A reference document used to keep track of the reading group meetings. Pull requests are welcome. Meeting Link
| Date | Presenter | Paper | Why should we care | Appear in | Repo | Blogs |
|---|---|---|---|---|---|---|
| 22/01/24 | Marian-Andrei Rizoiu | Causally estimating the effect of YouTube's recommender system using counterfactual bots | One of the first studies to estimate the effects of the YouTube recommender system on radicalisation using counterfactual viewing series, using bots. | arXiv | ||
| 05/02/24 | Rohit Ram | CA3 Dry Run | ||||
| 19/02/24 | Elaine Gong | E-BERT: Efficient-Yet-Effective Entity Embeddings for BERT | It introduces E-BERT, which aligns Wikipedia2Vec entity vectors with BERT's wordpiece vector space, enabling the use of entity vectors as if they were wordpiece vectors. | Association for Computational Linguistics 2020 | ||
| 04/03/24 | Pio Calderon | Inference for Mark-Censored Temporal Point Processes | The paper introduces a marginalization technique to performance inference at prediction time for a trained marked temporal point process when presented with mark censoring. | UAI 2023 | ||
| 18/03/24 | Frankie Yuan | Causality Guided Disentanglement for Cross-Platform Hate Speech Detection | The paper introduces a hate speech detection model that uses a VAE to extract causal embeddings separated into platform independent features useful for predicting presence of hate, and platform dependent features useful for predicting the target of hate. | WSDM 2024 | ||
| 01/04/24 | – | No reading due to public holiday | ||||
| 15/04/24 | Matthew Ghannoum | Auto-RPL Early Research Feedback | ||||
| 29/04/24 | – | Canceled | ||||
| 15/05/24 | Pio Calderon | ICWSM24 Lightning Talk Dry Run | ||||
| 29/05/24 | – | Research showcase dry run | ||||
| 12/06/24 | Lin Tian | Adaptable Transfer Learning Models for Online Misinformation Detection | Summary of Lin’s PhD work. | |||
| 26/06/24 | Elaine Gong | Demystifying embedding spaces using large language models | This paper introduces a novel framework to interpret and interact embedding vectors by leveraging the power of LLMs. | ICLR 2024 | ||
| 10/07/24 | Emily Booth | SoK: A Framework for Unifying At-Risk User Research | Paper linked during the Google Trust and Safety Awards information session and strongly suggested everyone read. They really liked how it understood ‘at risk’ users, and also the human and technical considerations it made. | IEEE Symposium on Security and Privacy’22 | ||
| 24/07/24 | Marian-Andrei Rizoiu | Dry run of the CMT II Keynote presentation. | The dry-run will happen in the UTS Data Arena facility – a 360° data visualisation facility designed for collaborative experiences. | CMT Information Integrity Forum | ||
| 07/08/24 | Frankie Yuan | Temporal dynamics of coordinated online behavior: Stability, archetypes, and influence | This paper examines temporal change in online coordinated communities and introduces temporal multiplex networks as a means of dynamically identifying communities engaging in coordinated behaviour leading up to elections in the UK and the US. | PNAS | ||
| 21/08/24 | Pio Calderon | ECML-PKDD24 Talk Dry Run | ||||
| 04/09/24 | Lin Tian | From Word Models to World Models: Translating from Natural Language to the Probabilistic Language of Thought | This paper takes you to the world model. How does language inform our downstream thinking? In particular how do humans make meaning from language–and how can we leverage a theory of linguistic meaning to build machines that think in more human-like ways? In this paper they propose rational meaning construction a computational framework for language-informed thinking that combines neural language models with probabilistic models for rational inference. | [NeurIPS'23] (https://proceedings.neurips.cc/paper_files/paper/2023/file/f9f54762cbb4fe4dbffdd4f792c31221-Paper-Conference.pdf) NeurIPS'23 | ||
| 18/09/24 | Canceled | |||||
| 02/10/24 | Canceled | |||||
| 16/10/24 | Matthew Ghannoum | Final Presentation | ||||
| 30/10/24 | Canceled | |||||
| 13/11/24 | Canceled | |||||
| 27/11/24 | Marian-Andrei Rizoiu | Unmasking the Web of Deceit: Uncovering Coordinated Activity to Expose Information Operations on Twitter | This paper delves into the detection of key players driving IOs by employing similarity graphs constructed from behavioral pattern data. | WWW’24 |