Ò»±¾µÀÎÞÂë

 

The “Disinformation Age" (Bennett and Livingston 2021) is a Frankenstein result of several deep and decades-old political currents (Brown 2019). The problem of disinformation brings together media, technology, and democracy. This is as much a political problem as it is a technological one, and the economics of media matter for any redress. While the digital affordances of social media platforms have offered unprecedented amplification and resonance of any messaging, social media activity is typically fueled by analog mass media (Benkler, Faris et al. 2018). While partisan and emotion-soaked messages blast out from organizations such as Sinclair Media and Fox, echoed in social-media platforms, fact-checked reporting and public service media also play an active role in fueling social-media activity. Public broadcasters have the strongest brands commanding public trust in the U.S.: NPR and PBS. Public broadcasters also showcase the work of some of the most trusted media-makers: independent documentarians such as Kartemquin filmmakers (Aufderheide, 2024). In this moment, can public broadcasting be a resource for democratic public-making and civil society?

Tutorials will be offered during the conference as parallel sessions. All attendees are able to attend tutorials.

Wednesday September 18th

9:30am-12:00pm - "How to evaluate your model" by Frank Ritter

“How to evaluate your model” introduces the basic concepts in evaluating a model. After debunking the concept of proving a model, this chapter presents the case that you would like to do two fundamental things: show that the model is worth taking seriously, both to yourself and to others, and to know where to improve it.I note non-numeric, numeric, and advanced methods that have been used, using a score card as a way to summarize the fit. It will also address interactions of these tasks with publishing your model. “How to publish your model” provides general comments on publishing reports of models and the steps in modeling and simulation. I note the importance of writing and of the final results. I provide a detailed process for handing the preparation, submission, and revision of a paper reporting a model, particularly about the importance of staying in touch with stakeholders.

1:30pm - 3:30pm - "The BotBuster Universe" by Lynnette Ng and Jeffrey Reminga

Thursday September 19th

10:00am-1:00pm "Defending Against Generative AI Threats in NLP" by Amrita Bhattacharjee

Generative AI and, in particular, Large Language Models (LLMs) have seen unprecedented advancements in the last few years. Given the ease of access of many state-of-the-art LLMs, these models have been heavily adopted and have entered workflows among professionals, academics and even enthusiastic lay users. With impressive performance on natural language and benchmarks, and even more complex tasks involving math and reasoning, LLMs are capable of being used as either generalist or topic-specific chatbots. While larger and more capable LLMs are being developed and used for a variety of high-impact use cases, these models are still susceptible to being misused and attacked by malicious entities. In this tutorial, we dive into the current state of LLM research and development, before exploring the types of threats and attacks that LLMs are susceptible to, and finally exploring the various defense methods that have been developed to tackle these threats, along with challenges and research directions that need attention from the community.

2:00pm-5:00pm "Introductory Tutorial on Agent-Based Modeling" by Charles M. Macal

Agent-based modeling (ABM) and simulation is an approach to modeling systems comprised of autonomous, interacting agents. The need for modeling complex and adaptive systems comprised of populations of natural (people, organizations, communities) and engineered (drones, robotic swarms) entities continues to drive the application of ABM in a variety of application areas, including those where simulation has not been extensively applied. Applications range from modeling agent behavior in supply chains, consumer goods markets, and financial markets, to predicting the spread of epidemics and providing insights on the factors responsible for the growth and fall of ancient civilizations. ABM is having far-reaching effects on the way that governments and businesses use computer models to support decision-making and how researchers use models as in silico electronic laboratories. Some contend that ABM “is a third way of doing science” and could augment traditional discovery methods for knowledge generation. This brief tutorial introduces agent-based modeling by describing key concepts of ABM, discussing some illustrative applications, and addressing toolkits and methods for developing agent-based models.