Final Project Reports for 2020

There were two options for the course project. Students either chose their own topic ("Custom Project"), or took part in a competition to build Question Answering models for the SQuAD 2.0 challenge ("Default Project"). This year saw fewer final projects than past years due to challenges caused by COVID-19.

You can find links to previous years' reports under Previous Offerings on the homepage.

Prize Winners

Congratulations to the following teams, who produced exceptional prize-winning projects! Thank you also to our sponsors, who generously donated prizes.

Prizes for Best Custom Project Reports

Prizes for Best Default Project Reports

Custom Projects

Project NameAuthors
Context-Based Models for Sarcasm Detection Nicholas Nestor Benavides, Angelo Reyes Ramos
Wikipedia Simplifier Daniel Kharitonov
Measuring Impact and Consumption of Local News in a Changing Environment Yu-chen Tuan, Austin Pennington, Anish Saha
Inclusivity-Exclusivity Inference Using Recurrent Neural Networks Santosh Murugan, Elissa Yiyi Li
Enhancing Arabic Diacritization with Word-Ending Specific Models Ahmed Abdallah Shuaibi, Hassan Hossam Fahmy
Music Recommendation for Podcast Scripts: Detecting Emotion from Text Anne Liu Lee, Shreya Ravi, Alex Ping Tsun
Couplet Scoring and Prediction using ALBERT Ju Huo
Can FOMC Minutes Predict the Federal Funds Rate? Ye Ye
DeepSumm - Deep Code Summaries using Neural Transformer Architecture Vivek Kumar Gupta
DeepClassic: Music Generation with Neural Networks Raphael Andreas Rene Abbou
How much does pre-trained information help? Partially re-initializing BERT during fine-tuning to analyze the contribution of layers Davide Giovanardi, Trisha Singh
Sentence Boundary Detection in Legal Texts Krithika Iyer
Keyword-based Code Auto-Complete Rui Aguiar, Faraz Fadavi Akhavan Bonab
Evaluation Metrics for Split and Rephrase Rui Yan
Idiomatic Language Translation and Transfer Learning David Andres Estrada-Arias, Jupinder Parmar
Improvement of deep learning techniques for ICD-9 code classification using MIMIC-III medical notes Derek Russell Jow, Anthony Li, Soham Uday Gadgil
DrCoding: Using Transformers to Predict ICD-9 Codes from Discharge Summaries Tom Jin
You Don’t Sound Like Me: Analyzing Hiring Based on Pre-hire Language Use Samaksh Ajay Goyal
Detecting Bias in Lending Data with NLP Models Orion Earl Darley, Senthil Selvaraj
Speech Recognition for Accented English Sasha Harrison, Claire Marika Pajot
Predicting ICD-9 Codes from Medical Notes - Does the Magic of BERT Applies Here? Yiyun Chen
Multilingual Customized Bert for Zero Shot Sequence Classification Yuehao Wu
Combining NLP and Computer Vision to Help Blind People Volha Leusha
Synth-tax: Modification of BERT Syntax Representations via Structural Probes Benjamin Louis Newman, Julia Xiaozhe Gong, Kai-Siang Ang
ClinicalBertSum: RCT Summarization by Using Clinical BERT Embeddings Xiaomeng Jin, Zihan Wang, Mingyi Lu
What errors are made by TyDi QA systems? Nivedita Rahurkar
More Abstractive Summarization with Pointer-Generator Networks Michael Andres Lin
Generate Symbolic Plans for Long-Horizon Robot Manipulation Tasks Karen Yang
DAMsL: A meta-learning based approach for dialogue state tracking Avanika Narayan, Josh Hedtke
Improving Language Generation with Sentence Coherence Objective Jie Yang, Ruixiao Sun, Mehrdad Yousefzadeh
Document Classification with DocBERT, et. Al. Dale Angus
BERT Goes to College: Understanding the Role of Pretrained Layers During Finetuning Daniel Huang, Jacky Lin, Shan Zhou
Inside-Out Code Auto-Completion Lauren Gillespie, Gabriel Poesia Reis e Silva, Scott Viteri
Ensemble-Learning for Sustainable NLP Surya Narayanan Hari, Elena Berman
Man vs. Machine: Master Debating with Argumentative Dialogue Jason Jin Zhao
What is a Fact? Juliette Elizabeth Love, Kevin Ji, Moritz Pascal Stephan
Deep learning-based detection of Dysarthric speech disability Siddhartha Prakash
Fine-Tuned BERT for the Detection of Political Ideology Maria Del Mar Castanos, Alexandre Simoes Gomes Junior
Modeling Long-Distance Dependencies with Second-Order LSTMs Yifeng Lin, Sam Alan Spinner, Noa Flora Bendit-Shtull
Hierarchical Reward Shaping Reinforcement Learning for Paragraph Generation Heejung Chung, Alex Nam
Plot Holes and Text Topology Egor Alimpiev, Vivek Robert Myers
Simple Transformers for PHI De-identification Aditya Khandelwal, Arjun Soin
Evaluating a relationship between mental health and wearable sensors using NLP Artem Alex Trotsyuk
Typologically Diverse QA: How many training examples do you need for a new language anyway? Caterina Wu, Tassica Linmei Lim, Tatiana Wu
A Comparative Study of Neural Question Generation Models Ankit Dwivedi, Arunothia Marappan
Know When To Abstain: Calibrating Question Answering System under Domain Shift Wanze Xie, Ruocheng Wang
A Close Examination of Factual Correctness Evaluation in Abstractive Summarization Yuhui Zhang
Second-order RNNs: Resurgence of Recurrence Shreya Singh, Sundararajan Renganathan

Default Projects

Project NameAuthors
Reimplementing QANet Adrian Ng
Applying Computer Vision Methodologies to Combat Adversarial Inputs for QA Reading Comprehension Stephanie Danielle Brito, Derek Bryan McCreight
Question Answering on SQuAD 2.0 Han Wu, Kevin Han, Yuqi Jin
Understanding and Improving Performance of QA Systems on Adversarial Evaluation Shai Limonchik, Andy Lapastora
Question Answering with Gated Attention and Multitask Learning Mingyang Ling, Lam Wing Chan
Applying and Adapting the Reformer as a Computationally Efficient Approach to the SQuAD 2.0 Question-Answering Task Hugo Clifford Kitano, Taide Ding
SQuAD 2.0 Based on ALBERT and Ensemble Yanpei Tian
Ensemble ALBERT on SQuAD 2.0 Veronica Peng, Renee Li, Allan Li
Intermediate Pretrained Contextual Embedding Models With Applications in Question Answering Ryan Silva
ALBERT Premium for SQuAD 2.0 Yibing Du, Yanhao Jiang
Answerability Verification in SQuAD Thatcher Solomon Freeman, Albert Zuo
Tackling SQuAD 2.0 with Ensemble Methods and Data Augmentation Elijah Amir Freeman, Emanuel Pinilla, Nikka Michelle Mofid
QANet implementation with multiple embedding methods for SQuAD Ying Guo, Jiachen Xu
Compressed SQuAD 2.0 Model With BERT Wenli Looi, Andrew Lawrence Deng
A PyTorch implementation of QANet Dat Pham Nguyen
Rare and novel words in SQuAD 2.0: An application of the form-context-model Axel Gross-Klussmann
Transformer-Based Models for SQuAD 2.0 Yuyan Wang, Lingjue Xie, Alex Xiaotong Gui
Machine Reading Comprehension with Abstention Verification Rakesh Kumar Grewal, Mara Rachel Finkelstein, Abhishek Rawat
Question Answering on Adversarial SQuAD 2.0 Ryan Guan
Neural Question Generation: Transformed Jeffrey Hu
Bert Mohith Rajendra
FusionNet: working smarter, not harder with SQuAD Sebastian Hurubaru, Francois Xavier Alain Henri Chesnay
Transferability of Contextual Representations for Question Answering Pranav Bhardwaj, Travis McGuire