09:00 - 10:30  PROJECT OVERVIEWS & Methods
09:00 - 09:20 Introduction to Global Contentious Politics (GLOCON) Project
09:20 - 09:40 Roger Bock and Elizabeth Boschee, Combining Machine Learning and Linguistic Intuition for Automatic Coding of Protest Events
09:40 - 10:00 Parang Saraf and Naren Ramakrishnan, A Semi-Automated Approach for Extracting Political Events from News Articles with High Recall Across Multiple Languages
10:00 - 10:30 Discussion - Source Selection
1. What are the respective challenges of working with global or local news sources in automated information extraction?
2. How should news articles in local sources that are about countries other than the focus country should be treated?[1]
3. How can we address possible problems of editorial bias and/or censorship in source selection?
10:30 - 11:00 Coffee break
11:00 - 12:30 DATA CHALLENGE
11:00 - 11:20 Ali HŸrriyetoğlu - Data challenge overview: results of Ali Safaya
11:20 - 11:40 Data challenge result 1: Anais Ollagnier and Hywel Williams, Using Neuronal Network Embeddings for Classification and Event Identification
11:40 - 12:00 Data challenge result 2: Hristo Tanev, Vanni Zavarella, Martin Atkinson and Jakub Piskorski, THE NEXUS EVENT DETECTION SYSTEM AT COPE 2019
12:00 - 12:30 Discussion - Information Extraction
1. How should texts containing information about multiple real world protest events be handled in event information extraction?
2. How can multiple real world events in single new articles be separated?
3. What are the challenges of event co-referencing?
4. What are the benefits of event deduplication?
5. Is event deduplication really worth it?
6. If deduplication is necessary, how can this be achieved cost-efficiently with minimum data loss?
12:30 - 14:00 Lunch
14:00 - 15:30 EVALUATION
14:00 - 14:20 Scott Althaus, Buddy Peyton and Dan Shalmon, Validating Contentious Political Event Data with Aggregate and Granular Comparisons: Methods and Application to NigeriaÕs Conflict with Boko Haram
14:20 - 14:40 Fuyuki Kurasawa, Analyzing Online Contention and Controversy: The Promises and Challenges of Deep Learning
14:40 - 15:30 DISCUSSION - Generalizability
1. How does variation in source language effect recall and precision performance of automated tools?
2. How would automated translation of source material effect classification and extraction performance?
3. How can automated translation of source material be optimized in terms of costs and performance?
4. What are the challenges of coming up with event ontologies that would accommodate variations in contentious politics over time and across country contexts?
5. What are the challenges of dealing with various editorial styles and/or web content formats (i.e. in tokenization) in text extraction?
6. What kind of automated text processing tools would best suit unsupervised domain adaptation?
15:30 - 16:00 Coffee break
16:00 - 16:30 Discussion - Ethics
1. What ethical concerns would automated information extraction from protest event news raise?
2. What are the ethical responsibilities of researchers working on automated protest event data extraction towards social groups and individuals engaging in protest?
3. How could we address and resolve possible ethical issue raised by automated information extraction from protest event news?
16:30 - 17:30 Discussion - Prospects
How can we organize publication of the workshop proceedings? Some options are CEUR, a special issue in a journal, etc.
What issue can we address together in order to advance state-of-the-art for automated protest event information collection?
What kind of collaborative effort could be organized? Some options are: a) creating a dataset together, e.g. annotating future vs. other events, creating a gold standard corpus of contentious events database info, sharing language or computing resources
Are there any venues we can write proposals for funding or resources?

[1]
ben bu soruyu pek anlamadõm? ne demek isteniyor burada?
-‚ağrõ Yoltar
Hindistan'dan haber işaretlenirken, pakistan'dan haber õkarsa, onu işaretlemiyoruz. irrelevant country deyip geiyoruz. Toollarõmõz bunu daha tam beceremiyor. insanlar bu konuda ne yapmõş?
-Ali HŸrriyetoğlu