1 |
Delving Deeper into Cross-lingual Visual Question Answering ...
|
|
|
|
BASE
|
|
Show details
|
|
2 |
Combating Temporal Drift in Crisis with Adapted Embeddings ...
|
|
|
|
BASE
|
|
Show details
|
|
3 |
Annotation Curricula to Implicitly Train Non-Expert Annotators ...
|
|
|
|
BASE
|
|
Show details
|
|
4 |
Smelting Gold and Silver for Improved Multilingual AMR-to-Text Generation ...
|
|
|
|
BASE
|
|
Show details
|
|
6 |
BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models ...
|
|
|
|
BASE
|
|
Show details
|
|
7 |
Avoiding Inference Heuristics in Few-shot Prompt-based Finetuning ...
|
|
|
|
BASE
|
|
Show details
|
|
9 |
GPL: Generative Pseudo Labeling for Unsupervised Domain Adaptation of Dense Retrieval ...
|
|
|
|
BASE
|
|
Show details
|
|
10 |
Modeling Global and Local Node Contexts for Text Generation from Knowledge Graphs
|
|
|
|
In: EISSN: 2307-387X ; Transactions of the Association for Computational Linguistics ; https://hal.archives-ouvertes.fr/hal-03020314 ; Transactions of the Association for Computational Linguistics, The MIT Press, 2020, 8, ⟨10.1162/tacl_a_00332⟩ (2020)
|
|
BASE
|
|
Show details
|
|
11 |
Making Monolingual Sentence Embeddings Multilingual using Knowledge Distillation ...
|
|
|
|
Abstract:
We present an easy and efficient method to extend existing sentence embedding models to new languages. This allows to create multilingual versions from previously monolingual models. The training is based on the idea that a translated sentence should be mapped to the same location in the vector space as the original sentence. We use the original (monolingual) model to generate sentence embeddings for the source language and then train a new system on translated sentences to mimic the original model. Compared to other methods for training multilingual sentence embeddings, this approach has several advantages: It is easy to extend existing models with relatively few samples to new languages, it is easier to ensure desired properties for the vector space, and the hardware requirements for training is lower. We demonstrate the effectiveness of our approach for 50+ languages from various language families. Code to extend sentence embeddings models to more than 400 languages is publicly available. ... : Accepted at EMNLP 2020 ...
|
|
Keyword:
Computation and Language cs.CL; FOS Computer and information sciences
|
|
URL: https://arxiv.org/abs/2004.09813 https://dx.doi.org/10.48550/arxiv.2004.09813
|
|
BASE
|
|
Hide details
|
|
12 |
How to Probe Sentence Embeddings in Low-Resource Languages: On Structural Design Choices for Probing Task Evaluation ...
|
|
|
|
BASE
|
|
Show details
|
|
13 |
MAD-X: An Adapter-Based Framework for Multi-Task Cross-Lingual Transfer ...
|
|
|
|
BASE
|
|
Show details
|
|
14 |
Predicting the Humorousness of Tweets Using Gaussian Process Preference Learning ...
|
|
|
|
BASE
|
|
Show details
|
|
15 |
How Good is Your Tokenizer? On the Monolingual Performance of Multilingual Language Models ...
|
|
|
|
BASE
|
|
Show details
|
|
16 |
UNKs Everywhere: Adapting Multilingual Language Models to New Scripts ...
|
|
|
|
BASE
|
|
Show details
|
|
17 |
PuzzLing Machines: A Challenge on Learning From Small Data ...
|
|
|
|
BASE
|
|
Show details
|
|
18 |
A Matter of Framing: The Impact of Linguistic Formalism on Probing Results ...
|
|
|
|
BASE
|
|
Show details
|
|
19 |
Modeling Global and Local Node Contexts for Text Generation from Knowledge Graphs ...
|
|
|
|
BASE
|
|
Show details
|
|
20 |
Empowering Active Learning to Jointly Optimize System and User Demands ...
|
|
|
|
BASE
|
|
Show details
|
|
|
|