README.md

NLP Reading Group

The target audience is all the members of the NLP group and other possible interested participants.

The meeting takes place weekly during term for one hour usually on Mondays, 14:00-15:00 UK Time. See below a list of the upcoming and past meetings.

The meetings of the group are informal and no necessary preparation will be required with the exception of the moderator reading the current paper and the rest having at least a brief overview of it.

There is also a #readinggroup channel in the NLP group's unofficial Slack channel: https://usfd-nlp.slack.com/messages

A list of past meetings before 2018/19 can be found here.

---

Upcoming Meeting(s)

Will restart after summer break.

Past Meetings

• Mon 30 May 2022

  • Paper: Context-aware Decoder for Neural Machine Translation using a Target-side Document-Level Language Model)
  • Moderator: Sebastian Vincent

• Mon 16 May 2022

  • Paper: Word Order Does Matter (And Shuffled Language Models Know It)
  • Moderator: Edward Gow-Smith

• Mon 9 May 2022

  • Paper: Chain of Thought Prompting Elicits Reasoning in Large Language Models
  • Moderator: Jasivan A Sivakumar

• Mon 11th Apr 2022

  • Paper: Refocusing on Relevance: Personalization in NLG
  • Moderator: Tomas Goldsack

• Mon 4th Apr 2022

  • Paper: Are we there yet? Exploring clinical domain knowledge of BERT models
  • Moderator: Dylan RS Phelps

• Mon 7th Mar 2022

  • Paper: I Wish I Would Have Loved This One, But I Didn't -- A Multilingual Dataset for Counterfactual Detection in Product Reviews
  • Moderator: Mali Jin

• Mon 7th Mar 2022

  • Paper: Superbizarre Is Not Superb: Derivational Morphology Improves BERT's Interpretation of Complex Words
  • Moderator: Edward Gow-Smith

• Mon 28th Feb 2022

  • Paper: Text Classification Using Label Names Only: A Language Model Self-Training Approach
  • Moderator: Danae Sanchez Villegas

• Mon 21st Feb 2022

  • Paper: Finetuned Language Models Are Zero-Shot Learners
  • Moderator: Katerina Margatina

• Mon 14th Feb 2022

  • Paper: ProFormer: Towards On-Device LSH Projection Based Transformers
  • Moderator: Huiyin

• Mon 07th Feb 2022

  • Paper: Structurizing Misinformation Stories via Rationalizing Fact-Checks
  • Moderator: Cass Zhao

• Mon 31st Jan 2022

  • Paper: Data2vec: A General Framework for Self-supervised Learning in Speech, Vision and Language
  • Moderator: Xutan

• Mon 24th Jan 2022

  • Paper: A Survey on Dialogue Summarization: Recent Advances and New Frontiers
  • Moderator: Guanyu Huang

• Mon 06th Dec 2021

  • Paper:
  • Moderator: Zeerak

• Mon 29th Nov 2021

  • Paper:
  • Moderator: Dylan

• Mon 22nd Nov 2021

  • Paper: Structure Mapping in Analogy and Similarity (American Psychologist 1997)
  • Moderator: Xutan

• Mon 08th Nov 2021

  • Review Session for ACL Submissions

• Mon 15th Nov 2021

  • ACL Deadline

• Mon 18th Oct 2021

  • Paper: The R-U-A-Robot Dataset: Helping Avoid Chatbot Deception by Detecting User Questions About Human or Non-Human Identity (ACL'21)
  • Moderator: Yida

• Mon 11th Oct 2021

  • Paper: Debugging Tests for Model Explanations
  • Moderator: George

• Mon 04th Oct 2021

  • Paper: Predicting emergent linguistic compositions through time: Syntactic frame extension via multimodal chaining (EMNLP'21)
  • Moderator: Danae

• Mon 27th Sep 2021

  • Paper: Does Pretraining for Summarization Require Knowledge Transfer? (EMNLP'21 - Findings)
  • Moderator: Katerina

• Mon 20th Sep 2021

  • Paper: Paragraph-level Simplification of Medical Texts (NAACL'21)
  • Moderator: Fernando

• Mon 16th Aug 2021

  • Paper: From characters to words: the turning point of BPE merges (EACL'21)
  • Moderator: Xutan

• Mon 28th June 2021

  • Paper: Can a Fruit Fly Learn Word Embeddings? (ICLR'21)
  • Moderator: Yida M

• Mon 21st June 2021

  • Paper: Does syntax matter? A strong baseline for Aspect-based Sentiment Analysis with RoBERTa (NAACL'21)
  • Moderator: Mali

• Mon 14th June 2021

  • Paper: Learning The Difference That Makes A Difference With Counterfactually-Augmented Data (ICLR'20)
  • Moderator: Katerina

• Mon 7th June 2021

  • Paper: -
  • Moderator: -

• Mon 26th April 2021

  • Paper: Estimating predictive uncertainty for rumour verification models (ACL'20)
  • Moderator: Wenzhe

• Mon 22nd March 2021

  • Paper: Summarising Historical Text in Modern Languages (EACL'21)
  • Moderator: Xutan

• Mon 15th March 2021

  • Paper: LAMBDA NETWORKS: MODELING LONG-RANGE INTERACTIONS WITHOUT ATTENTION
  • Moderator: George

• Mon 8th March 2021

  • Paper: Grounded Conversation Generation as Guided Traverses in Commonsense Knowledge Graphs
  • Moderator: Ruizhe Li

• Mon 1st March 2021

  • Paper: CogLTX: Applying BERT to Long Texts (NeurIPS'20)
  • Moderator: Yida Mu

• Mon 22nd Feb 2021

  • Paper: Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting (AAAI'21 Best Paper)
  • Moderator: Xutan

• Mon 8th Feb 2021

  • Paper: Cross-Media Keyphrase Prediction: A Unified Framework with Multi-Modality Multi-Head Attention and Image Wordings
  • Moderator: Danae

• Fri 29th Jan 2021

  • ACL Review/Feedback session

• Mon 18th Jan 2021

  • Paper: Affective and Contextual Embedding for Sarcasm Detection
  • Moderator: Mali

• Mon 11th Jan 2021

  • Paper: Unsupervised Data Augmentation for Consistency Training
  • Moderator: Katerina

• Mon 4th Jan 2021

  • Paper: A Diagnostic Study of Explainability Techniques for Text Classification
  • Moderator: George

• Mon 16th Nov 2020

  • Paper: Heterogeneous Graph Neural Networks for Extractive Document SummarizationSlide
  • Moderator: Wenzhe

• Mon 9th Nov 2020

  • Paper: Self-Attention Guided Copy Mechanism for Abstractive Summarization
  • Moderator: Hardy

• Mon 2th Nov 2020

  • Paper: Where Are the Facts? Searching for Fact-checked Information to Alleviate the Spread of Fake News
  • Moderator: Yida Mu

• Mon 26th Oct 2020

  • Paper: Cold-start Active Learning through Self-supervised Language Modeling
  • Moderator: Katerina

• Mon 19th Oct 2020

  • Paper: What Was Written vs. Who Read It: News Media Profiling Using Text Analysis and Social Media Context
  • Moderator: Danae

• Mon 12th Oct 2020

  • Paper: Joint Modelling of Emotion and Abusive Language Detection
  • Moderator: Mali

• Mon 5th Oct 2020

  • EACL Paper Review Session

• Mon 28th Sep 2020

  • Paper: Towards Faithfully Interpretable NLP Systems
  • Moderator: George

• Mon 20th July 2020

  • Paper: Longformer: The Long-Document Transformer
  • Moderator: Xutan

• Mon 13th July 2020

  • Paper: Latent Space Factorisation and Manipulation via Matrix Subspace Projection
  • Moderator: Ruizhe

• Mon 29 June 2020

  • Paper: Context-aware monolingual repair for neural machine translation
  • Moderator: Sebastian

• Mon 22 June 2020

  • Paper: Relational inductive biases, deep learning, and graph networks
  • Moderator: Peter

• Mon 15 June 2020

  • Paper: Multimodal Quality Estimation for Machine Translation
  • Moderator: Shu Okabe / Fred

• Mon 8 June 2020

  • Paper: Analyzing Political Parody in Social Media
  • Moderator: Danae

• Fri 29 May 2020

  • EMNLP Review/Feedback session

• Mon 18 May 2020

  • Paper: Learning to Faithfully Rationalize by Construction
  • Moderator: George

• Mon 11 May 2020

  • Paper: Aspect Sentiment Classification Towards Question-Answering with Reinforced Bidirectional Attention Network.
  • Moderator: Mali
  • Room: Google Hangouts

• Mon 04 May 2020

  • Paper: BLEU might be Guilty but References are not Innocent
  • Moderator: Fernando
  • Room: Google Hangouts

• Mon 27 Apr 2020

  • Paper: “Trust me, I have a Ph.D.”: A Propensity Score Analysis on the Halo Effect ofDisclosing One’s Offline Social Status in Online Communities (ICWSM-20)
  • Moderator: Yida
  • Room: Google Hangouts

• Mon 20 Apr 2020

  • Paper: How to (Properly) Evaluate Cross-Lingual Word Embeddings: On Strong Baselines, Comparative Analyses, and Some Misconceptions (ACL'19)
  • Moderator: Xutan
  • Room: Google Hangouts

• Mon 6 Apr 2020

  • Paper: Many Faces of Feature Importance: Comparing Built-in and Post-hocFeature Importance in Text Classification
  • Moderator: George

• Mon 16 Mar 2020 Now Mon 30 Mar 2020 (due to COVID-19)

  • Paper: Lagging Inference Networks and Posterior Collapse in Variational Autoencoders
  • Moderator: Ruizhe
  • Room: G12-Blue

• Mon 9 Mar 2020

  • Paper: BatchBALD: Efficient and Diverse Batch Acquisition for Deep Bayesian Active Learning
  • Moderator: Katerina
  • Room: G12-Blue

• Mon 2 Mar 2020

  • Paper: Integrating Text and Image: Determining Multimodal Document Intent in Instagram Posts
  • Moderator: Danae
  • Room: G12-Blue

• Mon 24 Feb 2020

  • Paper: Interpretable emoji prediction via label-wise attention LSTMs
  • Moderator: George
  • Room: G12-Blue

• Mon 17 Feb 2020

  • Paper: Detecting Customer Complaint Escalation with Recurrent Neural Networks and Manually-Engineered Features
  • Moderator: Mali
  • Room: G12-Blue

• Mon 10 Feb 2020

  • Paper: Hubless Nearest Neighbor Search for Bilingual Lexicon Induction (ACL 19) (paper | code)
  • Moderator: Xutan
  • Room: G12-Blue

• Mon 3 Feb 2020

  • Paper: Affect-Driven Dialog Generation
  • Moderator: Danae
  • Room: G12-Blue

• Mon 27 Jan 2020

  • Paper: BERTScore: Evaluating Text Generation with BERT
  • Moderator: Fernando
  • Room: G12-Blue

• Mon 20 Jan 2020

  • Paper: Uncertainty-Aware Attention for Reliable Interpretation and Prediction
  • Moderator: George
  • Room: G12-Blue

• Fri 06 Dec 2019

  • **ACL Review/Feedback session
  • Room: G12-Blue

• Mon 02 Dec 2019

  • Paper:
  • Moderator: Cancel due to ACL deadline
  • Room: G12-Blue

• Mon 25 Nov 2019

  • Paper: What Does This Word Mean? Explaining Contextualized Embeddings with Natural Language Definition
  • Moderator: Varvara
  • Room: G12-Blue COM-108 - Ada Lovelace Room

• Mon 18 Nov 2019

  • Paper: Hierarchical Transfer Learning for Multi-label Text Classification
  • Moderator: Mali
  • Room: G12-Blue

• Mon 11 Nov 2019

  • Paper: Incorporating Priors with Feature Attribution on Text Classification (https://www.aclweb.org/anthology/P19-1631)
  • Moderator: -
  • Room: G12-Blue

• Mon 4 Nov 2019

  • Paper: DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter. NeurIPS workshop paper 2019 (paper|blog|slides)
  • Moderator: Katerina
  • Room: G12-Blue

• Mon 28 Oct 2019

  • Paper: A robust self-learning method for fully unsupervised cross-lingual mappings of word embeddings (link|video|slides)
  • Moderator: Xutan
  • Room: G12-Blue

• Mon 21 Oct 2019

  • Paper: Fine-Grained Analysis of Propaganda in News Articles (https://arxiv.org/pdf/1910.02517.pdf)
  • Moderator: Yida
  • Room: G12-Blue

• Mon 14 Oct 2019

  • Paper: MeanSum: A Neural Model for Unsupervised Multi-document Abstractive Summarization
  • Moderator: Hardy
  • Room: G12-Blue

• Mon 7 Oct 2019

  • Paper: We need to talk about standard splits (ACL 2019 - Outstanding Paper Award)
  • Approximate Statistical Tests for Comparing Supervised Classification Learning Algorithms
  • Moderator: Fernando
  • Room: G12-Blue

• Mon 30 Sep 2019

  • Paper: Weight Uncertainty in Neural Networks (Blundell et al. 2015) (http://proceedings.mlr.press/v37/blundell15.pdf)
  • Moderator: George
  • Room: G12-Blue

• Mon 8 Jul 2019

  • Paper: Tay et al. KDD 2018. Multi-Pointer Co-Attention Networks for Recommendation. (https://arxiv.org/abs/1801.09251)
  • Moderator: Haiyang
  • Room: G12-Blue

• **Mon 1 Jul 2019&&

  • Early Rumour Detection (Zhou et al., 2019) NAACL 2019. (https://www.aclweb.org/anthology/N19-1163)
  • Moderator: Yida
  • Room: G12-Blue

• Mon 24 Jun 2019

  • Paper: (Mao et al., 2018), ACL -- Word Embedding and WordNet Based Metaphor Identification and Interpretation
  • Moderator: Varvara
  • Room: G12-Blue

• **Mon 17 Jun 2019&&

  • Paper: What’s in a Name? Reducing Bias in Bios Without Access to Protected Attributes Alexey Romanov, Maria De-Arteaga, Hanna Wallach, Jennifer Chayes, Christian Borgs, Alexandra Chouldechova, Sahin Geyik, Krishnaram Kenthapadi, Anna Rumshisky and Adam Kalai (https://arxiv.org/abs/1904.05233v1)
  • Moderator: Zeerak
  • Room: TBA

• Mon 10 Jun 2019

  • TBD
  • Moderator: George
  • Room: G12-Blue

• Mon 3 Jun 2019

  • (*ACL) Paper writing tips
  • Moderator: Nikos
  • Room: G12-Blue

• Fri 17 May 2019

  • EMNLP Review session

• Mon 13 May 2019

  • Paper: (Lei et al., 2017), ArXiv -- Simple Recurrent Units for Highly Parallelizable Recurrence
  • Moderator: Abiola
  • Room: G25

• Mon 06 May 2019 Cancelled: Early May bank holiday

  • Paper: TBA
  • Moderator: TBA
  • Room: TBA

• Mon 29 Apr 2019

  • Paper: TBA
  • Moderator: Albiola
  • Room: TBA

• Mon 22 Apr 2019 Cancelled: Easter Monday

  • Paper: TBA
  • Moderator: TBA
  • Room: TBA

• Mon 15 Apr 2019

  • Paper: Tolga Bolukbasi, Kai-Wei Chang, James Zou, Venkatesh Saligrama, and Adam Kalai. 2016. Man is to computer programmer as woman is to homemaker? Debiasing word embeddings. (https://dl.acm.org/citation.cfm?id=3157584)
  • Moderator: Alison
  • Room: TBA

• Mon 08 Apr 2019

  • Paper: Identifying and understanding user reactions to deceptive and trusted social news sources (https://arxiv.org/pdf/1805.12032.pdf) How Humans versus Bots React to Deceptive and Trusted News Sources: A Case Study of Active Users (https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8508700)
  • Moderator: Yida
  • Room: G12-Blue

• Mon 01 Apr 2019

  • Paper: Sarthak Jain and Byron C. Wallace (ArXiv) [Attention is not Explanation] (https://arxiv.org/pdf/1902.10186.pdf)
  • Moderator: George
  • Room: G12-Blue (Lewing Lab)

• Mon 25 Mar 2019

  • Paper: Rebekah Overdorf, Bogdan Kulynych, Ero Balsa, Carmela Troncoso, Seda Gürses (ArXiv) POTs: Protective Optimization Technologies
  • Moderator: Zeerak
  • Room: G12-Blue (Lewin Lab)

• Mon 18 Mar 2019

  • Paper: Coherence Aware Topic Modeling (EMNLP 2018)
  • Moderator: Areej
  • Room: G22-Orange

• Mon 11 Mar 2019

  • Paper: Gregor Wiedemann, Eugen Ruppert, Raghav Jindal, Chris Biemann (GermEval, 2018) Transfer Learning from LDA to BiLSTM-CNN for Offensive Language Detection in Twitter
  • Moderator: Cass
  • Room: G12-Blue (Lewin Lab)

• Fri 01 Mar 2019

  • ACL Review Session
  • Time: 13:00-15:00
  • Room: G12-Blue (Lewin Lab)

• Mon 25 Feb 2019

  • Paper: (Leila Arras et al, 2017), ACL -- Explaining Recurrent Neural Network Predictions in Sentiment Analysis
  • Moderator: George
  • Room: G25

• Mon 18 Feb 2019

  • Paper: (Dehghani et al., 2019), ArXiv -- Universal Transformers
  • Moderator: Hardy
  • Room: G25

• Mon 11 Feb 2019

  • Paper: (Houlsby et al., 2019), ArXiv -- Parameter-Efficient Transfer Learning for NLP
  • Moderator: Fred
  • Room: G12-Blue

• Mon 04 Feb 2019

  • Paper: Howard and Ruder (ACL, 2018) Universal Language Model Fine-tuning for Text Classification
  • Moderator: Nikos
  • Room: G25

• Mon 28 Jan 2019

  • Paper: Surya et al. (2019). Unsupervised Neural Text Simplification
  • Moderator: Fernando
  • Room: G25

• Wed 12 Dec 2018

  • Paper:
    • Wu etal (2018): Word Mover's Embedding: From Word2Vec to Document Embedding https://arxiv.org/abs/1811.01713
    • Wu etal (2018): D2KE: From Distance to Kernel and Embedding
  • Moderator: Johann
  • Room: COM-G25

• Wed 05 Dec 2018

  • NAACL submissions review (deadline: Dec 10)
  • Room: G-25

• Wed 28 Nov 2018

  • Paper: I'd like to try and look at the following three because they are all somewhat related, but with the focus on BERT and the transformer-based architectures and representations:
    • Devlin etal (2018): BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding https://arxiv.org/abs/1810.04805
    • Cer etal (2018): Universal Sentence Encoder https://arxiv.org/abs/1803.11175
    • Yang etal (2018): Learning Semantic Textual Similarity from Conversations https://arxiv.org/abs/1804.07754
  • Note: it will help to know the paper: Vaswani etal (2017): Attention is all you need https://arxiv.org/abs/1706.03762
  • Moderator: Johann
  • Room: COM-G25

• Wed 21 Nov 2018

  • Paper: Peters et al (2018), Deep contextualized word representations, NAACL 2018
  • Moderator: Carol
  • Room: COM-G25
  • Related papers mentioned:
    • word senses in embeddings: Arora etal 2016: Linear Algebraic Structure of Word Senses, with Applications to Polysemy. http://arxiv.org/abs/1601.03764
    • combining different vector spaces: Coates and Bollegala 2018: Frustratingly Easy Meta-Embedding -- Computing Meta-Embeddings by Averaging Source Word Embeddings http://aclweb.org/anthology/N18-2031
    • BPEs for MT: Sennrich et al. (2016): Neural Machine Translation of Rare Words with Subword Units. http://www.aclweb.org/anthology/P16-1162

• Wed 14 Nov 2018

  • Paper: EMNLP recap: papers presented
  • Moderator: Zeerak
  • Room: COM-G22 Blue

• Wed 07 Nov 2018

  • NLP seminar (RG is postponed)

• Wed 31 Oct 2018

  • Paper: Zheng et al (2018), Multi-Reference Training with Pseudo-References for Neural Translation and Text Generation, in EMNLP 2018
  • Moderator: Makis
  • Room: COM-G25

• Wed 24 Oct 2018

  • Paper: Xing and Paul (2018), Diagnosing and Improving Topic Models by Analyzing Posterior Variability, In AAAI
  • Moderator: Areej
  • Room: COM-G25

• Wed 17 Oct 2018

  • Paper: Dong, Quirk and Lapata (2018), Confidence Modeling for Neural Semantic Parsing, In ACL
  • Moderator: Nikos
  • Room: COM-G25

• Wed 10 Oct 2018

  • Kick-off meeting
  • Room: COM-G22 Blue