Current Schedule: Link
| Date | Presenter | Paper | Why should we care | Appear in | Repo | Blogs |
|---|---|---|---|---|---|---|
| 9/01/2023 | Frankie Yuan | Training language models to follow instructions with human feedback and How would Stance Detection Techniques Evolve after the Launch of ChatGPT? | InstructGPT (precursor model to ChatGPT) and the use of ChatGPT and large language models for explainable stance detection | InstructGPT ChatGPT | ||
| 16/01/2023 | Amelie Girard | NLP hands-on series | In this hands-on session Abubakar Abid showcases tools and resources for machine and deep learning on the Hugging Face Hub. Practical use cases include searching for loading and deploying models as well as datasets for various machine learning tasks. | |||
| 23/01/2023 | Rohit Ram | Meta Pseudo Labels | This is a development in the learning to learn paradigm and is a good way to situate at DL paradigms generally | CVPR | https://github.com/google-research/google-research/tree/master/meta_pseudo_labels | |
| 30/01/2023 | Guest lecture: Juliette Unwin | Title: Developing methods for infectious disease modelling | Bio: Imperial College Research Fellow (ICRF) based in the MRC centre for Global Infectious Disease Analysis in the School of Public Health at Imperial College, London. I am interested in applying and developing novel methods for outbreak analysis to help inform policy makers in real time. I am currently part of the Imperial College COVID-19 response team looking at real time modelling of Rt across Europe and the USA. | |||
| 6/02/2023 | Pio Calderon | Identifying Competition and Mutualism between Online Groups | Application of population ecology and community ecology approaches to identify competition and mutualism between online groups | ICWSM 2022 | ||
| 13/02/2023 | Daniela Elia | Trend-Aware Tensor Factorization for Job Skill Demand Analysis | Given a job position, how to identify the right job skill demand and its evolving trend becomes critically important for both job seekers and employers in the fast-paced job market. | IJCAI'19 | ||
| 20/02/2023 | Elaine Gong | Topic Modeling in Embedding Spaces | This paper proposes a method called Topic Embedding Model (TEM) which uses a neural network to learn topic embeddings | TACL 2020 | ||
| 27/02/2023 | Jooyoung Lee | Whose Advantage? Measuring Attention Dynamics across YouTube and Twitter on Controversial Topics | Whether an ideologicalgroup garners more attention across platforms and/or topics and how the attention dynamics evolve over time have notbeen explored. In this work we present a quantitative studythat links collective attention across two social platforms –YouTube and Twitter centered on online activities surround-ing popular videos of three controversial political topics in-cluding Abortion Gun control and Black Lives Matter over 16 months. | ICWSM 2022 | ||
| 6/03/2023 | Rohit Ram | Addressing Annotation Complexity: The Case of Annotating Ideological Perspective in Egyptian Social Media | Showcases the difficulty of annotating ideology online. | ACL'16 | ||
| 13/03/2023 | Frankie Yuan | Reconsidering Tweets:Intervening during Tweet Creation Decreases Offensive Content | Case study on effectiveness of interventions at posting time in preventing abusive and offensive posting on Twitter. | ICWSM 2022 | Blog | |
| 20/03/2023 | Amelie Girard | Engaging Gentrification as a Social Justice Issue in HCI | The article proposes that gentrification is a social justice issue and argues that HCI needs to engage with it. It suggests six research areas for HCI scholars to counter gentrification and discusses how contemporary socio-technical systems mediate the consumption side dynamics of gentrification. | ACM 2019 | ||
| 27/03/2023 | Rohit Ram | Individuals with depression express more distorted thinking on social media | This article looks at the prevalence of congnitive distortions in depressed people on Twitter using a dictionary approach. | |||
| 3/04/2023 | Marian-Andrei Rizoiu | Interval-censored Transformer Hawkes: Detecting Information Operations using the Reaction of Social Systems | WWW’23 presentation dry run. | WWW’23 | ||
| 10/04/2023 | – | No reading due to public holiday | ||||
| 17/04/2023 | Pio Calderon | Tweedie-Hawkes Processes: Interpreting the Phenomena of Outbreaks | This article introduces the Tweedie-Hawkes process to model the influence of event-level features on outbreaks | AAAI 2020 | ||
| 24/04/2023 | – | Reading Cancelled | ||||
| 1/05/2023 | Daniela Elia | Occupational profiling driven by online job advertisements: Taking the data analysis and processing engineering technicians as an example | This paper uses Convolutional Neural Networks for Sentence Classification (TextCNN) to extract named entities from job postings and profile occupations | PLOS ONE | https://doi.org/10.1371/journal.pone.0253308 | |
| 8/05/2023 | – | Cancelled due to family emergency | ||||
| 15/05/2023 | Elaine Gong | Topic Discovery via Latent Space Clustering of Pretrained Language Model Representations | This paper introduces a framework that combines pre-trained language models (PLMs) with latent space learning and clustering for topic discovery in text corpora. The approach outperforms traditional topic models generating more coherent and diverse topics while providing improved document representations. | Proceedings of the ACM Web Conference 2022 | ||
| 22/05/2023 | Jooyoung Lee | Reinforcement Learning-based Counter-Misinformation Response Generation: A case Study of COVID-19 Vaccine Misinformation | This paper presents RL based counter-misinformation response generation model with desirable properties. | |||
| 29/05/2023 | Elaine Gong | ISPIM Conference deck - Linking Leading R&D Firms and Emerging Technologies | This paper mainly leverages pre-trained model to explore the latent knowledge of emerging technologies | |||
| 5/06/2023 | Amelie Girard | Causal Machine Learning: A Survey and Open Problems | The paper discusses Causal Machine Learning (CausalML) which is a set of machine learning methods that use a structural causal model (SCM) to formalize the data-generation process. This approach enables the analysis of the effects of interventions and counterfactuals. The paper categorizes the methods into five groups: (1)causal supervised learning (2)causal generative modeling (3)causal explanations (4)causal fairness and (5)causal reinforcement learning. The authors provide a systematic comparison of the methods in each category and discuss open problems. The paper also reviews applications in computer vision natural language processing and graph representation learning. The authors conclude with an overview of causal benchmarks and a critical discussion of the state of the field along with recommendations for future work. | |||
| 12/06/2023 | – | No reading due to public holiday | ||||
| 19/06/2023 | Pio Calderon | ABC Learning of Hawkes Processes with Missing or Noisy Event Times | This work introduces a likelihood-free approach based on Approximate Bayesian Computation (ABC) to fit the Hawkes process to distorted data (ex. missing events noisily observed events). | |||
| 26/06/2023 | – | Cancelled | ||||
| 3/07/2023 | Marian-Andrei Rizoiu | Digital Services Act: Estimate the Effectiveness of Moderating Harmful Online Content | This paper explores the likely effectiveness of online social media moderation by deploying point process modelling. | PNAS 2023 | GitHub repo | |
| 10/07/2023 | Daniela Elia | Exploring the UK Cyber Skills Gap through a mapping of active job listings to the Cyber Security Body of Knowledge (CyBOK) | This paper presents a methodology to link job descriptions to specific knowledge areas for cybersecurity roles in the UK | |||
| 17/07/2023 | Frankie Yuan | Finding Qs: Profiling QAnon Supporters on Parler | This paper examines identifying and profiling Qanon supporters on Parler, a relatively understudied far-right social media platform. | ICWSM 2023 | ||
| 31/07/2023 | Jooyoung Lee | Misleading Repurposing on Twitter | This paper defines misleading repurposing and presents evidences of such repurposed accounts in social media. | |||
| 7/08/2023 | Frankie Yuan | Happenstance: Utilizing Semantic Search to Track Russian State Media Narratives about the Russo-Ukrainian War On Reddit | The paper looks narratives pushed by Russian state media during the start of the Russo-Ukrainian war. They propose a sentence level embedding pipeline for topic extraction. They further examine the spread of the identified narratives on reddit using semantic search. | ICWSM 2023 | ||
| 14/08/2023 | Amelie Girard | |||||
| 21/08/2023 | Pio Calderon | Popularity Prediction for Social Media over Arbitrary Time Horizons | This paper introduces a Hawkes-based model to predict popularity of an online item at any arbitrary time horizon in the future using a combination of static features (content and user) and observed popularity growth | VLDB 2022 | ||
| 28/08/2023 | – | Cancelled | ||||
| 4/09/2023 | Elaine Gong | Dynamic ContextualizedWord Embeddings | The paper introduces a novel method of representing words as functions of both linguistic and extralinguistic context, based on a pretrained language model, and demonstrates its potential applications for various NLP tasks that involve semantic variability | ACL Anthology 2021 | ||
| 11/09/2023 | – | Cancelled | ||||
| 18/09/2023 | Jooyoung Lee | The Morbid Realities of Social Media: An Investigation into the Narratives Shared by the Deceased Victims of COVID-19 | The paper characterizes the dominant themes and sources present in the victim’s posts along with identifying the role of the platform in handling deadly narratives. | |||
| 25/09/2023 | Elaine Gong | CA1 Presentation | ||||
| 2/10/2023 | – | No reading due to public holiday | ||||
| 9/10/2023 | Pio Calderon | Predicting the popularity of tweets using internal and external knowledge: an empirical Bayes type approach | The paper introduces an empirical Bayes approach to enhance the accuracy of tweet popularity prediction, combining historical retweet data with external knowledge to improve forecasting. | AStA (Advances in Statistical Analysis) | ||
| 16/10/2023 | Rohit Ram | Empirically Measuring Online Social Influence | Social influence is linked to opinion formation and is known to modulate political and geostrategic processes. Worryingly, influence mechanisms are also used to sow mistrust in democratic institutions, health organisations or the effectiveness of life-saving vaccines. | Psyber Conference Dry-run | ||
| 23/10/2023 | Frankie Yuan | Non-polar Opposites: Analyzing the Relationship between Echo Chambers and Hostile Intergroup Interactions on Reddit | Echo chambers in online social media communities have been a hot topic as of late. This paper investigates the correlation between users’ engagement with echo-chambers and their toxicity when posting in outside communities. | ICWSM 2023 | ||
| 30/10/2023 | – | No reading (bi-weekly) | ||||
| 6/11/2023 | Pio Calderon | The Geometry of Misinformation: Embedding Twitter Networks of Users WhoSpread Fake News in Geometrical Opinion Spaces | This paper studies fake news sharers by leveraging multidimensional ideological embeddings, representing stances/attitudes towards multiple issues, not just the traditional left-right polarization. | ICWSM 2023 | ||
| 13/11/2023 | – | No reading (bi-weekly) | ||||
| 20/11/2023 | Matthew Ghannoum | Fast and Accurate Network Embeddings via Very Sparse Random Projection | FastRP is a node embedding algorithm that uses Random Projections. Not only does FastRP achieve comparable performance to existing methods such as DeepWalk and Node2Vec, but in this paper it is shown to be over 4000 times faster. | Proceedings of the 28th ACM International Conference on Information and Knowledge Management | ||
| 27/11/2023 | – | No reading (bi-weekly) | ||||
| 4/12/2023 | Matthew Ghannoum & Frankie Yuan | Honours Project Presentation for BGI dry run / AT3 Research Presentation + Frankie ASQPS dry run | Will attach my paper/slides on presentation day. | |||
| 11/12/2023 | – | No reading (bi-weekly) | ||||
| 18/12/2023 | Marian-Andrei Rizoiu | |||||
| 25/12/2023 | – | No reading due to public holiday |