apologies for any typos in this or previous threads! #WIDH2018
From discussion following Eric Atwell’s keynote: a need for precision about what the cut-off for classic Arabic is. Arabic from the Quran is pre-Islamic Arabic; there is the possibility of two different forms of Arabic (pre-Islamic and classical) being conflated #WIDH2018
There are concerns about how meaning is decided within the training data, which is why this is closely tied to AI and ethics debate. The job of humanists is to work with computer scientists to consider potential noise within the data. #WIDH2018
One such conversation that Prof. Atwell brought up having had with his students re. the above: what is the gender of angels? There is conflicting information between the language construction (feminine) and tradition (heavenly beings are only male). #WIDH2018
Comment above re. AI, ethics, and the humanities was by @sarahsavant1. #WIDH2018
