Keynote Speakers
Emine Yilmaz, Amazon, UCL (UK)
Fact Checking Using Stance Detection and User Comments
Abstract:Misinformation regarding products could be a major problem that could have a significant effect on the trust and satisfaction of the users of e-commerce sites. Such misinformation could be present in two major forms: 1) Claims made about a product in product descriptions, and 2) Reviews posted about a product. In this talk, I will be talking about the methods we have developed for automatic detection of misinformation. I will first focus on the problem of stance detection, where the task is to categorize an overall position of a resource towards a claim into different classes such as agree, disagree, unrelated, etc. One of the major problems faced by current machine learning models used for stance detection is caused by a severe class imbalance among these classes. Hence, most models fail to correctly classify instances that fall into minority classes. In the first part of this talk, I will present a model that addresses this problem by proposing a hierarchical representation of these classes and show how such a model could achieve significant performance improvement especially in the classification of minority classes. In addition to stance detection, people’s comments or reviews are also quite informative regarding the truthfulness of the claim made. In the second part of this talk, I will present a model that uses information from people's replies to a claim that can be used to predict the truthfulness of the claims made, together with its uncertainty.Speaker Bio
Emine Yilmaz is a Professor and Turing Fellow at University College London (UCL), Department of Computer Science, as well as an Amazon Scholar at Amazon Alexa Shopping. Between 2012 and 2019, she also worked as a research consultant for Microsoft Research Cambridge, where she used to work as a full-time researcher prior to joining UCL. Her research until now has received several awards including a Bloomberg Data Science Research Award in 2018, the Karen Sparck Jones Award in 2015 and the Google Faculty Research Award in 2014. Emine's current research interests include automatic misinformation detection, fairness and bias in machine learning algorithms and conversational assistants. She has published research papers extensively at major venues such as ACM SIGIR, CIKM and WSDM, gave several tutorials as part of top conferences, and organised various workshops. She has served in various roles such as the Best Paper Awards Chair for SIGIR 2021, Panels Chair for The Web Conf 2021, PC Chair for ECIR 2019, and PC Chair for ACM SIGIR 2018.
Julia Lasserre, Zalando (Germany)
Fashion AI and Algorithmic size advice
Abstract: Fashion is an enormous industry and offers incredible challenges. As a result it is rapidly developing as an interest for the scientific community, particularly in the field of computer vision. Fashion is intimate, it touches our personality and our body, and it can't be avoided :) In the first part of the talk, we will discuss Fashion AI and what challenges lie ahead of us at Zalando. In the second part, we will talk about our algorithmic size advice: what data we use and what kind of products we build to support our customers in buying the right size. If time allows, we will have a deep-dive into one algorithm.Speaker Bio
Julia Lasserre has always been interested in the possibilities offered by machine learning and in applying various techniques to real-world problems. After receiving her PhD in machine learning and computer vision from Cambridge Uiversity, she did a post-doc in bioinformatics at the Max-Planck-Institute for Molecular Genetics, and then joined the industry. At Zalando, she has worked on different topics ranging from brand understanding to product tagging, active learning, representation learning and recommendations. She is currently an Principal Applied Scientist for the Size & Fit team and focuses on personalized size recommendations.
