Weight-sharing supernet has become a vital component for performance
est...
Ternary and binary neural networks enable multiplication-free computatio...
Several post-training quantization methods have been applied to large
la...
3D human modeling has been widely used for engaging interaction in gamin...
We propose a new two-stage pre-training framework for video-to-text
gene...
There is growing interest in searching for information from large video
...
Latent diffusion models for image generation have crossed a quality thre...
Indoor scene synthesis involves automatically picking and placing furnit...
Various techniques have been developed in recent years to improve dense
...
Multi-vector retrieval methods combine the merits of sparse (e.g. BM25) ...
Building dense retrievers requires a series of standard procedures, incl...
Existing hybrid retrievers which integrate sparse and dense retrievers, ...
Modern pre-trained transformers have rapidly advanced the state-of-the-a...
In order to address the increasing demands of real-world applications, t...
We propose DrBoost, a dense retrieval ensemble inspired by boosting. DrB...
Many NLP tasks require processing long contexts beyond the length limit ...
With the rise of large-scale pre-trained language models, open-domain
qu...
Despite their recent popularity and well known advantages, dense retriev...
Pre-training on larger datasets with ever increasing model size is now a...
Question answering (QA) is an important use case on voice assistants. A
...
We review the EfficientQA competition from NeurIPS 2020. The competition...
Retrieving relevant contexts from a large corpus is a crucial step for t...
Structured information is an important knowledge source for automatic
ve...
We study open-domain question answering (ODQA) with structured, unstruct...
We propose a simple and efficient multi-hop dense retrieval approach for...
Open-domain question answering relies on efficient passage retrieval to
...
Question answering (QA) models have shown rapid progress enabled by the
...
The scarcity of labeled training data often prohibits the
internationali...
We introduce PyText - a deep learning based NLP modeling framework built...