Xian Li

Multilingual Speech Translation with Efficient Finetuning of Pretrained Models

2020 · arXiv (Cornell University)

We present a simple yet effective approach to build multilingual speech-to-text (ST) translation by efficient transfer learning from pretrained speech encoder and text decoder. Our key finding is that a minimalistic LNA (LayerNorm and Attention) …

Towards Understanding the Behaviors of Optimal Deep Active Learning Algorithms

2020 · arXiv (Cornell University)

Active learning (AL) algorithms may achieve better performance with fewer data because the model guides the data selection process. While many algorithms have been proposed, there is little study on what the optimal AL algorithm …

Lifting the Curse of Multilinguality by Pre-training Modular Transformers

2022 · arXiv (Cornell University)

Multilingual pre-trained models are known to suffer from the curse of multilinguality, which causes per-language performance to drop as they cover more languages. We address this issue by introducing language-specific modules, which allows us to …

PGE

2022 · Proceedings of the VLDB Endowment

Although product graphs (PGs) have gained increasing attentions in recent years for their successful applications in product search and recommendations, the extensive power of PGs can be limited by the inevitable involvement of various kinds …

Deep Voice: Real-time Neural Text-to-Speech

2017 · arXiv (Cornell University)

We present Deep Voice, a production-quality text-to-speech system constructed entirely from deep neural networks. Deep Voice lays the groundwork for truly end-to-end neural speech synthesis. The system comprises five major building blocks: a segmentation model …

On Evaluation of Adversarial Perturbations for Sequence-to-Sequence Models

2019

Paul Michel, Xian Li, Graham Neubig, Juan Pino. Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers). 2019.

Multilingual Denoising Pre-training for Neural Machine Translation

2020 · Transactions of the Association for Computational Linguistics

This paper demonstrates that multilingual denoising pre-training produces significant performance gains across a wide variety of machine translation (MT) tasks. We present mBART—a sequence-to-sequence denoising auto-encoder pre-trained on large-scale monolingual corpora in many languages using …

Analysing Off-The-Shelf Options for Question Answering with Portuguese FAQs

2022 · DROPS (Schloss Dagstuhl – Leibniz Center for Informatics)

Following the current interest in developing automatic question answering systems, we analyse alternative approaches for finding suitable answers from a list of Frequently Asked Questions (FAQs), in Portuguese. These rely on different technologies, some more …

OPT-IML: Scaling Language Model Instruction Meta Learning through the Lens of Generalization

2022 · arXiv (Cornell University)

Recent work has shown that fine-tuning large pre-trained language models on a collection of tasks described via instructions, a.k.a. instruction-tuning, improves their zero and few-shot generalization to unseen tasks. However, there is a limited understanding …