Skip to main content ->
Ai2

Research - Papers

Explore a selection of our published work on a variety of key research challenges in AI.

Filter papers

NeuroLogic A*esque Decoding: Constrained Text Generation with Lookahead Heuristics

Ximing LuS. WelleckPeter WestYejin Choi
2022
NAACL

The dominant paradigm for neural text generation is left-to-right decoding from autoregressive language models. Constrained or controllable generation under complex lexical constraints, however,… 

Time Waits for No One! Analysis and Challenges of Temporal Misalignment

Kelvin LuuDaniel KhashabiSuchin GururanganNoah A. Smith
2022
NAACL

When an NLP model is trained on text data from one time period and tested or deployed on data from another, the resulting temporal misalignment can degrade end-task performance. In this work, we… 

Transparent Human Evaluation for Image Captioning

Jungo KasaiKeisuke SakaguchiLavinia DunaganNoah A. Smith
2022
NAACL

We establish a rubric-based human evaluation protocol for image captioning models. Our scoring rubrics and their definitions are carefully developed based on machineand humangenerated captions on… 

Data Governance in the Age of Large-Scale Data-Driven Language Technology

Yacine JerniteHuu NguyenStella Rose BidermanMargaret Mitchell
2022
FAccT

The recent emergence and adoption of Machine Learning technology, and specifically of Large Language Models, has drawn attention to the need for systematic and transparent management of language… 

Measuring the Carbon Intensity of AI in Cloud Instances

Jesse DodgeTaylor PrewittRémi Tachet des CombesWill Buchanan
2022
FAccT

The advent of cloud computing has provided people around the world with unprecedented access to computational power and enabled rapid growth in technologies such as machine learning, the… 

Domain Mismatch Doesn’t Always Prevent Cross-Lingual Transfer Learning

Daniel EdmistonPhillip KeungNoah A. Smith
2022
LREC

Cross-lingual transfer learning without labeled target language data or parallel text has been surprisingly effective in zero-shot cross-lingual classification, question answering, unsupervised… 

What Language Model to Train if You Have One Million GPU Hours?

Teven Le ScaoThomas WangDaniel HesslowIz Beltagy
2022
EMNLP

The crystallization of modeling methods around the Transformer architecture has been a boon for practitioners. Simple, well-motivated architectural variations that transfer across tasks and scale,… 

Retrieval Data Augmentation Informed by Downstream Question Answering Performance

James FergusonPradeep DasigiTushar KhotHannaneh Hajishirzi
2022
ACL • FEVER

Training retrieval models to fetch contexts for Question Answering (QA) over large corpora requires labeling relevant passages in those corpora. Since obtaining exhaustive manual annotations of all… 

NaturalProver: Grounded Mathematical Proof Generation with Language Models

S. WelleckJiacheng LiuXiming LuYejin Choi
2022
arXiv

Theorem proving in natural mathematical language - the mixture of symbolic and natural language used by humans - plays a central role in mathematical advances and education, and tests aspects of… 

Zero- and Few-Shot NLP with Pretrained Language Models

Iz BeltagyArman CohanRobert Logan IVSameer Singh
2022
ACL, tutorial

The ability to efficiently learn from little-to-no data is critical to applying NLP to tasks where data collection is costly or otherwise difficult. This is a challenging setting both academically…