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

FaVIQ: FAct Verification from Information-seeking Questions

Jungsoo ParkSewon MinJaewoo KangHannaneh Hajishirzi
2022
ACL

Despite significant interest in developing general purpose fact checking models, it is challenging to construct a large-scale fact verification dataset with realistic real-world claims. Existing… 

MetaICL: Learning to Learn In Context

Sewon MinM. LewisLuke ZettlemoyerHannaneh Hajishirzi
2022
NAACL

We introduce MetaICL (Meta-training for In-Context Learning), a new meta-training framework for few-shot learning where a pretrained language model is tuned to do in-context learning on a large set… 

Noisy Channel Language Model Prompting for Few-Shot Text Classification

Sewon MinMichael LewisHannaneh HajishirziLuke Zettlemoyer
2022
ACL

We introduce a noisy channel approach for language model prompting in few-shot text classification. Instead of computing the likelihood of the label given the input (referred as direct models),… 

Robust fine-tuning of zero-shot models

Mitchell WortsmanGabriel IlharcoMike LiLudwig Schmidt
2022
CVPR

Large pre-trained models such as CLIP or ALIGN offer consistent accuracy across a range of data distributions when performing zero-shot inference (i.e., without fine-tuning on a specific dataset).… 

Annotators with Attitudes: How Annotator Beliefs And Identities Bias Toxic Language Detection

Maarten SapSwabha SwayamdiptaLaura ViannaNoah A. Smith
2022
NAACL

Warning : this paper discusses and contains content that is offensive or upsetting. The perceived toxicity of language can vary based on someone’s identity and beliefs, but this variation is often… 

Bidimensional Leaderboards: Generate and Evaluate Language Hand in Hand

Jungo KasaiKeisuke SakaguchiRonan Le BrasNoah A. Smith
2022
NAACL

Natural language processing researchers have identified limitations of evaluation methodology for generation tasks, with new questions raised about the validity of automatic metrics and of… 

DEMix Layers: Disentangling Domains for Modular Language Modeling

Suchin GururanganMichael LewisAri HoltzmanLuke Zettlemoyer
2022
NAACL

We introduce a new domain expert mixture (DEMIX) layer that enables conditioning a language model (LM) on the domain of the input text. A DEMIX layer is a collection of expert feedforward networks,… 

Efficient Hierarchical Domain Adaptation for Pretrained Language Models

Alexandra ChronopoulouMatthew E. PetersJesse Dodge
2022
NAACL

The remarkable success of large language models has been driven by dense models trained on massive unlabeled, unstructured corpora. These corpora typically contain text from diverse, heterogeneous… 

Few-Shot Self-Rationalization with Natural Language Prompts

Ana MarasovićIz BeltagyDoug DowneyMatthew E. Peters
2022
Findings of NAACL

Self-rationalization models that predict task labels and generate free-text elaborations for their predictions could enable more intuitive interaction with NLP systems. These models are, however,… 

MultiVerS: Improving scientific claim verification with weak supervision and full-document context

David WaddenKyle LoLucy Lu WangHannaneh Hajishirzi
2022
Findings of NAACL

The scientific claim verification task requires an NLP system to label scientific documents which Support or Refute an input claim, and to select evidentiary sentences (or rationales) justifying…