Artificial Intelligence (AI) is making a profound impact in almost every domain. A vital enabler
of its great success is the availability of abundant and high-quality data for building machine …

Annotated data plays a critical role in Natural Language Processing (NLP) in training
models and evaluating their performance. Given recent developments in Large Language …

Weakly-supervised learning (WSL) has shown promising results in addressing label scarcity
on many NLP tasks, but manually designing a comprehensive, high-quality labeling rule set …

We propose a new strategy for applying large pre-trained language models to novel tasks
when labeled training data is limited. Rather than apply the model in a typical zero-shot or …

To leverage the copious amount of data from source tasks and overcome the scarcity of the
target task samples, representation learning based on multi-task pretraining has become a …

We present PATRON, a prompt-based data selection method for pre-trained language
model fine-tuning under cold-start scenarios, ie, no initial labeled data are available. In …

Labeling training data is a critical and expensive step in producing high accuracy ML
models, whether training from scratch or fine-tuning. To make labeling more efficient, two …

Constructing supervised machine learning models for real-world video analysis require
substantial labeled data, which is costly to acquire due to scarce domain expertise and …

Large language models (LLMs) have demonstrated remarkable capabilities out of box for a
wide range of applications, yet accuracy still remains a major growth area, especially in …

Large Language Models have demonstrated remarkable few-shot performance, but the
performance can be sensitive to the selection of few-shot instances. We propose PATRON, a …