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
...
Various techniques have been developed in recent years to improve dense
...
Existing language models (LMs) predict tokens with a softmax over a fini...
Multi-vector retrieval methods combine the merits of sparse (e.g. BM25) ...
We study the problem of retrieval with instructions, where users of a
re...
Existing hybrid retrievers which integrate sparse and dense retrievers, ...
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...
We propose pre-finetuning, an additional large-scale learning stage betw...
We review the EfficientQA competition from NeurIPS 2020. The competition...
We study open-domain question answering (ODQA) with structured, unstruct...
Task-oriented semantic parsing is a critical component of virtual assist...
Modern natural language processing and understanding applications have
e...
Multilingual Word Embeddings (MWEs) represent words from multiple langua...
Many text classification tasks are known to be highly domain-dependent.
...
In recent years deep neural networks have achieved great success in sent...