Skip to search formSkip to main contentSkip to account menu
  • Publications
  • Influence
DExperts: Decoding-Time Controlled Text Generation with Experts and Anti-Experts
TLDR
This work highlights the promise of tuning small LMs on text with (un)desirable attributes for efficient decoding-time steering and applies DExperts to language detoxification and sentiment-controlled generation, where it outperform existing controllable generation methods on both automatic and human evaluations.
View on ACL
Multi-sense Definition Modeling using Word Sense Decompositions
TLDR
This paper introduces a new method that enables definition modeling for multiple senses of the same word, and shows how a Gumble-Softmax approach outperforms baselines at matching sense-specific embeddings to definitions during training.
View PDF on arXiv
WANLI: Worker and AI Collaboration for Natural Language Inference Dataset Creation
TLDR
This work introduces a novel approach for dataset creation based on worker and AI collaboration, which brings together the generative strength of language models and the evaluative power of humans.
View PDF on arXiv
On-the-Fly Controlled Text Generation with Experts and Anti-Experts
TLDR
This work proposes DEXPERTS: Decoding-time Experts, a decodingtime method for controlled text generation which combines a pretrained language model with “experts” and/or “anti-experts" in an ensemble of language models.
Generated Knowledge Prompting for Commonsense Reasoning
TLDR
Generated knowledge prompting develops generated knowledge prompting, which consists of generating knowledge from a language model, then providing the knowledge as additional input when answering a question, and improves performance of large-scale, state-of-the-art models on four commonsense reasoning tasks.
View on ACL
Incorporating Music Knowledge in Continual Dataset Augmentation for Music Generation
TLDR
Aug-Gen is presented, a method of dataset augmentation for any music generation system trained on a resource-constrained domain that shows that the training data for a generative system can be augmented by examples the system produces during the course of training, provided these examples are of sufficiently high quality and variety.
View PDF on arXiv
Model selection for deep audio source separation via clustering analysis
TLDR
This work presents a confidence measure that does not require ground truth to estimate quality, given a deep model and audio mixture and uses this confidence measure to automatically select the model output with the best predicted separation quality.
View PDF on arXiv
Bach or Mock? A Grading Function for Chorales in the Style of J.S. Bach
TLDR
This paper introduces a grading function that evaluates four-part chorales in the style of J.S. Bach along important musical features, and shows that the function is both interpretable and outperforms human experts at discriminating Bach chorale from model-generated ones.
View PDF on arXiv
Abstract 3030: Integrated in vitro and in silico screening of breast cancer reveals serine metabolism as a precision oncology target
TLDR
Integrated in vitro and in silico screening of breast cancer reveals serine metabolism as a precision oncology target and demonstrates that specific metabolic targets can be identified through this integrated screening.
View via Publisher
Comparison of Discourse Surrounding CRISPR/Cas9 in the Media and Peer-Re- viewed Literature
TLDR
Evaluating how the discussion of CRISPR/Cas9 in the mainstream media reflects and compares to that of the academic literature can give an understanding of the similarities and differences between expert and public discussion on the topic of CRisPR/ Cas9 and gene editing more broadly.
By clicking accept or continuing to use the site, you agree to the terms outlined in our Privacy Policy, Terms of Service, and Dataset License