9th International Workshop on Mining Actionable Insights from Social Networks
Special Edition on
Deep Learning Methods for Social Media Mining
Invited Talks
Title: Towards XAI on Knowledge Graphs
(Recorded Video)
Abstract. Deep learning based models have been effectively applied to tackle various problems in many disciplines. Yet, their predictions are often at most post-hoc and locally explainable. In contrast, predicted class expressions in description logics are ante-hoc and globally explainable. However, such models do not yet scale to real-world knowledge graphs with billions of triples. In this talk, we firstly present recent deep learning based models and their applications in large knowledge graphs. Thereafter, we discuss recent progress towards explainable AI on knowledge graphs. Particularly, we focus on the intersection of description logic concept learning and deep learning on knowledge graphs.
Bio. Caglar Demir is a researcher at the Data Science Group of Paderborn University, Germany. Caglar studied Computer Engineering in Istanbul and Computer Science in Paderborn. His PhD thesis (submitted in March 2023) was on learning continuous representations for knowledge graphs. His research interest centers around designing scalable ML algorithms on graph-structured data. Caglar has published various research papers in top-tier conferences including ICML, IJCAI, WWW, ISWC, ESWC. He is a co-founder of Tentris Start-up that works on providing the fastest-backend for RDF knowledge graphs..
Ferdinando Fioretto
Title: The Unintended Societal Effects of Privacy in decision and learning tasks
Abstract. Differential Privacy has become the go-to approach for protecting sensitive information in data releases and learning tasks that are used for critical decision processes. For example, census data is used to allocate funds and distribute benefits, while several corporations use machine learning systems for criminal assessments, hiring decisions, and more. While this privacy notion provides strong guarantees, we will show that it may also induce biases and fairness issues in downstream decision processes. These issues may adversely affect many individuals’ health, well-being, and sense of belonging, and are currently poorly understood. In this talk, we delve into the intersection of privacy, fairness, and decision processes, with a focus on understanding and addressing these fairness issues. We first provide an overview of Differential Privacy and its applications in data release and learning tasks. Next, we examine the societal impacts of privacy through a fairness lens and present a framework to illustrate what aspects of the private algorithms and/or data may be responsible for exacerbating unfairness. We hence show how to extend this framework to assess the disparate impacts arising in Machine Learning tasks. Finally, we propose a path to partially mitigate these fairness issues and discuss grand challenges that require further exploration.
Bio. Ferdinando Fioretto is an assistant professor at Syracuse University. He works at the juncture of Machine Learning, optimization, privacy, and ethics focusing on two themes: (1) Responsible AI: it analyzes the equity of AI systems in support of decision-making and learning tasks and designs algorithms that better align with societal values and (2) ML for Science and Engineering: it develops the foundation to blend deep learning with mathematical optimization to enable the integration of knowledge, constraints, and physical principles into learning models. He is a recipient of the 2022 NSF CAREER award, the 2022 Amazon Research Award, the 2022 Google Research Scholar Award, the 2022 Caspar Bowden PET award, the 2021 ISSNAF Mario Gerla Young Investigator Award, the 2021 ACP Early Career Researcher Award, the 2017 AI*AI Best AI dissertation award, and several best paper awards. He is also actively involved in the organization of several events, including the Privacy-Preserving Artificial Intelligence workshop at AAAI, the Algorithmic Fairness through the lens of Causality and Privacy at NeurIPS, and the Optimization and Learning in multiagent systems workshop at AAMAS.
Ruizhe Li
Title: Shaping Dialogue: Exploring Text Style Transfer and Empathetic AI
Abstract. Human-computer interaction is rapidly evolving, and two fields at the forefront of this evolution are text style transfer and empathetic dialogue response generation. In this talk, we'll dive deep into these burgeoning research areas, examining their potential to reshape our interactions with AI systems. Text style transfer, the technology that enables the modification of a text's style while maintaining its original content, has fascinating implications. It broadens horizons for personalization, sentiment modification, and data augmentation. We'll survey the latest breakthroughs in this field, spotlighting innovative techniques and models that have significantly improved style transfer accuracy. Our exploration then shifts to empathetic dialogue response generation. As AI systems become integral to our lives, there's a growing need for them to understand and respond to human emotions effectively. By incorporating empathy into these systems, we can foster more engaging and supportive interactions. We'll discuss recent advancements in training AI models to generate responses that align with the user's emotional state, providing not just understanding, but also appropriate emotional responses. Throughout the talk, we will summarize recent related works in both fields, offering a comprehensive overview of the current research landscape and the challenges and opportunities it presents.
Bio. Ruizhe Li is an assistant professor in the Department of Computing Science at the University of Aberdeen. He was a postdoc research fellow in the Web Intelligence Group, affiliated with the Center for Artificial Intelligence at the University College London (UCL). He received his PhD from the Department of Computer Science at the University of Sheffield. His research interests focus on dialogue systems, multi-modality NLP, knowledge graphs, and deep latent variable models. He has published several papers at top-tier conferences, e.g., ACL, EMNLP, IJCAI, and ICML.