Papers by Scott Yih
CCQA: A New Web-Scale Question Answering Dataset for Model Pre-Training (2022.findings-naacl)
Copied to clipboard
| Challenge: | Existing approaches to answer open domain questions rely on unlabeled text or synthetically generated question-answer pairs. |
| Approach: | They propose a large-scale open-domain question-answering dataset based on the Common Crawl project that can be used to in-domain pre-train popular language models. |
| Outcome: | The proposed dataset achieves promising results in zero-shot, low resource and fine-tuned settings across multiple tasks, models and benchmarks. |
FiD-Ex: Improving Sequence-to-Sequence Models for Extractive Rationale Generation (2021.emnlp-main)
Copied to clipboard
| Challenge: | Pre-trained sequence to sequence models are effective in making and generating NL explanations, but they have many shortcomings. |
| Approach: | They propose a model that uses sentence markers to eliminate explanation fabrication . they use fusion-in-decoder architecture to handle long input contexts . |
| Outcome: | The proposed model significantly improves on the ERASER explainability benchmark. |
DiffCSE: Difference-based Contrastive Learning for Sentence Embeddings (2022.naacl-main)
Copied to clipboard
Yung-Sung Chuang, Rumen Dangovski, Hongyin Luo, Yang Zhang, Shiyu Chang, Marin Soljacic, Shang-Wen Li, Scott Yih, Yoon Kim, James Glass
| Challenge: | Recent work shows that finetuning pretrained models with contrastive learning makes it possible to learn good sentence embeddings without labeled data. |
| Approach: | They propose an unsupervised contrastive learning framework for learning sentence embeddings . they use a masked language model to mask out the edited sentence . |
| Outcome: | The proposed framework outperforms SimCSE on semantic textual similarity tasks by 2.3 absolute points. |
Boosted Dense Retriever (2022.naacl-main)
Copied to clipboard
| Challenge: | DrBoost is a dense retrieval ensemble that is trained in stages to correct retrieval mistakes . it produces representations which are 4x more compact, while delivering comparable retrieval results. |
| Approach: | They propose a dense retrieval ensemble inspired by boosting that is trained in stages . they produce representations which are 4x more compact, while delivering comparable retrieval results . |
| Outcome: | The proposed model performs surprisingly well under approximate search with coarse quantization, reducing latency and bandwidth needs by another 4x. |
UniPELT: A Unified Framework for Parameter-Efficient Language Model Tuning (2022.acl-long)
Copied to clipboard
| Challenge: | Existing methods for parameter-efficient language model tuning (PELT) match the performance of fine-tuning with fewer trainable parameters. |
| Approach: | They propose a framework which integrates different PELT methods as submodules and learns to activate the ones that best suit the current data or task setup via gating mechanism. |
| Outcome: | The proposed framework outperforms fine-tuning methods on the GLUE benchmark and achieves 14% gains over the best individual PELT method. |
Adapting Pretrained Text-to-Text Models for Long Text Sequences (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing short-context models are limited in their domain coverage and can be used for long-sequence inputs. |
| Approach: | They propose to replace full attention in transformers with pooling-augmented blockwise attention and pretrain the model with a masked-span prediction task with spans of varying lengths. |
| Outcome: | The proposed model outperforms existing models on long-sequence summarization tasks and achieves competitive performance on long document corpora. |
UniK-QA: Unified Representations of Structured and Unstructured Knowledge for Open-Domain Question Answering (2022.findings-naacl)
Copied to clipboard
Barlas Oguz, Xilun Chen, Vladimir Karpukhin, Stan Peshterliev, Dmytro Okhonko, Michael Schlichtkrull, Sonal Gupta, Yashar Mehdad, Scott Yih
| Challenge: | a recent study aims to answer factual questions using a structured knowledge base (KBQA). |
| Approach: | They propose a unifying approach that homogenizes all knowledge sources by reducing them to text . they demonstrate that UniK-QA is a simple and yet effective way to combine heterogeneous sources of knowledge. |
| Outcome: | The proposed approach improves state-of-the-art results on knowledge-base QA tasks by 11 points compared to graph-based methods. |
On Continual Model Refinement in Out-of-Distribution Data Streams (2022.acl-long)
Copied to clipboard
| Challenge: | Existing continual learning (CL) problems cannot cover real-world scenarios such as out-of-distribution errors. |
| Approach: | They propose a continual model refinement problem formulation to solve this problem . they extend several existing continual learning approaches to the CMR problem based on a general sampling algorithm . |
| Outcome: | The proposed model refinement solution improves on existing models and their performance metrics. |
Simple Local Attentions Remain Competitive for Long-Context Tasks (2022.naacl-main)
Copied to clipboard
Wenhan Xiong, Barlas Oguz, Anchit Gupta, Xilun Chen, Diana Liskovich, Omer Levy, Scott Yih, Yashar Mehdad
| Challenge: | Existing models for NLP tasks require long text sequences beyond the length limit of pretrained models. |
| Approach: | They propose to pretrain large-size NLP models using the same long-doc corpus and fine tune them for real-world long-context tasks. |
| Outcome: | The proposed models can perform better under standard pretraining paradigms than longformer and Longformer. |
Domain-matched Pre-training Tasks for Dense Retrieval (2022.findings-naacl)
Copied to clipboard
Barlas Oguz, Kushal Lakhotia, Anchit Gupta, Patrick Lewis, Vladimir Karpukhin, Aleksandra Piktus, Xilun Chen, Sebastian Riedel, Scott Yih, Sonal Gupta, Yashar Mehdad
| Challenge: | Existing approaches to improve performance of pre-training tasks are needed. |
| Approach: | They propose to pre-train large bi-encoder models on a recently released set of 65 millionsynthetically generated questions and 200 million post-comment pairs from a preexisting reddit conversation dataset. |
| Outcome: | The proposed model can be pre-trained on a set of 65 millionsynthetically generated questions and 200 million post-comment pairs from a preexisting dataset of Reddit conversations. |