Current Schedule: Link
Date | Presenter | Paper | Why should we care | Appear in | Repo | Blogs |
---|---|---|---|---|---|---|
10/1/2022 | Pio Calderon | Correlated Cascades: Compete or cooperate | A model of behavior adoption in the presence of multiple cooperating or competing products | AAAI 2017 | 1 | |
17/01/22 | Marian-Andrei Rizoiu | End-of-history illusion | A series of psychological studies that show that humans are not able to predict how much they will change in the future. | Science 2013 | TED talk | |
24/01/22 | Dima Galat | Synthesizing Adversarial Negative Responses for Robust ResponseRanking and Evaluation | Context dependent garbage in - strong predictor out | ACL2021 | 1 | |
31/01/22 | Frankie Yuan | Challenges in Detoxifying Language Models | Side effects of trying to prevent certain language in language models | Findings of EMNLP 2021 | ||
7/2/2022 | Rohit Ram | Polarization and tipping points | An explanation for political polarization and why it might be difficult to reverse | PNAS | https://github.com/mongooma/polarizationTipping | |
14/02/22 | Quyu Kong | Learning Temporal Point Processes for Efficient Retrieval of Continuous Time Event Sequences | A model to retrieve and rank a relevant set of event sequences for a given query sequence | AAAI 2022 | 1 | |
21/02/22 | Callum Pastuszak | Multi-modal Identification of State-Sponsored Propaganda on Social Media | A model for detecting state sponsored propaganda on Twitter combining textual and visual analysis | ICPR 2020 | ||
28/02/22 | Rohit Ram | Polarized Ukraine 2014: opinion and territorial split demonstrated with the bounded confidence XY model parametrized by Twitter data | RSOS | demo | ||
7/3/2022 | Pio Calderon | Modeling Adoption and Usage of Competing Products | An extension of multivariate Hawkes to model the adoption and frequency of use of competing products | ICDM 2015 | ||
14/03/22 | Daniela Elia | Ontology-guided Job Market Demand Analysis: A Cross-Sectional Study for the Data Science field | An Ontology-based Information Extraction (OBIE) method for job and skill data | Semantics2017 | ||
21/03/22 | – | No group meeting; TRP intensive course for several students. | ||||
28/03/22 | Marian-Andrei Rizoiu | Misinformation in and about science | A sharp and sober overview of what can go wrong in science and publishing: from predatory journal to salami publishing and p-value hacking. | PNAS | tools for calling bullsh*t | |
4/4/2022 | Pio Calderon | CA1 dry run | ||||
11/4/2022 | Frankie Yuan | Racism is a virus: anti-asian hate and counterspeech in social media during the COVID-19 crisis | Labelling and analysis of interaction between hatespeech and counterspeech during the COVID-19 pandemic | ASONAM 2021 | repo | |
18/04/22 | – | No group meeting: public holiday (Easter Monday) | ||||
25/04/22 | – | No group meeting: public holiday (ANZAC day) | ||||
2/5/2022 | Amelie Girard | Fast Sparse Decision Tree Optimization via Reference Ensembles | Decision Trees are a family of interpretable models. When Models are used for high Stake decisions they need to be vetted properly | AAAI 2022 | 1 | |
9/5/2022 | Quyu Kong | CCGL: Contrastive Cascade Graph Learning | A novel framework for information cascade graph learning in a contrastive self-supervised and task-agnostic way. | TKDE 2022 | 1 | |
16/05/22 | – | No group meeting: 3MT presentations for compulsory coursework for half of the group. | ||||
23/05/22 | Callum Pastuszak | The many dimensions of truthfulness: Crowdsourcing misinformation assessments on a multidimensional scale | Comparing the effectiveness of crowdsourced fact-checking vs expert fact-checking | IP&M 2021 | ||
30/05/22 | Amelie Girard | Bias in data-driven artificial intelligence systems—An introductory survey | The goal of this survey is to provide a broad multidisciplinary overview of the area of bias in AI systems and to focus on technical challenges and solutions as well as to suggest new research directions towards approaches well-grounded in a legal frame | European Commission Grant/Award | ||
6/6/2022 | Marian-Andrei Rizoiu | Network constraints on worker mobility | The paper proposes an employment-weighted embeddedness measure that the authors show is correlated with wage premium and optimized over the worker’s career. | Nature Human Behavior | ||
13/06/22 | – | No group meeting: public holiday (Queen’s birthday) | ||||
20/06/22 | Daniela Elia | Using LinkedIn Endorsements to Reinforce an Ontology and Machine Learning-Based Recommender System to Improve Professional Skills | This paper presents a Lifelong Learning course recommendation system that incorporates LinkedIn endorsements | |||
27/06/22 | Frankie Yuan | Reinforcement Guided Multi-Task Learning Framework for Low-Resource Stereotype Detection | Interesting model framework which trains an RL agent to select training instances for transfer learning. | ACL 2022 | ||
4/7/2022 | – | No group meeting (wild weather) | ||||
11/7/2022 | – | BDS social event: Gaddie pitching and project matching & VR room | ||||
18/07/22 | Rohit Ram | TIMME: Twitter Ideology-detection via Multi-task Multi-relational Embedding | A method for recovering political ideology based on different relations graphs based on a GCN approach. | KDD20 | https://github.com/PatriciaXiao/TIMME | |
25/07/22 | Elaine Gong | The building blocks of economic complexity | ||||
1/8/2022 | Pio Calderon | Information Interaction Profile of Choice Adoption | A method to estimate the interaction profile between entities given sequences of contagions and exposures | ECML PKDD 2021 | https://github.com/GaelPouxMedard/InterRate | |
8/8/2022 | Guest lecture: Swapnil Mishra | Title: Machine Learning; Bayesian Inference and Point Processes for Public Good: Understanding and quantifying the uncertainty | Bio: Assistant Professor at University of Copenhagen (UCPH) working at intersection of public health; machine learning and Bayeasian modelling. Before that he was a Post-doc at School of Public Health @ Imperial College London | Machine Learning & Global Health Network (mlgh.net) | ||
15/08/22 | Daniela Elia | Practical Skills Demand Forecasting via Representation Learning of Temporal Dynamics | Forecasting skills using deep learning for time series analysis | AIES 2022 | ||
22/08/22 | Guest lecture: Philipp Schneider | Title: Cross-Excitation and Influence Pathways in Networks: Propagation of COVID-19 and Public Intervention in Switzerland | Bio: PhD student at the chair of Operations, Economics and Strategy at EPFL in Switzerland. Philipp’s current work focuses on leveraging data, optimization, and machine learning, to solve practical problems that matter to society. | |||
29/08/22 | Frankie Yuan | Computational Analysis of Bot Activity in the Asia-Pacific: A Comparative Study of Four National Elections | Case study of bot activities in elections of 4 Asia-Pacific countries | ICWSM 2021 | ||
5/9/2022 | Rohit Ram | Predicting Opinion Dynamics via Sociologically-Informed Neural Networks | This paper presents an opinion dynamics model (i.e. predicting individuals opinions over time). They use a NN (which is socially informed) to predict opinions over time and show that regularisation by sociological opinion dynamics models converted to ODEs improves performance. | KDD22 | repo | |
12/9/2022 | Pio Calderon | On the Complexity of Opinions and Online Discussions | A method to (1) develop latent representations of online opinions using upvote/downvote data and (2) estimate the complexity (i.e. dimensionality) of the latent space. | KDD19 | repo | |
19/09/22 | Marian-Andrei Rizoiu | A causal test of the strength of weak ties | A new study of more than 20 million people, published in Science, shows that your close friends (on LinkedIn) are not your best bet: instead you should look to acquaintances you don’t know well enough to share a personal connection with | Science | The Conversation article | |
26/09/22 | Amelie Girard | Toward Personalized XAI: A Case Study in Intelligent Tutoring Systems | A rare case study aiming to deploy an XAI system | Journal of Artificial Intelligence | ||
3/10/2022 | – | No group meeting: public holiday (Labour Day) | ||||
10/10/2022 | Elaine Gong | Social capital I: measurement and associations with economic mobility | This paper introduces three types of social capital and explain how to measure them based on Facebook social network, after that they explore the correlation between social capital and economic mobility | [Nature] | Could find more analysis results in (https://opportunityinsights.org/) | |
17/10/22 | Daniela Elia | jobBERT: Understanding Job Titles through Skills | This paper presents a model to normalise job titles based on their skill content | FEAST, ECML-PKDD 2021 Workshop | ||
24/10/22 | Frankie Yuan | Unsupervised Domain Adaptation for Hate Speech Detection Using a Data Augmentation Approach | This paper presents some interesting tricks for domain adaptation in low resource hate speech datasets through data augmentation. They tag sentences from the target domain to generate contextual templates, which they then fill with key terms they identify from the source domain. | ICWSM 2022 | ||
31/10/22 | Amelie Girard | [Case study: Applied XAI to bio-informatics – Instance-based learning and belief rule-based XAI approach] | ||||
7/11/2022 | Rohit Ram | Fast Few shot Self-attentive Semi-supervised Political Inclination Prediction | This paper uses a fairly sophisticated few shot architechture for detecting political inclincation and achieves SOTA performance | |||
14/11/22 | Frankie Yuan | CA1 Dry Run | ||||
21/11/22 | Rohit Ram | DHSS Dry Run | ||||
28/11/22 | Pio Calderon | Modeling Collective Anticipation and Response on Wikipedia | A model of the dynamics around peaks of online popularity that can be used to model collective anticipation and response to planned events. | ICWSM 2021 | ||
5/12/22 | Elaine Gong | Learning interpretable word embeddings via bidirectional alignment of dimensions with semantic concepts | This paper propose a method called BiImp that makes dimensions of word embedding interpretable without making compromises on the semantic task performance | |||
12/12/2022 | Daniela Elia | Skill requirements in job advertisements: A comparison of skill-categorization methods based on wage regressions | This paper compares different methods to extract skill demand from job posting and their performance in explaining wage variation |